Modelling_Normal.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Modelling_Normal.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "environment": {
      "name": "common-cpu.m48",
      "type": "gcloud",
      "uri": "gcr.io/deeplearning-platform-release/base-cpu:m48"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ashishthomaschempolil/Santander-Customer-Satisfaction/blob/new_1/Modelling_Normal.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wkI7lsxLeC79",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        },
        "outputId": "ab763582-5b23-4886-f406-5937ee16d1be"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "QVQr_7USbczr",
        "colab": {}
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set(palette='muted',style='whitegrid')\n",
        "import sklearn\n",
        "from sklearn import metrics,ensemble,model_selection,linear_model,tree,calibration\n",
        "import xgboost as xgb\n",
        "import lightgbm as lgb\n",
        "import scipy\n",
        "import random\n",
        "from tqdm import tqdm\n",
        "import os\n",
        "np.random.seed(13154)"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "hRB4PxAYa5xu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        },
        "outputId": "8fc4c032-b0d9-4671-8ded-90046f1a9cb1"
      },
      "source": [
        "file = [k for k in os.listdir('/content/drive/My Drive/Santander_Customer_Satisfaction') if (('train' in k)|('test' in k))&('X' not in k)] #getting all the necessary files for modelling\n",
        "file"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['train_Normal.pkl',\n",
              " 'test_Normal.pkl',\n",
              " 'train_Normal with Response encoding.pkl',\n",
              " 'test_Normal with Response encoding.pkl',\n",
              " 'train_Normal with One Hot Encoding.pkl',\n",
              " 'test_Normal with One Hot Encoding.pkl',\n",
              " 'train_Log Transformed.pkl',\n",
              " 'test_Log Transformed.pkl',\n",
              " 'train_Log Transformed with Response encoding.pkl',\n",
              " 'test_Log Transformed with Response encoding.pkl',\n",
              " 'train_Log Transformed with One Hot Encoding.pkl',\n",
              " 'test_Log Transformed with One Hot Encoding.pkl']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "mV0hmG66S7ZC",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "d803ed3f-7523-4d16-df68-0ce8a35e31d8"
      },
      "source": [
        "file_path = '/content/drive/My Drive/Santander_Customer_Satisfaction/'\n",
        "train_file = file_path+'train_Normal.pkl' #Normal\n",
        "test_file = file_path+'test_Normal.pkl'\n",
        "dataset= \"Normal\"\n",
        "train = pd.read_pickle(train_file)\n",
        "test = pd.read_pickle(test_file)\n",
        "X_train = train.drop(['ID','TARGET'],axis=1)\n",
        "y_train = train['TARGET'].values\n",
        "X_test = test.drop('ID',axis=1)\n",
        "del train,test\n",
        "X_train,X_val,y_train,y_val = model_selection.train_test_split(X_train,y_train,stratify=y_train,test_size=0.15)\n",
        "X_train.shape,X_val.shape,X_test.shape"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((64617, 359), (11403, 359), (75818, 359))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "v4yLi9O_CtFR",
        "colab": {}
      },
      "source": [
        "global i\n",
        "i=0\n",
        "def plot_auc(y_true,y_pred,label,dataset=dataset):\n",
        "  \"\"\"\n",
        "  Plots auc curve when y_true and y_pred are given\n",
        "  dataset is string which tells which dataset were used.\n",
        "  label: tells which model was used\n",
        "  If label is a list plots all the auc curve for all labels (here y_pred will be a list of lists )\n",
        "  \"\"\"\n",
        "  \n",
        "  \n",
        "  if (type(label)!=list) & (type(label)!=np.array):\n",
        "    print(\"\\t\\t %s on %s dataset \\t\\t \\n\"%(label,dataset))\n",
        "    fpr,tpr,threshold = metrics.roc_curve(y_true,y_pred) #getting true positive rate, false positive rate\n",
        "    auc = metrics.roc_auc_score(y_true,y_pred) #getting auc score\n",
        "    log_loss = metrics.log_loss(y_true,y_pred) #getting logloss\n",
        "    label_1 = label+' AUC=%.3f'%(auc)\n",
        "\n",
        "    #plotting tpr vs fpr\n",
        "    sns.lineplot(fpr,tpr,label=label_1)\n",
        "    x=np.arange(0,1.1,0.1) #auc=0.5 line \n",
        "    sns.lineplot(x,x,label=\"AUC=0.5\")\n",
        "    plt.title(\"ROC on %s dataset\"%(dataset))\n",
        "    plt.xlabel('False Positive Rate')\n",
        "    plt.ylabel(\"True Positive Rate\")\n",
        "    plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.) #legend box outside graph\n",
        "    plt.show()\n",
        "    print(\"The log loss and auc was found to be %.3f and %.3f for %s on %s dataset.\"%(log_loss,auc,label,dataset))\n",
        "    #creating a dataframe of the result\n",
        "    result_dict = {\"Model\":label,'Dataset':dataset,'log_loss':log_loss,'auc':auc}\n",
        "    \n",
        "    return pd.DataFrame(result_dict,index=[i])\n",
        "\n",
        "  else:\n",
        "    plt.figure(figsize=(12,8))\n",
        "    for k,y in enumerate(y_pred):\n",
        "      fpr,tpr,threshold = metrics.roc_curve(y_true,y)\n",
        "      auc = metrics.roc_auc_score(y_true,y)\n",
        "      label_ = label[k]+' AUC=%.3f'%(auc)\n",
        "      #plotting tpr vs fpr\n",
        "      sns.lineplot(fpr,tpr,label=label_)\n",
        "\n",
        "    x=np.arange(0,1.1,0.1)\n",
        "    sns.lineplot(x,x,label=\"AUC=0.5\")\n",
        "    plt.title(\"Combined ROC\")\n",
        "    plt.xlabel('False Positive Rate')\n",
        "    plt.ylabel(\"True Positive Rate\")\n",
        "    plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.) #legend box outside graph\n",
        "    plt.show()"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "spCK7X7beIiy",
        "colab": {}
      },
      "source": [
        "def find_best_params(model,params,cv=10,n_jobs=-1,X_train=X_train):\n",
        "  \"\"\"\n",
        "  Funcion which implements Random Seacrh CV and returns best model\n",
        "  \"\"\"\n",
        "  random_cv = model_selection.RandomizedSearchCV(model,param_distributions=params,scoring='roc_auc',n_jobs=n_jobs,cv=cv,verbose=2)\n",
        "  random_cv.fit(X_train,y_train)\n",
        "  print(\"The best auc score was %.3f\"%(random_cv.best_score_))\n",
        "  print(\"The best params were: %s\"%(random_cv.best_params_))\n",
        "  return random_cv.best_estimator_"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WPbh5JzratIv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def plot_feature_importance(df,feat_imp,top=20,size=(6,6)):\n",
        "    \"\"\"\n",
        "    getting most important top features if df and feature_importance array is passed\n",
        "    \"\"\"\n",
        "    top_indices = np.argsort(feat_imp)[::-1][:top] #getting the indices with top feature importace\n",
        "    most_important_feat = df.columns[top_indices] #getiing the columns names\n",
        "    plt.figure(figsize=size)\n",
        "    sns.barplot(x=feat_imp[top_indices],y=most_important_feat) #plotting feature importance \n",
        "    plt.title('Feature Importance')\n",
        "    plt.xlabel('Importance')\n",
        "    plt.ylabel(\"Feature names\")\n",
        "    plt.show()"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b5DrW1M0IDSQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_top_features(df,feat_imp,top=50):\n",
        "  \"\"\"\n",
        "  gets the top features.\n",
        "  if top is of integer dtype it will return top features,\n",
        "  if top is between 0 and 1, it will return features that have at least feature importance value= top\n",
        "  \"\"\"\n",
        "  if (top>0.)&(top<=1.): #getting the features that have feature importance value greater than top\n",
        "    feature_to_consider = [1 if i>=top else 0 for i in feat_imp]\n",
        "    most_important_feat = [i  for i,j in zip(df.columns,feature_to_consider) if (j==1)]#getiing the columns names\n",
        "  else:\n",
        "    top_indices = np.argsort(feat_imp)[::-1][:top] #getting the indices with top feature importace\n",
        "    most_important_feat = df.columns[top_indices] #getiing the columns names\n",
        "  return most_important_feat  "
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "vvLQsSsVAsUA"
      },
      "source": [
        "# Modeling"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "uLh2j2XvAxeO"
      },
      "source": [
        "# Logistic Regression"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "nsbNtvxAYB2O",
        "colab": {}
      },
      "source": [
        "model = linear_model.LogisticRegression(class_weight='balanced')\n",
        "params = {'penalty':['l2','l1'], 'C':[10.**i for i in np.arange(-3,3,1)], \n",
        "          'fit_intercept':[True,False] ,\n",
        "          }\n"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "mXTZa5oPYB2R",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        },
        "outputId": "baebe65a-c1a5-4b8c-d3da-ec38aaaf9b3b"
      },
      "source": [
        "%%time\n",
        "find_best_params(model,params)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 10 folds for each of 10 candidates, totalling 100 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done  37 tasks      | elapsed:  1.3min\n",
            "[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed:  4.2min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "The best auc score was 0.798\n",
            "The best params were: {'penalty': 'l2', 'fit_intercept': True, 'C': 0.001}\n",
            "CPU times: user 9.1 s, sys: 1.09 s, total: 10.2 s\n",
            "Wall time: 4min 14s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LogisticRegression(C=0.001, class_weight='balanced', dual=False,\n",
              "                   fit_intercept=True, intercept_scaling=1, l1_ratio=None,\n",
              "                   max_iter=100, multi_class='auto', n_jobs=None, penalty='l2',\n",
              "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
              "                   warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "eYp1oCU5YB2U",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        },
        "outputId": "d724f879-f82e-403f-f27a-b1c467b4e50c"
      },
      "source": [
        "model = linear_model.LogisticRegression(C=0.001, class_weight='balanced')\n",
        "\n",
        "model.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LogisticRegression(C=0.001, class_weight='balanced', dual=False,\n",
              "                   fit_intercept=True, intercept_scaling=1, l1_ratio=None,\n",
              "                   max_iter=100, multi_class='auto', n_jobs=None, penalty='l2',\n",
              "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
              "                   warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "DlDUHTI-YB2X",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "ee0f3cb6-9275-4beb-d10a-433979c06f8b"
      },
      "source": [
        "i=0\n",
        "labels=[] #storing all labels\n",
        "y_preds = [] #storing all y_pred\n",
        "y_pred = model.predict_proba(X_val)[:,1]\n",
        "label=\"Logisic Regression\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "result = plot_auc(y_val,y_pred,label,dataset)\n",
        "result_df = result\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t Logisic Regression on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.542 and 0.812 for Logisic Regression on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "0y9PvYUNX-t2"
      },
      "source": [
        "# Decision Trees"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "0kVyDscneGv3",
        "colab": {}
      },
      "source": [
        "model = tree.DecisionTreeClassifier(class_weight='balanced')\n",
        "params = { \n",
        "          'max_depth':[10,100,500,1000,-1], \n",
        "         'min_samples_split':[2,5,10,100,500], \n",
        "          'min_samples_leaf':[1,3,5,7,10],\n",
        "          'max_leaf_nodes':[100,500,1000,5000,-1] \n",
        "          }\n"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "OWmU9ZdnjHEk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "380951a8-d438-488d-aae3-2821de152a2f"
      },
      "source": [
        "%%time\n",
        "find_best_params(model,params)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 10 folds for each of 10 candidates, totalling 100 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done  37 tasks      | elapsed:  1.8min\n",
            "[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed:  5.9min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "The best auc score was 0.787\n",
            "The best params were: {'min_samples_split': 5, 'min_samples_leaf': 5, 'max_leaf_nodes': 100, 'max_depth': 10}\n",
            "CPU times: user 4.84 s, sys: 367 ms, total: 5.21 s\n",
            "Wall time: 5min 57s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeClassifier(ccp_alpha=0.0, class_weight='balanced', criterion='gini',\n",
              "                       max_depth=10, max_features=None, max_leaf_nodes=100,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=5, min_samples_split=5,\n",
              "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
              "                       random_state=None, splitter='best')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "11n9ZcTNirnF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "ca69e782-280d-4799-ea10-d266964e1244"
      },
      "source": [
        "model = tree.DecisionTreeClassifier(ccp_alpha=0.0, class_weight='balanced', criterion='gini',\n",
        "                       max_depth=10, max_features=None, max_leaf_nodes=500,\n",
        "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
        "                       min_samples_leaf=10, min_samples_split=5,\n",
        "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
        "                      splitter='best')\n",
        "\n",
        "model.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeClassifier(ccp_alpha=0.0, class_weight='balanced', criterion='gini',\n",
              "                       max_depth=10, max_features=None, max_leaf_nodes=500,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=10, min_samples_split=5,\n",
              "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
              "                       random_state=None, splitter='best')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "IqO7PiHUhFIR",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "outputId": "6bc4c02e-c62b-4156-f4a5-ccd44bb9a532"
      },
      "source": [
        "cc_model = calibration.CalibratedClassifierCV(model,cv='prefit')\n",
        "cc_model.fit(X_train,y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "CalibratedClassifierCV(base_estimator=DecisionTreeClassifier(ccp_alpha=0.0,\n",
              "                                                             class_weight='balanced',\n",
              "                                                             criterion='gini',\n",
              "                                                             max_depth=10,\n",
              "                                                             max_features=None,\n",
              "                                                             max_leaf_nodes=500,\n",
              "                                                             min_impurity_decrease=0.0,\n",
              "                                                             min_impurity_split=None,\n",
              "                                                             min_samples_leaf=10,\n",
              "                                                             min_samples_split=5,\n",
              "                                                             min_weight_fraction_leaf=0.0,\n",
              "                                                             presort='deprecated',\n",
              "                                                             random_state=None,\n",
              "                                                             splitter='best'),\n",
              "                       cv='prefit', method='sigmoid')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CIxYiOfoih44",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "1bb3fd5e-8d5b-4fa3-d2c1-2a46642943f3"
      },
      "source": [
        "i+=1\n",
        "y_pred = cc_model.predict_proba(X_val)[:,1]\n",
        "label=\"Decision Tree Classifier\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "result = plot_auc(y_val,y_pred,label)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t Decision Tree Classifier on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.148 and 0.779 for Decision Tree Classifier on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "qusVRO8HdpHW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "2aec827f-0c90-4ce7-e4d3-cda162e79922"
      },
      "source": [
        "result_df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                      Model Dataset  log_loss       auc\n",
              "0        Logisic Regression  Normal  0.542129  0.812101\n",
              "1  Decision Tree Classifier  Normal  0.148082  0.779135"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 72
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "msPtDUHT0sQN"
      },
      "source": [
        "# Random Forest"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "OwHc1zRG0sQP",
        "colab": {}
      },
      "source": [
        "model = ensemble.RandomForestClassifier(class_weight='balanced')\n",
        "params = { 'n_estimators':[1000,2000],\n",
        "          'max_depth':[1000,2000], \n",
        "         'min_samples_split':[100,500],\n",
        "          'min_samples_leaf':[3,5],\n",
        "          'max_leaf_nodes':[100,250] \n",
        "          }\n"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "zDOPEuJy0sQS",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "outputId": "1ce8b255-eba0-4019-e628-f7cbe0d7fc60"
      },
      "source": [
        "%%time\n",
        "find_best_params(model,params,cv=3)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 3 folds for each of 10 candidates, totalling 30 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done  30 out of  30 | elapsed: 74.5min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "The best auc score was 0.817\n",
            "The best params were: {'n_estimators': 2000, 'min_samples_split': 500, 'min_samples_leaf': 3, 'max_leaf_nodes': 250, 'max_depth': 1000}\n",
            "CPU times: user 6min 31s, sys: 1.01 s, total: 6min 32s\n",
            "Wall time: 1h 21min 4s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight='balanced',\n",
              "                       criterion='gini', max_depth=1000, max_features='auto',\n",
              "                       max_leaf_nodes=250, max_samples=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=3, min_samples_split=500,\n",
              "                       min_weight_fraction_leaf=0.0, n_estimators=2000,\n",
              "                       n_jobs=None, oob_score=False, random_state=None,\n",
              "                       verbose=0, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "RBeRKL7R0sQW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        },
        "outputId": "81ca2890-02a1-4e99-deb0-a4b169baaeb6"
      },
      "source": [
        "model = ensemble.RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight='balanced',\n",
        "                       criterion='gini', max_depth=2000, max_features='auto',\n",
        "                       max_leaf_nodes=250, max_samples=None,\n",
        "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
        "                       min_samples_leaf=3, min_samples_split=500,\n",
        "                       min_weight_fraction_leaf=0.0, n_estimators=2000,\n",
        "                       n_jobs=10, oob_score=False,random_state=42,\n",
        "                       verbose=0, warm_start=False)\n",
        "\n",
        "\n",
        "model.fit(X_train,y_train)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight='balanced',\n",
              "                       criterion='gini', max_depth=2000, max_features='auto',\n",
              "                       max_leaf_nodes=250, max_samples=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=3, min_samples_split=500,\n",
              "                       min_weight_fraction_leaf=0.0, n_estimators=2000,\n",
              "                       n_jobs=10, oob_score=False, random_state=42, verbose=0,\n",
              "                       warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "oJLF56Ts0sQY",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "9ced86b3-20ff-4134-c548-fffd10b586b1"
      },
      "source": [
        "cc_model = calibration.CalibratedClassifierCV(model,cv='prefit')\n",
        "cc_model.fit(X_train,y_train)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "CalibratedClassifierCV(base_estimator=RandomForestClassifier(bootstrap=True,\n",
              "                                                             ccp_alpha=0.0,\n",
              "                                                             class_weight='balanced',\n",
              "                                                             criterion='gini',\n",
              "                                                             max_depth=2000,\n",
              "                                                             max_features='auto',\n",
              "                                                             max_leaf_nodes=250,\n",
              "                                                             max_samples=None,\n",
              "                                                             min_impurity_decrease=0.0,\n",
              "                                                             min_impurity_split=None,\n",
              "                                                             min_samples_leaf=3,\n",
              "                                                             min_samples_split=500,\n",
              "                                                             min_weight_fraction_leaf=0.0,\n",
              "                                                             n_estimators=2000,\n",
              "                                                             n_jobs=10,\n",
              "                                                             oob_score=False,\n",
              "                                                             random_state=42,\n",
              "                                                             verbose=0,\n",
              "                                                             warm_start=False),\n",
              "                       cv='prefit', method='sigmoid')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4Jct69qq0sQa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "c5c5fb05-94ab-44ad-b01d-3d9b91e73561"
      },
      "source": [
        "i+=1\n",
        "y_pred = cc_model.predict_proba(X_val)[:,1]\n",
        "label=\"Random Forest Classifier\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "result = plot_auc(y_val,y_pred,label)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t Random Forest Classifier on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.137 and 0.833 for Random Forest Classifier on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "0igsmQs_0sQd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "outputId": "32779d8d-a051-43cf-f326-822faed7555e"
      },
      "source": [
        "result_df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137490</td>\n",
              "      <td>0.832512</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                      Model Dataset  log_loss       auc\n",
              "0        Logisic Regression  Normal  0.542129  0.812101\n",
              "1  Decision Tree Classifier  Normal  0.148082  0.779135\n",
              "2  Random Forest Classifier  Normal  0.137490  0.832512"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GZqVnGrga8Tu",
        "colab_type": "text"
      },
      "source": [
        "# Selecting top features"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tFB132azbBQS",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 736
        },
        "outputId": "8b742793-9c58-48f9-a8ed-1e6bfe9554ad"
      },
      "source": [
        "plot_feature_importance(X_train,model.feature_importances_,top=50,size=(7,12))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x864 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XxwkRX5gnXWi",
        "colab_type": "text"
      },
      "source": [
        "We will create new dataset with less features."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CfArYVshrZNh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "47db821d-163a-4447-b860-48c1b780653e"
      },
      "source": [
        "#getting top k features\n",
        "top=250\n",
        "features = get_top_features(X_train,model.feature_importances_,top=top)\n",
        "print(\"Taking top %i features\"%(top))\n",
        "X_train_200 = X_train[features].copy()\n",
        "X_val_200 = X_val[features].copy()\n",
        "X_test_200 = X_test[features].copy()"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Taking top 250 features\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "U85-m-LIEmmc"
      },
      "source": [
        "# XGB Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Ssc-G7LSEmme",
        "colab": {}
      },
      "source": [
        "model = xgb.XGBClassifier(n_jobs=-1,\n",
        "                      nthread=-1,\n",
        "                      scale_pos_weight=1.,\n",
        "                      learning_rate=0.01,  \n",
        "                      colsample_bytree = 0.5,\n",
        "                      subsample = 0.9,\n",
        "                      objective='binary:logistic', \n",
        "                      n_estimators=1000,\n",
        "                      reg_alpha = 0.3,\n",
        "                      max_depth=5, \n",
        "                      gamma=5,random_state=42\n",
        "                     )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T0CEyW37Dai9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "e6839ea9-fc26-42bf-efff-c82bd3458b54"
      },
      "source": [
        "#https://towardsdatascience.com/fine-tuning-xgboost-in-python-like-a-boss-b4543ed8b1e\n",
        "\"\"\"\n",
        "manually changed each parameters 1 by 1 and compared the validation auc\n",
        "and selected the best parameter value\n",
        "\"\"\"\n",
        "eval_metric = ['error','auc']\n",
        "%time model.fit(X_train,y_train,eval_set=[(X_val,y_val)],eval_metric=eval_metric,verbose=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.789452\n",
            "[1]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.813654\n",
            "[2]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.819006\n",
            "[3]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.821242\n",
            "[4]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822229\n",
            "[5]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822625\n",
            "[6]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.824054\n",
            "[7]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.824851\n",
            "[8]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.824043\n",
            "[9]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823426\n",
            "[10]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823687\n",
            "[11]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822861\n",
            "[12]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822598\n",
            "[13]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822548\n",
            "[14]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822354\n",
            "[15]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822713\n",
            "[16]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823039\n",
            "[17]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.821369\n",
            "[18]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823301\n",
            "[19]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823705\n",
            "[20]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823539\n",
            "[21]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823454\n",
            "[22]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82295\n",
            "[23]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.822838\n",
            "[24]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823105\n",
            "[25]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8236\n",
            "[26]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823449\n",
            "[27]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82358\n",
            "[28]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823801\n",
            "[29]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823831\n",
            "[30]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.823724\n",
            "[31]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82427\n",
            "[32]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.824563\n",
            "[33]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.824939\n",
            "[34]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825219\n",
            "[35]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825621\n",
            "[36]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825246\n",
            "[37]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825621\n",
            "[38]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825157\n",
            "[39]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825031\n",
            "[40]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825521\n",
            "[41]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825953\n",
            "[42]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825996\n",
            "[43]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825881\n",
            "[44]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826124\n",
            "[45]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826139\n",
            "[46]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825987\n",
            "[47]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826531\n",
            "[48]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8262\n",
            "[49]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82587\n",
            "[50]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826027\n",
            "[51]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826057\n",
            "[52]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826906\n",
            "[53]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826732\n",
            "[54]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826801\n",
            "[55]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826797\n",
            "[56]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826427\n",
            "[57]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826739\n",
            "[58]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826648\n",
            "[59]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826399\n",
            "[60]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826192\n",
            "[61]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826105\n",
            "[62]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826103\n",
            "[63]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825937\n",
            "[64]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825508\n",
            "[65]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825978\n",
            "[66]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826012\n",
            "[67]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826098\n",
            "[68]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826247\n",
            "[69]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826161\n",
            "[70]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826434\n",
            "[71]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826579\n",
            "[72]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826419\n",
            "[73]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826387\n",
            "[74]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826397\n",
            "[75]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826222\n",
            "[76]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826286\n",
            "[77]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826341\n",
            "[78]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826138\n",
            "[79]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82607\n",
            "[80]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826454\n",
            "[81]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826645\n",
            "[82]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826524\n",
            "[83]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826551\n",
            "[84]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826386\n",
            "[85]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826371\n",
            "[86]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826357\n",
            "[87]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826256\n",
            "[88]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826343\n",
            "[89]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826183\n",
            "[90]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826017\n",
            "[91]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826223\n",
            "[92]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825981\n",
            "[93]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825732\n",
            "[94]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825722\n",
            "[95]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825682\n",
            "[96]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825935\n",
            "[97]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825937\n",
            "[98]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82585\n",
            "[99]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.825922\n",
            "[100]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82634\n",
            "[101]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826361\n",
            "[102]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826611\n",
            "[103]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826597\n",
            "[104]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826448\n",
            "[105]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826814\n",
            "[106]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826681\n",
            "[107]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827037\n",
            "[108]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827051\n",
            "[109]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827186\n",
            "[110]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827099\n",
            "[111]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826931\n",
            "[112]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826923\n",
            "[113]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82677\n",
            "[114]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.826701\n",
            "[115]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827154\n",
            "[116]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827191\n",
            "[117]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827239\n",
            "[118]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827383\n",
            "[119]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827416\n",
            "[120]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827514\n",
            "[121]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827767\n",
            "[122]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827885\n",
            "[123]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827696\n",
            "[124]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827606\n",
            "[125]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827882\n",
            "[126]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828132\n",
            "[127]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828227\n",
            "[128]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828048\n",
            "[129]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828208\n",
            "[130]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828047\n",
            "[131]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827836\n",
            "[132]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827838\n",
            "[133]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.82802\n",
            "[134]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827828\n",
            "[135]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828037\n",
            "[136]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827979\n",
            "[137]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828102\n",
            "[138]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828047\n",
            "[139]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828003\n",
            "[140]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828136\n",
            "[141]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827986\n",
            "[142]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827901\n",
            "[143]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827731\n",
            "[144]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.827804\n",
            "[145]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828116\n",
            "[146]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828517\n",
            "[147]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828724\n",
            "[148]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828764\n",
            "[149]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828703\n",
            "[150]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.828852\n",
            "[151]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829165\n",
            "[152]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829077\n",
            "[153]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829196\n",
            "[154]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829359\n",
            "[155]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829145\n",
            "[156]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829309\n",
            "[157]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829334\n",
            "[158]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829546\n",
            "[159]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829663\n",
            "[160]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829829\n",
            "[161]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829857\n",
            "[162]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829743\n",
            "[163]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829859\n",
            "[164]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829686\n",
            "[165]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829605\n",
            "[166]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829997\n",
            "[167]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829913\n",
            "[168]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.829991\n",
            "[169]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83032\n",
            "[170]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830213\n",
            "[171]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830402\n",
            "[172]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830485\n",
            "[173]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83037\n",
            "[174]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830637\n",
            "[175]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830856\n",
            "[176]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830917\n",
            "[177]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830935\n",
            "[178]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831145\n",
            "[179]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830979\n",
            "[180]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830825\n",
            "[181]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830972\n",
            "[182]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831009\n",
            "[183]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830886\n",
            "[184]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830849\n",
            "[185]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831006\n",
            "[186]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831238\n",
            "[187]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831198\n",
            "[188]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831125\n",
            "[189]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83099\n",
            "[190]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.830854\n",
            "[191]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831099\n",
            "[192]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83097\n",
            "[193]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831197\n",
            "[194]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831065\n",
            "[195]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831067\n",
            "[196]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831259\n",
            "[197]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831222\n",
            "[198]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831137\n",
            "[199]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831154\n",
            "[200]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831154\n",
            "[201]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831298\n",
            "[202]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831189\n",
            "[203]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831524\n",
            "[204]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831473\n",
            "[205]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831499\n",
            "[206]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831465\n",
            "[207]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831771\n",
            "[208]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.831892\n",
            "[209]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832026\n",
            "[210]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83223\n",
            "[211]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832148\n",
            "[212]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832394\n",
            "[213]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832364\n",
            "[214]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832447\n",
            "[215]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832382\n",
            "[216]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83256\n",
            "[217]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832646\n",
            "[218]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832441\n",
            "[219]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832425\n",
            "[220]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832353\n",
            "[221]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832569\n",
            "[222]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832663\n",
            "[223]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832598\n",
            "[224]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832651\n",
            "[225]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832746\n",
            "[226]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832823\n",
            "[227]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832929\n",
            "[228]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832994\n",
            "[229]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832995\n",
            "[230]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.832948\n",
            "[231]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833013\n",
            "[232]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833162\n",
            "[233]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833113\n",
            "[234]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833205\n",
            "[235]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833263\n",
            "[236]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833307\n",
            "[237]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833406\n",
            "[238]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833389\n",
            "[239]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833588\n",
            "[240]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833539\n",
            "[241]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833435\n",
            "[242]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833366\n",
            "[243]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833556\n",
            "[244]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833769\n",
            "[245]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83368\n",
            "[246]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833858\n",
            "[247]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833843\n",
            "[248]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833756\n",
            "[249]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833883\n",
            "[250]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833943\n",
            "[251]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833995\n",
            "[252]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.833996\n",
            "[253]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834091\n",
            "[254]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834194\n",
            "[255]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83415\n",
            "[256]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834196\n",
            "[257]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834209\n",
            "[258]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834249\n",
            "[259]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834278\n",
            "[260]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834174\n",
            "[261]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834303\n",
            "[262]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834185\n",
            "[263]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834384\n",
            "[264]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834393\n",
            "[265]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834346\n",
            "[266]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834499\n",
            "[267]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834471\n",
            "[268]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834421\n",
            "[269]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834347\n",
            "[270]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83439\n",
            "[271]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834445\n",
            "[272]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834465\n",
            "[273]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83445\n",
            "[274]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834624\n",
            "[275]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834734\n",
            "[276]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83488\n",
            "[277]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.834992\n",
            "[278]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835068\n",
            "[279]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835186\n",
            "[280]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835295\n",
            "[281]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835283\n",
            "[282]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835396\n",
            "[283]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835422\n",
            "[284]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835297\n",
            "[285]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835363\n",
            "[286]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835484\n",
            "[287]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835465\n",
            "[288]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835496\n",
            "[289]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835457\n",
            "[290]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835555\n",
            "[291]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835613\n",
            "[292]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835688\n",
            "[293]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835829\n",
            "[294]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835934\n",
            "[295]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.835999\n",
            "[296]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836027\n",
            "[297]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836008\n",
            "[298]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836053\n",
            "[299]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83614\n",
            "[300]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836206\n",
            "[301]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836341\n",
            "[302]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836498\n",
            "[303]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836388\n",
            "[304]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836461\n",
            "[305]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836545\n",
            "[306]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836765\n",
            "[307]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836892\n",
            "[308]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836926\n",
            "[309]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.836958\n",
            "[310]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837016\n",
            "[311]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837196\n",
            "[312]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837249\n",
            "[313]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83724\n",
            "[314]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837248\n",
            "[315]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83734\n",
            "[316]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837477\n",
            "[317]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83758\n",
            "[318]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837474\n",
            "[319]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837553\n",
            "[320]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837576\n",
            "[321]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837628\n",
            "[322]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837586\n",
            "[323]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837568\n",
            "[324]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837491\n",
            "[325]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837692\n",
            "[326]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837702\n",
            "[327]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837817\n",
            "[328]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838026\n",
            "[329]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.837989\n",
            "[330]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838101\n",
            "[331]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838244\n",
            "[332]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838268\n",
            "[333]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838301\n",
            "[334]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838215\n",
            "[335]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83825\n",
            "[336]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838323\n",
            "[337]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838478\n",
            "[338]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838479\n",
            "[339]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83852\n",
            "[340]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838512\n",
            "[341]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838556\n",
            "[342]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838609\n",
            "[343]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838703\n",
            "[344]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838685\n",
            "[345]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838747\n",
            "[346]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838807\n",
            "[347]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838879\n",
            "[348]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838918\n",
            "[349]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.838995\n",
            "[350]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839026\n",
            "[351]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839086\n",
            "[352]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83915\n",
            "[353]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839279\n",
            "[354]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839311\n",
            "[355]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839421\n",
            "[356]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839398\n",
            "[357]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839386\n",
            "[358]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839496\n",
            "[359]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839531\n",
            "[360]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839552\n",
            "[361]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839567\n",
            "[362]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839451\n",
            "[363]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839456\n",
            "[364]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839453\n",
            "[365]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839564\n",
            "[366]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839633\n",
            "[367]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839685\n",
            "[368]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839743\n",
            "[369]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839773\n",
            "[370]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839799\n",
            "[371]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.83986\n",
            "[372]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839911\n",
            "[373]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839838\n",
            "[374]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839866\n",
            "[375]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839857\n",
            "[376]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839937\n",
            "[377]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.839989\n",
            "[378]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840009\n",
            "[379]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840036\n",
            "[380]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8401\n",
            "[381]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840179\n",
            "[382]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840512\n",
            "[383]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840499\n",
            "[384]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840493\n",
            "[385]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840513\n",
            "[386]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840489\n",
            "[387]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840461\n",
            "[388]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840535\n",
            "[389]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84062\n",
            "[390]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840643\n",
            "[391]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840656\n",
            "[392]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840705\n",
            "[393]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840782\n",
            "[394]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840844\n",
            "[395]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840837\n",
            "[396]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840842\n",
            "[397]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840929\n",
            "[398]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840994\n",
            "[399]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840945\n",
            "[400]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840989\n",
            "[401]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841006\n",
            "[402]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841021\n",
            "[403]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840886\n",
            "[404]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840903\n",
            "[405]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840934\n",
            "[406]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840946\n",
            "[407]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.840913\n",
            "[408]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841077\n",
            "[409]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84111\n",
            "[410]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841165\n",
            "[411]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841213\n",
            "[412]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841266\n",
            "[413]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841267\n",
            "[414]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841233\n",
            "[415]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841289\n",
            "[416]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841305\n",
            "[417]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841433\n",
            "[418]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841375\n",
            "[419]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841421\n",
            "[420]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841458\n",
            "[421]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841567\n",
            "[422]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841596\n",
            "[423]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841575\n",
            "[424]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841577\n",
            "[425]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841535\n",
            "[426]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841533\n",
            "[427]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841477\n",
            "[428]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841544\n",
            "[429]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841582\n",
            "[430]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841623\n",
            "[431]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84164\n",
            "[432]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84164\n",
            "[433]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841657\n",
            "[434]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841653\n",
            "[435]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841668\n",
            "[436]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841572\n",
            "[437]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841647\n",
            "[438]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841734\n",
            "[439]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841713\n",
            "[440]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841762\n",
            "[441]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841835\n",
            "[442]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841887\n",
            "[443]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841901\n",
            "[444]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841888\n",
            "[445]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841837\n",
            "[446]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841798\n",
            "[447]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841866\n",
            "[448]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841848\n",
            "[449]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841844\n",
            "[450]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841857\n",
            "[451]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841845\n",
            "[452]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841861\n",
            "[453]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841942\n",
            "[454]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841966\n",
            "[455]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841972\n",
            "[456]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841973\n",
            "[457]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841933\n",
            "[458]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841961\n",
            "[459]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841904\n",
            "[460]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841888\n",
            "[461]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841904\n",
            "[462]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841916\n",
            "[463]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841946\n",
            "[464]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.841993\n",
            "[465]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842071\n",
            "[466]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842093\n",
            "[467]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842124\n",
            "[468]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842221\n",
            "[469]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842265\n",
            "[470]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842301\n",
            "[471]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842342\n",
            "[472]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8424\n",
            "[473]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84245\n",
            "[474]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842423\n",
            "[475]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84248\n",
            "[476]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842463\n",
            "[477]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842528\n",
            "[478]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842523\n",
            "[479]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842527\n",
            "[480]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842612\n",
            "[481]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842627\n",
            "[482]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842607\n",
            "[483]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842644\n",
            "[484]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842651\n",
            "[485]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842668\n",
            "[486]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842717\n",
            "[487]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842787\n",
            "[488]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842759\n",
            "[489]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842775\n",
            "[490]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842737\n",
            "[491]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842733\n",
            "[492]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842729\n",
            "[493]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842801\n",
            "[494]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842852\n",
            "[495]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842902\n",
            "[496]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842889\n",
            "[497]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842858\n",
            "[498]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842917\n",
            "[499]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842922\n",
            "[500]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84287\n",
            "[501]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842874\n",
            "[502]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842855\n",
            "[503]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842947\n",
            "[504]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842943\n",
            "[505]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.842974\n",
            "[506]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84309\n",
            "[507]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843117\n",
            "[508]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843145\n",
            "[509]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8432\n",
            "[510]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843223\n",
            "[511]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843279\n",
            "[512]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843281\n",
            "[513]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843357\n",
            "[514]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843338\n",
            "[515]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843344\n",
            "[516]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843296\n",
            "[517]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843309\n",
            "[518]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843325\n",
            "[519]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843337\n",
            "[520]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843304\n",
            "[521]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843339\n",
            "[522]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843377\n",
            "[523]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843339\n",
            "[524]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843318\n",
            "[525]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843357\n",
            "[526]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843378\n",
            "[527]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843375\n",
            "[528]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843456\n",
            "[529]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843512\n",
            "[530]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843509\n",
            "[531]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843491\n",
            "[532]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843521\n",
            "[533]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84353\n",
            "[534]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843534\n",
            "[535]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843555\n",
            "[536]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843573\n",
            "[537]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843535\n",
            "[538]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843606\n",
            "[539]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843601\n",
            "[540]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84363\n",
            "[541]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843654\n",
            "[542]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843662\n",
            "[543]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84368\n",
            "[544]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843669\n",
            "[545]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843661\n",
            "[546]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843734\n",
            "[547]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843783\n",
            "[548]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843823\n",
            "[549]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843833\n",
            "[550]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843817\n",
            "[551]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843777\n",
            "[552]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843766\n",
            "[553]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843786\n",
            "[554]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8438\n",
            "[555]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843794\n",
            "[556]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843799\n",
            "[557]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843787\n",
            "[558]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84377\n",
            "[559]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843802\n",
            "[560]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843777\n",
            "[561]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843813\n",
            "[562]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843837\n",
            "[563]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84381\n",
            "[564]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843799\n",
            "[565]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843844\n",
            "[566]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843863\n",
            "[567]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843891\n",
            "[568]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843891\n",
            "[569]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843939\n",
            "[570]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843937\n",
            "[571]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843901\n",
            "[572]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843875\n",
            "[573]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843835\n",
            "[574]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843886\n",
            "[575]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843892\n",
            "[576]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843893\n",
            "[577]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843951\n",
            "[578]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843939\n",
            "[579]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84392\n",
            "[580]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843914\n",
            "[581]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843916\n",
            "[582]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843979\n",
            "[583]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84399\n",
            "[584]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843999\n",
            "[585]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.843985\n",
            "[586]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844013\n",
            "[587]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844109\n",
            "[588]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844155\n",
            "[589]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844155\n",
            "[590]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844187\n",
            "[591]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844177\n",
            "[592]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844198\n",
            "[593]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844171\n",
            "[594]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84422\n",
            "[595]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844234\n",
            "[596]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844216\n",
            "[597]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844237\n",
            "[598]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84435\n",
            "[599]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844388\n",
            "[600]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844344\n",
            "[601]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844388\n",
            "[602]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84439\n",
            "[603]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844415\n",
            "[604]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844437\n",
            "[605]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844399\n",
            "[606]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844363\n",
            "[607]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844348\n",
            "[608]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844408\n",
            "[609]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844404\n",
            "[610]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844444\n",
            "[611]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844438\n",
            "[612]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844423\n",
            "[613]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844404\n",
            "[614]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844442\n",
            "[615]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844456\n",
            "[616]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844512\n",
            "[617]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844535\n",
            "[618]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844564\n",
            "[619]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844611\n",
            "[620]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844632\n",
            "[621]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844597\n",
            "[622]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844598\n",
            "[623]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844603\n",
            "[624]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844607\n",
            "[625]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844575\n",
            "[626]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844536\n",
            "[627]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844555\n",
            "[628]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844577\n",
            "[629]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84457\n",
            "[630]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844633\n",
            "[631]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844633\n",
            "[632]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844613\n",
            "[633]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844644\n",
            "[634]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844632\n",
            "[635]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844654\n",
            "[636]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844655\n",
            "[637]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844667\n",
            "[638]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844665\n",
            "[639]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844664\n",
            "[640]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844684\n",
            "[641]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844646\n",
            "[642]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844652\n",
            "[643]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84467\n",
            "[644]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844691\n",
            "[645]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844738\n",
            "[646]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84476\n",
            "[647]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844738\n",
            "[648]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844727\n",
            "[649]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844752\n",
            "[650]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844748\n",
            "[651]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844801\n",
            "[652]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84477\n",
            "[653]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844801\n",
            "[654]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844796\n",
            "[655]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844811\n",
            "[656]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844801\n",
            "[657]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844776\n",
            "[658]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844756\n",
            "[659]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844783\n",
            "[660]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844785\n",
            "[661]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844845\n",
            "[662]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844849\n",
            "[663]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844897\n",
            "[664]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844906\n",
            "[665]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844934\n",
            "[666]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844948\n",
            "[667]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844954\n",
            "[668]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84499\n",
            "[669]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84497\n",
            "[670]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844965\n",
            "[671]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844949\n",
            "[672]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844939\n",
            "[673]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844921\n",
            "[674]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84492\n",
            "[675]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844927\n",
            "[676]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844909\n",
            "[677]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844919\n",
            "[678]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844936\n",
            "[679]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844945\n",
            "[680]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844929\n",
            "[681]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844922\n",
            "[682]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8449\n",
            "[683]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844945\n",
            "[684]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844949\n",
            "[685]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844924\n",
            "[686]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844929\n",
            "[687]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844921\n",
            "[688]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84492\n",
            "[689]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8449\n",
            "[690]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844905\n",
            "[691]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844935\n",
            "[692]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844934\n",
            "[693]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844958\n",
            "[694]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844965\n",
            "[695]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844999\n",
            "[696]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844984\n",
            "[697]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844993\n",
            "[698]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845004\n",
            "[699]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845003\n",
            "[700]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845004\n",
            "[701]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.844997\n",
            "[702]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845046\n",
            "[703]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845096\n",
            "[704]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845126\n",
            "[705]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845141\n",
            "[706]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845143\n",
            "[707]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845194\n",
            "[708]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845179\n",
            "[709]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845228\n",
            "[710]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845241\n",
            "[711]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845233\n",
            "[712]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845216\n",
            "[713]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845223\n",
            "[714]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845213\n",
            "[715]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845217\n",
            "[716]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845211\n",
            "[717]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845212\n",
            "[718]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845219\n",
            "[719]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845275\n",
            "[720]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845279\n",
            "[721]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845307\n",
            "[722]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845343\n",
            "[723]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845345\n",
            "[724]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845348\n",
            "[725]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845357\n",
            "[726]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845396\n",
            "[727]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845373\n",
            "[728]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845374\n",
            "[729]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845368\n",
            "[730]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845385\n",
            "[731]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845396\n",
            "[732]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845401\n",
            "[733]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845393\n",
            "[734]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845358\n",
            "[735]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845418\n",
            "[736]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845395\n",
            "[737]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84538\n",
            "[738]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845403\n",
            "[739]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84541\n",
            "[740]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845411\n",
            "[741]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845407\n",
            "[742]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845383\n",
            "[743]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845367\n",
            "[744]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845365\n",
            "[745]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845362\n",
            "[746]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845372\n",
            "[747]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84536\n",
            "[748]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845371\n",
            "[749]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84537\n",
            "[750]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845397\n",
            "[751]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845419\n",
            "[752]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845429\n",
            "[753]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845434\n",
            "[754]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845416\n",
            "[755]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845388\n",
            "[756]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845384\n",
            "[757]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845385\n",
            "[758]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845375\n",
            "[759]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845417\n",
            "[760]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845419\n",
            "[761]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845437\n",
            "[762]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845456\n",
            "[763]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84544\n",
            "[764]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845385\n",
            "[765]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845379\n",
            "[766]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845336\n",
            "[767]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845342\n",
            "[768]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845334\n",
            "[769]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845318\n",
            "[770]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845312\n",
            "[771]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845306\n",
            "[772]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845281\n",
            "[773]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845312\n",
            "[774]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845298\n",
            "[775]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8453\n",
            "[776]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845306\n",
            "[777]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845322\n",
            "[778]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845348\n",
            "[779]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84538\n",
            "[780]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845416\n",
            "[781]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845409\n",
            "[782]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845439\n",
            "[783]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845433\n",
            "[784]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845465\n",
            "[785]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84544\n",
            "[786]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84543\n",
            "[787]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84543\n",
            "[788]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845424\n",
            "[789]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845409\n",
            "[790]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845403\n",
            "[791]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845418\n",
            "[792]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845415\n",
            "[793]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845431\n",
            "[794]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84544\n",
            "[795]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845432\n",
            "[796]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845435\n",
            "[797]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845437\n",
            "[798]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845439\n",
            "[799]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845433\n",
            "[800]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845452\n",
            "[801]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845465\n",
            "[802]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845499\n",
            "[803]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845506\n",
            "[804]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845506\n",
            "[805]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845501\n",
            "[806]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845536\n",
            "[807]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845564\n",
            "[808]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845571\n",
            "[809]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845583\n",
            "[810]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845561\n",
            "[811]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84556\n",
            "[812]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845553\n",
            "[813]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845614\n",
            "[814]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845622\n",
            "[815]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845605\n",
            "[816]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845605\n",
            "[817]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845603\n",
            "[818]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845589\n",
            "[819]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845583\n",
            "[820]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84557\n",
            "[821]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845583\n",
            "[822]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845595\n",
            "[823]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845601\n",
            "[824]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845605\n",
            "[825]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845631\n",
            "[826]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845624\n",
            "[827]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845617\n",
            "[828]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845622\n",
            "[829]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845618\n",
            "[830]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845647\n",
            "[831]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845674\n",
            "[832]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845675\n",
            "[833]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845689\n",
            "[834]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845665\n",
            "[835]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845649\n",
            "[836]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845641\n",
            "[837]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845646\n",
            "[838]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845659\n",
            "[839]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845657\n",
            "[840]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845663\n",
            "[841]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845658\n",
            "[842]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84563\n",
            "[843]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845657\n",
            "[844]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845657\n",
            "[845]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845667\n",
            "[846]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845672\n",
            "[847]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845671\n",
            "[848]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845685\n",
            "[849]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845679\n",
            "[850]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845662\n",
            "[851]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845655\n",
            "[852]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845663\n",
            "[853]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845649\n",
            "[854]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845636\n",
            "[855]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845678\n",
            "[856]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845675\n",
            "[857]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845681\n",
            "[858]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845666\n",
            "[859]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845666\n",
            "[860]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845685\n",
            "[861]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845673\n",
            "[862]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845674\n",
            "[863]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845684\n",
            "[864]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845684\n",
            "[865]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845673\n",
            "[866]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845676\n",
            "[867]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845708\n",
            "[868]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84571\n",
            "[869]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845685\n",
            "[870]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845685\n",
            "[871]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845698\n",
            "[872]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84568\n",
            "[873]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845666\n",
            "[874]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845693\n",
            "[875]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845694\n",
            "[876]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845694\n",
            "[877]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845678\n",
            "[878]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845666\n",
            "[879]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845698\n",
            "[880]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845711\n",
            "[881]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845681\n",
            "[882]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845658\n",
            "[883]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845652\n",
            "[884]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845646\n",
            "[885]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845645\n",
            "[886]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845616\n",
            "[887]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845594\n",
            "[888]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845581\n",
            "[889]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84556\n",
            "[890]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845559\n",
            "[891]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845593\n",
            "[892]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845596\n",
            "[893]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845613\n",
            "[894]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845618\n",
            "[895]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845628\n",
            "[896]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845637\n",
            "[897]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845641\n",
            "[898]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845652\n",
            "[899]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845647\n",
            "[900]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845635\n",
            "[901]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84564\n",
            "[902]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84564\n",
            "[903]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845614\n",
            "[904]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845612\n",
            "[905]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845596\n",
            "[906]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845596\n",
            "[907]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845601\n",
            "[908]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845602\n",
            "[909]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8456\n",
            "[910]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845616\n",
            "[911]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845618\n",
            "[912]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845607\n",
            "[913]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845603\n",
            "[914]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845602\n",
            "[915]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845595\n",
            "[916]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8456\n",
            "[917]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84559\n",
            "[918]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845578\n",
            "[919]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845578\n",
            "[920]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845578\n",
            "[921]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845563\n",
            "[922]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845552\n",
            "[923]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84557\n",
            "[924]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845575\n",
            "[925]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845575\n",
            "[926]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845568\n",
            "[927]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845559\n",
            "[928]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845576\n",
            "[929]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845576\n",
            "[930]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845584\n",
            "[931]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845623\n",
            "[932]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84562\n",
            "[933]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845628\n",
            "[934]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845624\n",
            "[935]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845626\n",
            "[936]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845664\n",
            "[937]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845678\n",
            "[938]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845687\n",
            "[939]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845683\n",
            "[940]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845681\n",
            "[941]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845702\n",
            "[942]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845685\n",
            "[943]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845699\n",
            "[944]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845696\n",
            "[945]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845701\n",
            "[946]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845701\n",
            "[947]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845722\n",
            "[948]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845686\n",
            "[949]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845686\n",
            "[950]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84568\n",
            "[951]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845709\n",
            "[952]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845695\n",
            "[953]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845691\n",
            "[954]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84571\n",
            "[955]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845709\n",
            "[956]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845662\n",
            "[957]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845656\n",
            "[958]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845639\n",
            "[959]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845649\n",
            "[960]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845649\n",
            "[961]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845624\n",
            "[962]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845622\n",
            "[963]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845609\n",
            "[964]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845597\n",
            "[965]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845588\n",
            "[966]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845583\n",
            "[967]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845583\n",
            "[968]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845588\n",
            "[969]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845566\n",
            "[970]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845553\n",
            "[971]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845543\n",
            "[972]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845533\n",
            "[973]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845516\n",
            "[974]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845499\n",
            "[975]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845499\n",
            "[976]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845492\n",
            "[977]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845492\n",
            "[978]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845508\n",
            "[979]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845494\n",
            "[980]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.8455\n",
            "[981]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845498\n",
            "[982]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845499\n",
            "[983]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.84549\n",
            "[984]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845478\n",
            "[985]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845478\n",
            "[986]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845488\n",
            "[987]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845493\n",
            "[988]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845486\n",
            "[989]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845462\n",
            "[990]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845504\n",
            "[991]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845504\n",
            "[992]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845504\n",
            "[993]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845508\n",
            "[994]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845502\n",
            "[995]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845548\n",
            "[996]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845558\n",
            "[997]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845557\n",
            "[998]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845543\n",
            "[999]\tvalidation_0-error:0.039551\tvalidation_0-auc:0.845507\n",
            "CPU times: user 12min 52s, sys: 1.31 s, total: 12min 53s\n",
            "Wall time: 6min 37s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
              "              colsample_bynode=1, colsample_bytree=0.5, gamma=5,\n",
              "              learning_rate=0.01, max_delta_step=0, max_depth=5,\n",
              "              min_child_weight=1, missing=None, n_estimators=1000, n_jobs=-1,\n",
              "              nthread=-1, objective='binary:logistic', random_state=42,\n",
              "              reg_alpha=0.3, reg_lambda=1, scale_pos_weight=1.0, seed=None,\n",
              "              silent=None, subsample=0.9, verbosity=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 82
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "EcGsADLLEmmt",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "f3a3ccf4-3e87-4c04-c844-fce9a31bd146"
      },
      "source": [
        "y_pred = model.predict_proba(X_val)[:,1]\n",
        "label=\"XGBoost Classifer\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "i+=1\n",
        "result = plot_auc(y_val,y_pred,label)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t XGBoost Classifer on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.133 and 0.846 for XGBoost Classifer on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "kn5XWIAPEmmx",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        },
        "outputId": "9d168357-56c6-4987-87ae-b5b26e542eee"
      },
      "source": [
        "result_df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137490</td>\n",
              "      <td>0.832512</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>XGBoost Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.133102</td>\n",
              "      <td>0.845507</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                      Model Dataset  log_loss       auc\n",
              "0        Logisic Regression  Normal  0.542129  0.812101\n",
              "1  Decision Tree Classifier  Normal  0.148082  0.779135\n",
              "2  Random Forest Classifier  Normal  0.137490  0.832512\n",
              "3         XGBoost Classifer  Normal  0.133102  0.845507"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 84
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "LeV2q75ULzvm"
      },
      "source": [
        "# LightGBM Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "32dzyVRQLzvn",
        "colab": {}
      },
      "source": [
        "model = lgb.LGBMClassifier(n_jobs=-1,\n",
        "                      nthread=-1,\n",
        "                      scale_pos_weight=1.,\n",
        "                      learning_rate=0.01,  \n",
        "                      colsample_bytree = 0.5,\n",
        "                      subsample = 0.8,\n",
        "                      objective='binary', \n",
        "                      n_estimators=1000,\n",
        "                      reg_alpha = 0.3,\n",
        "                      max_depth=7, \n",
        "                      random_state=42\n",
        "                     )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "yJS6sV0qLzvr",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "fa3a9f88-a519-4194-8696-f9a3912ccff0"
      },
      "source": [
        "#https://towardsdatascience.com/fine-tuning-xgboost-in-python-like-a-boss-b4543ed8b1e\n",
        "\"\"\"\n",
        "manually changed each parameters 1 by 1 and compared the validation auc\n",
        "and selected the best parameter value\n",
        "\"\"\"\n",
        "eval_metric = ['auc','error']\n",
        "%time model.fit(X_train,y_train,eval_set=[(X_val,y_val)],eval_metric=eval_metric,verbose=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1]\tvalid_0's auc: 0.822169\tvalid_0's binary_logloss: 0.165704\n",
            "[2]\tvalid_0's auc: 0.820962\tvalid_0's binary_logloss: 0.16502\n",
            "[3]\tvalid_0's auc: 0.818058\tvalid_0's binary_logloss: 0.164387\n",
            "[4]\tvalid_0's auc: 0.825008\tvalid_0's binary_logloss: 0.163781\n",
            "[5]\tvalid_0's auc: 0.828321\tvalid_0's binary_logloss: 0.163151\n",
            "[6]\tvalid_0's auc: 0.824818\tvalid_0's binary_logloss: 0.162605\n",
            "[7]\tvalid_0's auc: 0.823404\tvalid_0's binary_logloss: 0.162124\n",
            "[8]\tvalid_0's auc: 0.828613\tvalid_0's binary_logloss: 0.161517\n",
            "[9]\tvalid_0's auc: 0.831049\tvalid_0's binary_logloss: 0.160938\n",
            "[10]\tvalid_0's auc: 0.830558\tvalid_0's binary_logloss: 0.160472\n",
            "[11]\tvalid_0's auc: 0.828204\tvalid_0's binary_logloss: 0.159971\n",
            "[12]\tvalid_0's auc: 0.826689\tvalid_0's binary_logloss: 0.159496\n",
            "[13]\tvalid_0's auc: 0.826417\tvalid_0's binary_logloss: 0.159061\n",
            "[14]\tvalid_0's auc: 0.826723\tvalid_0's binary_logloss: 0.15865\n",
            "[15]\tvalid_0's auc: 0.826422\tvalid_0's binary_logloss: 0.158216\n",
            "[16]\tvalid_0's auc: 0.825522\tvalid_0's binary_logloss: 0.157836\n",
            "[17]\tvalid_0's auc: 0.827403\tvalid_0's binary_logloss: 0.157354\n",
            "[18]\tvalid_0's auc: 0.82868\tvalid_0's binary_logloss: 0.156924\n",
            "[19]\tvalid_0's auc: 0.827641\tvalid_0's binary_logloss: 0.156595\n",
            "[20]\tvalid_0's auc: 0.827807\tvalid_0's binary_logloss: 0.156226\n",
            "[21]\tvalid_0's auc: 0.827683\tvalid_0's binary_logloss: 0.155875\n",
            "[22]\tvalid_0's auc: 0.828027\tvalid_0's binary_logloss: 0.155538\n",
            "[23]\tvalid_0's auc: 0.829065\tvalid_0's binary_logloss: 0.155173\n",
            "[24]\tvalid_0's auc: 0.830076\tvalid_0's binary_logloss: 0.154777\n",
            "[25]\tvalid_0's auc: 0.82991\tvalid_0's binary_logloss: 0.154476\n",
            "[26]\tvalid_0's auc: 0.829647\tvalid_0's binary_logloss: 0.154182\n",
            "[27]\tvalid_0's auc: 0.829948\tvalid_0's binary_logloss: 0.153868\n",
            "[28]\tvalid_0's auc: 0.830514\tvalid_0's binary_logloss: 0.153531\n",
            "[29]\tvalid_0's auc: 0.830957\tvalid_0's binary_logloss: 0.153164\n",
            "[30]\tvalid_0's auc: 0.831333\tvalid_0's binary_logloss: 0.152828\n",
            "[31]\tvalid_0's auc: 0.831108\tvalid_0's binary_logloss: 0.152548\n",
            "[32]\tvalid_0's auc: 0.830801\tvalid_0's binary_logloss: 0.152312\n",
            "[33]\tvalid_0's auc: 0.830668\tvalid_0's binary_logloss: 0.152049\n",
            "[34]\tvalid_0's auc: 0.831259\tvalid_0's binary_logloss: 0.151724\n",
            "[35]\tvalid_0's auc: 0.831924\tvalid_0's binary_logloss: 0.151406\n",
            "[36]\tvalid_0's auc: 0.83211\tvalid_0's binary_logloss: 0.151099\n",
            "[37]\tvalid_0's auc: 0.832027\tvalid_0's binary_logloss: 0.150867\n",
            "[38]\tvalid_0's auc: 0.832718\tvalid_0's binary_logloss: 0.15056\n",
            "[39]\tvalid_0's auc: 0.833345\tvalid_0's binary_logloss: 0.150285\n",
            "[40]\tvalid_0's auc: 0.833632\tvalid_0's binary_logloss: 0.150061\n",
            "[41]\tvalid_0's auc: 0.833459\tvalid_0's binary_logloss: 0.149845\n",
            "[42]\tvalid_0's auc: 0.83388\tvalid_0's binary_logloss: 0.149577\n",
            "[43]\tvalid_0's auc: 0.834212\tvalid_0's binary_logloss: 0.149313\n",
            "[44]\tvalid_0's auc: 0.833894\tvalid_0's binary_logloss: 0.149101\n",
            "[45]\tvalid_0's auc: 0.834182\tvalid_0's binary_logloss: 0.148854\n",
            "[46]\tvalid_0's auc: 0.834439\tvalid_0's binary_logloss: 0.148605\n",
            "[47]\tvalid_0's auc: 0.834806\tvalid_0's binary_logloss: 0.148349\n",
            "[48]\tvalid_0's auc: 0.83462\tvalid_0's binary_logloss: 0.14815\n",
            "[49]\tvalid_0's auc: 0.834581\tvalid_0's binary_logloss: 0.147966\n",
            "[50]\tvalid_0's auc: 0.834973\tvalid_0's binary_logloss: 0.147731\n",
            "[51]\tvalid_0's auc: 0.834909\tvalid_0's binary_logloss: 0.147574\n",
            "[52]\tvalid_0's auc: 0.835053\tvalid_0's binary_logloss: 0.147382\n",
            "[53]\tvalid_0's auc: 0.834948\tvalid_0's binary_logloss: 0.147231\n",
            "[54]\tvalid_0's auc: 0.835011\tvalid_0's binary_logloss: 0.147045\n",
            "[55]\tvalid_0's auc: 0.835158\tvalid_0's binary_logloss: 0.146839\n",
            "[56]\tvalid_0's auc: 0.835588\tvalid_0's binary_logloss: 0.146627\n",
            "[57]\tvalid_0's auc: 0.835472\tvalid_0's binary_logloss: 0.146479\n",
            "[58]\tvalid_0's auc: 0.835723\tvalid_0's binary_logloss: 0.146283\n",
            "[59]\tvalid_0's auc: 0.835702\tvalid_0's binary_logloss: 0.146128\n",
            "[60]\tvalid_0's auc: 0.835837\tvalid_0's binary_logloss: 0.145948\n",
            "[61]\tvalid_0's auc: 0.836148\tvalid_0's binary_logloss: 0.145767\n",
            "[62]\tvalid_0's auc: 0.835897\tvalid_0's binary_logloss: 0.145633\n",
            "[63]\tvalid_0's auc: 0.836022\tvalid_0's binary_logloss: 0.145443\n",
            "[64]\tvalid_0's auc: 0.836225\tvalid_0's binary_logloss: 0.14528\n",
            "[65]\tvalid_0's auc: 0.83621\tvalid_0's binary_logloss: 0.145124\n",
            "[66]\tvalid_0's auc: 0.836386\tvalid_0's binary_logloss: 0.144951\n",
            "[67]\tvalid_0's auc: 0.836509\tvalid_0's binary_logloss: 0.144824\n",
            "[68]\tvalid_0's auc: 0.836397\tvalid_0's binary_logloss: 0.144697\n",
            "[69]\tvalid_0's auc: 0.836553\tvalid_0's binary_logloss: 0.144566\n",
            "[70]\tvalid_0's auc: 0.836558\tvalid_0's binary_logloss: 0.144422\n",
            "[71]\tvalid_0's auc: 0.83682\tvalid_0's binary_logloss: 0.144292\n",
            "[72]\tvalid_0's auc: 0.836702\tvalid_0's binary_logloss: 0.144166\n",
            "[73]\tvalid_0's auc: 0.836715\tvalid_0's binary_logloss: 0.144049\n",
            "[74]\tvalid_0's auc: 0.836906\tvalid_0's binary_logloss: 0.1439\n",
            "[75]\tvalid_0's auc: 0.836844\tvalid_0's binary_logloss: 0.143775\n",
            "[76]\tvalid_0's auc: 0.83693\tvalid_0's binary_logloss: 0.143618\n",
            "[77]\tvalid_0's auc: 0.837342\tvalid_0's binary_logloss: 0.143456\n",
            "[78]\tvalid_0's auc: 0.837252\tvalid_0's binary_logloss: 0.143346\n",
            "[79]\tvalid_0's auc: 0.83745\tvalid_0's binary_logloss: 0.143239\n",
            "[80]\tvalid_0's auc: 0.837498\tvalid_0's binary_logloss: 0.14309\n",
            "[81]\tvalid_0's auc: 0.837652\tvalid_0's binary_logloss: 0.142955\n",
            "[82]\tvalid_0's auc: 0.837716\tvalid_0's binary_logloss: 0.142819\n",
            "[83]\tvalid_0's auc: 0.837744\tvalid_0's binary_logloss: 0.142724\n",
            "[84]\tvalid_0's auc: 0.837841\tvalid_0's binary_logloss: 0.142588\n",
            "[85]\tvalid_0's auc: 0.838006\tvalid_0's binary_logloss: 0.142471\n",
            "[86]\tvalid_0's auc: 0.83822\tvalid_0's binary_logloss: 0.142331\n",
            "[87]\tvalid_0's auc: 0.838163\tvalid_0's binary_logloss: 0.142239\n",
            "[88]\tvalid_0's auc: 0.838194\tvalid_0's binary_logloss: 0.142116\n",
            "[89]\tvalid_0's auc: 0.838111\tvalid_0's binary_logloss: 0.142027\n",
            "[90]\tvalid_0's auc: 0.838471\tvalid_0's binary_logloss: 0.1419\n",
            "[91]\tvalid_0's auc: 0.838513\tvalid_0's binary_logloss: 0.141817\n",
            "[92]\tvalid_0's auc: 0.838462\tvalid_0's binary_logloss: 0.141731\n",
            "[93]\tvalid_0's auc: 0.838353\tvalid_0's binary_logloss: 0.141649\n",
            "[94]\tvalid_0's auc: 0.838506\tvalid_0's binary_logloss: 0.141522\n",
            "[95]\tvalid_0's auc: 0.838495\tvalid_0's binary_logloss: 0.141439\n",
            "[96]\tvalid_0's auc: 0.838457\tvalid_0's binary_logloss: 0.141355\n",
            "[97]\tvalid_0's auc: 0.838582\tvalid_0's binary_logloss: 0.141265\n",
            "[98]\tvalid_0's auc: 0.838548\tvalid_0's binary_logloss: 0.141184\n",
            "[99]\tvalid_0's auc: 0.838548\tvalid_0's binary_logloss: 0.141104\n",
            "[100]\tvalid_0's auc: 0.838373\tvalid_0's binary_logloss: 0.14104\n",
            "[101]\tvalid_0's auc: 0.838707\tvalid_0's binary_logloss: 0.140935\n",
            "[102]\tvalid_0's auc: 0.838949\tvalid_0's binary_logloss: 0.140821\n",
            "[103]\tvalid_0's auc: 0.838989\tvalid_0's binary_logloss: 0.140746\n",
            "[104]\tvalid_0's auc: 0.8389\tvalid_0's binary_logloss: 0.140674\n",
            "[105]\tvalid_0's auc: 0.838818\tvalid_0's binary_logloss: 0.140602\n",
            "[106]\tvalid_0's auc: 0.838803\tvalid_0's binary_logloss: 0.140531\n",
            "[107]\tvalid_0's auc: 0.838982\tvalid_0's binary_logloss: 0.140435\n",
            "[108]\tvalid_0's auc: 0.839076\tvalid_0's binary_logloss: 0.140343\n",
            "[109]\tvalid_0's auc: 0.839278\tvalid_0's binary_logloss: 0.14026\n",
            "[110]\tvalid_0's auc: 0.839225\tvalid_0's binary_logloss: 0.140194\n",
            "[111]\tvalid_0's auc: 0.839351\tvalid_0's binary_logloss: 0.140128\n",
            "[112]\tvalid_0's auc: 0.839315\tvalid_0's binary_logloss: 0.140059\n",
            "[113]\tvalid_0's auc: 0.839395\tvalid_0's binary_logloss: 0.139965\n",
            "[114]\tvalid_0's auc: 0.839429\tvalid_0's binary_logloss: 0.1399\n",
            "[115]\tvalid_0's auc: 0.839554\tvalid_0's binary_logloss: 0.139802\n",
            "[116]\tvalid_0's auc: 0.839534\tvalid_0's binary_logloss: 0.139729\n",
            "[117]\tvalid_0's auc: 0.839638\tvalid_0's binary_logloss: 0.139643\n",
            "[118]\tvalid_0's auc: 0.839642\tvalid_0's binary_logloss: 0.139588\n",
            "[119]\tvalid_0's auc: 0.839672\tvalid_0's binary_logloss: 0.139503\n",
            "[120]\tvalid_0's auc: 0.839595\tvalid_0's binary_logloss: 0.139437\n",
            "[121]\tvalid_0's auc: 0.839747\tvalid_0's binary_logloss: 0.139378\n",
            "[122]\tvalid_0's auc: 0.839762\tvalid_0's binary_logloss: 0.139301\n",
            "[123]\tvalid_0's auc: 0.839889\tvalid_0's binary_logloss: 0.139238\n",
            "[124]\tvalid_0's auc: 0.839969\tvalid_0's binary_logloss: 0.139153\n",
            "[125]\tvalid_0's auc: 0.840041\tvalid_0's binary_logloss: 0.139097\n",
            "[126]\tvalid_0's auc: 0.840124\tvalid_0's binary_logloss: 0.139047\n",
            "[127]\tvalid_0's auc: 0.840015\tvalid_0's binary_logloss: 0.139004\n",
            "[128]\tvalid_0's auc: 0.84013\tvalid_0's binary_logloss: 0.138915\n",
            "[129]\tvalid_0's auc: 0.840083\tvalid_0's binary_logloss: 0.138874\n",
            "[130]\tvalid_0's auc: 0.840213\tvalid_0's binary_logloss: 0.13881\n",
            "[131]\tvalid_0's auc: 0.840335\tvalid_0's binary_logloss: 0.138734\n",
            "[132]\tvalid_0's auc: 0.840168\tvalid_0's binary_logloss: 0.138688\n",
            "[133]\tvalid_0's auc: 0.840079\tvalid_0's binary_logloss: 0.138649\n",
            "[134]\tvalid_0's auc: 0.840142\tvalid_0's binary_logloss: 0.13858\n",
            "[135]\tvalid_0's auc: 0.840249\tvalid_0's binary_logloss: 0.138512\n",
            "[136]\tvalid_0's auc: 0.840201\tvalid_0's binary_logloss: 0.138452\n",
            "[137]\tvalid_0's auc: 0.840206\tvalid_0's binary_logloss: 0.138394\n",
            "[138]\tvalid_0's auc: 0.840282\tvalid_0's binary_logloss: 0.138334\n",
            "[139]\tvalid_0's auc: 0.840431\tvalid_0's binary_logloss: 0.138261\n",
            "[140]\tvalid_0's auc: 0.840361\tvalid_0's binary_logloss: 0.138224\n",
            "[141]\tvalid_0's auc: 0.840431\tvalid_0's binary_logloss: 0.138161\n",
            "[142]\tvalid_0's auc: 0.840388\tvalid_0's binary_logloss: 0.138124\n",
            "[143]\tvalid_0's auc: 0.840337\tvalid_0's binary_logloss: 0.138086\n",
            "[144]\tvalid_0's auc: 0.840372\tvalid_0's binary_logloss: 0.13803\n",
            "[145]\tvalid_0's auc: 0.840411\tvalid_0's binary_logloss: 0.137976\n",
            "[146]\tvalid_0's auc: 0.840419\tvalid_0's binary_logloss: 0.137936\n",
            "[147]\tvalid_0's auc: 0.840373\tvalid_0's binary_logloss: 0.137898\n",
            "[148]\tvalid_0's auc: 0.840475\tvalid_0's binary_logloss: 0.137853\n",
            "[149]\tvalid_0's auc: 0.840449\tvalid_0's binary_logloss: 0.137812\n",
            "[150]\tvalid_0's auc: 0.840639\tvalid_0's binary_logloss: 0.137747\n",
            "[151]\tvalid_0's auc: 0.840711\tvalid_0's binary_logloss: 0.137686\n",
            "[152]\tvalid_0's auc: 0.840725\tvalid_0's binary_logloss: 0.137652\n",
            "[153]\tvalid_0's auc: 0.840743\tvalid_0's binary_logloss: 0.137613\n",
            "[154]\tvalid_0's auc: 0.840741\tvalid_0's binary_logloss: 0.137576\n",
            "[155]\tvalid_0's auc: 0.840681\tvalid_0's binary_logloss: 0.137534\n",
            "[156]\tvalid_0's auc: 0.840624\tvalid_0's binary_logloss: 0.1375\n",
            "[157]\tvalid_0's auc: 0.840778\tvalid_0's binary_logloss: 0.137432\n",
            "[158]\tvalid_0's auc: 0.840722\tvalid_0's binary_logloss: 0.137391\n",
            "[159]\tvalid_0's auc: 0.840606\tvalid_0's binary_logloss: 0.137365\n",
            "[160]\tvalid_0's auc: 0.8407\tvalid_0's binary_logloss: 0.137302\n",
            "[161]\tvalid_0's auc: 0.840834\tvalid_0's binary_logloss: 0.137265\n",
            "[162]\tvalid_0's auc: 0.840966\tvalid_0's binary_logloss: 0.1372\n",
            "[163]\tvalid_0's auc: 0.840952\tvalid_0's binary_logloss: 0.137164\n",
            "[164]\tvalid_0's auc: 0.840951\tvalid_0's binary_logloss: 0.137112\n",
            "[165]\tvalid_0's auc: 0.840948\tvalid_0's binary_logloss: 0.137083\n",
            "[166]\tvalid_0's auc: 0.84103\tvalid_0's binary_logloss: 0.137026\n",
            "[167]\tvalid_0's auc: 0.841119\tvalid_0's binary_logloss: 0.136976\n",
            "[168]\tvalid_0's auc: 0.841138\tvalid_0's binary_logloss: 0.136938\n",
            "[169]\tvalid_0's auc: 0.841217\tvalid_0's binary_logloss: 0.136889\n",
            "[170]\tvalid_0's auc: 0.841282\tvalid_0's binary_logloss: 0.136837\n",
            "[171]\tvalid_0's auc: 0.841204\tvalid_0's binary_logloss: 0.136803\n",
            "[172]\tvalid_0's auc: 0.841273\tvalid_0's binary_logloss: 0.136762\n",
            "[173]\tvalid_0's auc: 0.841292\tvalid_0's binary_logloss: 0.136725\n",
            "[174]\tvalid_0's auc: 0.841299\tvalid_0's binary_logloss: 0.136691\n",
            "[175]\tvalid_0's auc: 0.841326\tvalid_0's binary_logloss: 0.136641\n",
            "[176]\tvalid_0's auc: 0.841438\tvalid_0's binary_logloss: 0.136594\n",
            "[177]\tvalid_0's auc: 0.841568\tvalid_0's binary_logloss: 0.136563\n",
            "[178]\tvalid_0's auc: 0.841504\tvalid_0's binary_logloss: 0.136533\n",
            "[179]\tvalid_0's auc: 0.841675\tvalid_0's binary_logloss: 0.136488\n",
            "[180]\tvalid_0's auc: 0.841672\tvalid_0's binary_logloss: 0.136463\n",
            "[181]\tvalid_0's auc: 0.841666\tvalid_0's binary_logloss: 0.136432\n",
            "[182]\tvalid_0's auc: 0.84169\tvalid_0's binary_logloss: 0.136395\n",
            "[183]\tvalid_0's auc: 0.841699\tvalid_0's binary_logloss: 0.136361\n",
            "[184]\tvalid_0's auc: 0.841679\tvalid_0's binary_logloss: 0.136336\n",
            "[185]\tvalid_0's auc: 0.841682\tvalid_0's binary_logloss: 0.136313\n",
            "[186]\tvalid_0's auc: 0.841741\tvalid_0's binary_logloss: 0.136289\n",
            "[187]\tvalid_0's auc: 0.841789\tvalid_0's binary_logloss: 0.136253\n",
            "[188]\tvalid_0's auc: 0.841938\tvalid_0's binary_logloss: 0.136215\n",
            "[189]\tvalid_0's auc: 0.841882\tvalid_0's binary_logloss: 0.136194\n",
            "[190]\tvalid_0's auc: 0.841993\tvalid_0's binary_logloss: 0.136151\n",
            "[191]\tvalid_0's auc: 0.841997\tvalid_0's binary_logloss: 0.136131\n",
            "[192]\tvalid_0's auc: 0.84204\tvalid_0's binary_logloss: 0.136099\n",
            "[193]\tvalid_0's auc: 0.84215\tvalid_0's binary_logloss: 0.136053\n",
            "[194]\tvalid_0's auc: 0.842027\tvalid_0's binary_logloss: 0.136032\n",
            "[195]\tvalid_0's auc: 0.84235\tvalid_0's binary_logloss: 0.135996\n",
            "[196]\tvalid_0's auc: 0.842423\tvalid_0's binary_logloss: 0.135968\n",
            "[197]\tvalid_0's auc: 0.842463\tvalid_0's binary_logloss: 0.135927\n",
            "[198]\tvalid_0's auc: 0.842512\tvalid_0's binary_logloss: 0.135904\n",
            "[199]\tvalid_0's auc: 0.842592\tvalid_0's binary_logloss: 0.135866\n",
            "[200]\tvalid_0's auc: 0.842626\tvalid_0's binary_logloss: 0.135829\n",
            "[201]\tvalid_0's auc: 0.842617\tvalid_0's binary_logloss: 0.135814\n",
            "[202]\tvalid_0's auc: 0.842642\tvalid_0's binary_logloss: 0.135791\n",
            "[203]\tvalid_0's auc: 0.842644\tvalid_0's binary_logloss: 0.135758\n",
            "[204]\tvalid_0's auc: 0.842723\tvalid_0's binary_logloss: 0.135716\n",
            "[205]\tvalid_0's auc: 0.842797\tvalid_0's binary_logloss: 0.135678\n",
            "[206]\tvalid_0's auc: 0.842829\tvalid_0's binary_logloss: 0.135658\n",
            "[207]\tvalid_0's auc: 0.842824\tvalid_0's binary_logloss: 0.135641\n",
            "[208]\tvalid_0's auc: 0.842901\tvalid_0's binary_logloss: 0.135593\n",
            "[209]\tvalid_0's auc: 0.842904\tvalid_0's binary_logloss: 0.135558\n",
            "[210]\tvalid_0's auc: 0.843002\tvalid_0's binary_logloss: 0.135537\n",
            "[211]\tvalid_0's auc: 0.843066\tvalid_0's binary_logloss: 0.1355\n",
            "[212]\tvalid_0's auc: 0.843025\tvalid_0's binary_logloss: 0.135489\n",
            "[213]\tvalid_0's auc: 0.843103\tvalid_0's binary_logloss: 0.135469\n",
            "[214]\tvalid_0's auc: 0.843154\tvalid_0's binary_logloss: 0.135433\n",
            "[215]\tvalid_0's auc: 0.843182\tvalid_0's binary_logloss: 0.135408\n",
            "[216]\tvalid_0's auc: 0.843226\tvalid_0's binary_logloss: 0.135382\n",
            "[217]\tvalid_0's auc: 0.84324\tvalid_0's binary_logloss: 0.135366\n",
            "[218]\tvalid_0's auc: 0.843288\tvalid_0's binary_logloss: 0.135333\n",
            "[219]\tvalid_0's auc: 0.843389\tvalid_0's binary_logloss: 0.135307\n",
            "[220]\tvalid_0's auc: 0.843502\tvalid_0's binary_logloss: 0.135273\n",
            "[221]\tvalid_0's auc: 0.843537\tvalid_0's binary_logloss: 0.135249\n",
            "[222]\tvalid_0's auc: 0.843518\tvalid_0's binary_logloss: 0.135232\n",
            "[223]\tvalid_0's auc: 0.843474\tvalid_0's binary_logloss: 0.135221\n",
            "[224]\tvalid_0's auc: 0.843564\tvalid_0's binary_logloss: 0.135181\n",
            "[225]\tvalid_0's auc: 0.843533\tvalid_0's binary_logloss: 0.135158\n",
            "[226]\tvalid_0's auc: 0.843634\tvalid_0's binary_logloss: 0.135136\n",
            "[227]\tvalid_0's auc: 0.843665\tvalid_0's binary_logloss: 0.135109\n",
            "[228]\tvalid_0's auc: 0.843667\tvalid_0's binary_logloss: 0.135102\n",
            "[229]\tvalid_0's auc: 0.843806\tvalid_0's binary_logloss: 0.135063\n",
            "[230]\tvalid_0's auc: 0.843834\tvalid_0's binary_logloss: 0.135049\n",
            "[231]\tvalid_0's auc: 0.843865\tvalid_0's binary_logloss: 0.135026\n",
            "[232]\tvalid_0's auc: 0.843906\tvalid_0's binary_logloss: 0.135007\n",
            "[233]\tvalid_0's auc: 0.843944\tvalid_0's binary_logloss: 0.13498\n",
            "[234]\tvalid_0's auc: 0.843922\tvalid_0's binary_logloss: 0.134974\n",
            "[235]\tvalid_0's auc: 0.843975\tvalid_0's binary_logloss: 0.134952\n",
            "[236]\tvalid_0's auc: 0.843977\tvalid_0's binary_logloss: 0.134931\n",
            "[237]\tvalid_0's auc: 0.843996\tvalid_0's binary_logloss: 0.134903\n",
            "[238]\tvalid_0's auc: 0.844014\tvalid_0's binary_logloss: 0.134891\n",
            "[239]\tvalid_0's auc: 0.844069\tvalid_0's binary_logloss: 0.134879\n",
            "[240]\tvalid_0's auc: 0.844144\tvalid_0's binary_logloss: 0.134864\n",
            "[241]\tvalid_0's auc: 0.844172\tvalid_0's binary_logloss: 0.134837\n",
            "[242]\tvalid_0's auc: 0.844192\tvalid_0's binary_logloss: 0.134811\n",
            "[243]\tvalid_0's auc: 0.844156\tvalid_0's binary_logloss: 0.134808\n",
            "[244]\tvalid_0's auc: 0.844157\tvalid_0's binary_logloss: 0.13479\n",
            "[245]\tvalid_0's auc: 0.844189\tvalid_0's binary_logloss: 0.134765\n",
            "[246]\tvalid_0's auc: 0.844157\tvalid_0's binary_logloss: 0.134754\n",
            "[247]\tvalid_0's auc: 0.844083\tvalid_0's binary_logloss: 0.134747\n",
            "[248]\tvalid_0's auc: 0.844145\tvalid_0's binary_logloss: 0.134718\n",
            "[249]\tvalid_0's auc: 0.844141\tvalid_0's binary_logloss: 0.134709\n",
            "[250]\tvalid_0's auc: 0.844183\tvalid_0's binary_logloss: 0.13468\n",
            "[251]\tvalid_0's auc: 0.844234\tvalid_0's binary_logloss: 0.134668\n",
            "[252]\tvalid_0's auc: 0.844222\tvalid_0's binary_logloss: 0.134649\n",
            "[253]\tvalid_0's auc: 0.844319\tvalid_0's binary_logloss: 0.134612\n",
            "[254]\tvalid_0's auc: 0.844328\tvalid_0's binary_logloss: 0.134605\n",
            "[255]\tvalid_0's auc: 0.844341\tvalid_0's binary_logloss: 0.134595\n",
            "[256]\tvalid_0's auc: 0.844355\tvalid_0's binary_logloss: 0.134579\n",
            "[257]\tvalid_0's auc: 0.844413\tvalid_0's binary_logloss: 0.134554\n",
            "[258]\tvalid_0's auc: 0.844429\tvalid_0's binary_logloss: 0.134546\n",
            "[259]\tvalid_0's auc: 0.844419\tvalid_0's binary_logloss: 0.134525\n",
            "[260]\tvalid_0's auc: 0.844478\tvalid_0's binary_logloss: 0.134503\n",
            "[261]\tvalid_0's auc: 0.844465\tvalid_0's binary_logloss: 0.134495\n",
            "[262]\tvalid_0's auc: 0.844435\tvalid_0's binary_logloss: 0.134487\n",
            "[263]\tvalid_0's auc: 0.844475\tvalid_0's binary_logloss: 0.134464\n",
            "[264]\tvalid_0's auc: 0.844533\tvalid_0's binary_logloss: 0.134447\n",
            "[265]\tvalid_0's auc: 0.844506\tvalid_0's binary_logloss: 0.134437\n",
            "[266]\tvalid_0's auc: 0.844555\tvalid_0's binary_logloss: 0.134421\n",
            "[267]\tvalid_0's auc: 0.844601\tvalid_0's binary_logloss: 0.134401\n",
            "[268]\tvalid_0's auc: 0.844556\tvalid_0's binary_logloss: 0.134389\n",
            "[269]\tvalid_0's auc: 0.84457\tvalid_0's binary_logloss: 0.134383\n",
            "[270]\tvalid_0's auc: 0.844605\tvalid_0's binary_logloss: 0.134365\n",
            "[271]\tvalid_0's auc: 0.84463\tvalid_0's binary_logloss: 0.134349\n",
            "[272]\tvalid_0's auc: 0.844692\tvalid_0's binary_logloss: 0.134332\n",
            "[273]\tvalid_0's auc: 0.844666\tvalid_0's binary_logloss: 0.134334\n",
            "[274]\tvalid_0's auc: 0.844719\tvalid_0's binary_logloss: 0.134319\n",
            "[275]\tvalid_0's auc: 0.844722\tvalid_0's binary_logloss: 0.134312\n",
            "[276]\tvalid_0's auc: 0.844746\tvalid_0's binary_logloss: 0.134287\n",
            "[277]\tvalid_0's auc: 0.844772\tvalid_0's binary_logloss: 0.13427\n",
            "[278]\tvalid_0's auc: 0.84478\tvalid_0's binary_logloss: 0.134252\n",
            "[279]\tvalid_0's auc: 0.844793\tvalid_0's binary_logloss: 0.134244\n",
            "[280]\tvalid_0's auc: 0.844767\tvalid_0's binary_logloss: 0.134237\n",
            "[281]\tvalid_0's auc: 0.844845\tvalid_0's binary_logloss: 0.134218\n",
            "[282]\tvalid_0's auc: 0.844854\tvalid_0's binary_logloss: 0.134202\n",
            "[283]\tvalid_0's auc: 0.844832\tvalid_0's binary_logloss: 0.134188\n",
            "[284]\tvalid_0's auc: 0.844759\tvalid_0's binary_logloss: 0.134187\n",
            "[285]\tvalid_0's auc: 0.844758\tvalid_0's binary_logloss: 0.134178\n",
            "[286]\tvalid_0's auc: 0.844799\tvalid_0's binary_logloss: 0.134164\n",
            "[287]\tvalid_0's auc: 0.844895\tvalid_0's binary_logloss: 0.134135\n",
            "[288]\tvalid_0's auc: 0.844901\tvalid_0's binary_logloss: 0.134131\n",
            "[289]\tvalid_0's auc: 0.844928\tvalid_0's binary_logloss: 0.134116\n",
            "[290]\tvalid_0's auc: 0.844962\tvalid_0's binary_logloss: 0.134104\n",
            "[291]\tvalid_0's auc: 0.844995\tvalid_0's binary_logloss: 0.134076\n",
            "[292]\tvalid_0's auc: 0.845002\tvalid_0's binary_logloss: 0.134069\n",
            "[293]\tvalid_0's auc: 0.84499\tvalid_0's binary_logloss: 0.134064\n",
            "[294]\tvalid_0's auc: 0.844949\tvalid_0's binary_logloss: 0.134067\n",
            "[295]\tvalid_0's auc: 0.844969\tvalid_0's binary_logloss: 0.13406\n",
            "[296]\tvalid_0's auc: 0.845004\tvalid_0's binary_logloss: 0.13404\n",
            "[297]\tvalid_0's auc: 0.845031\tvalid_0's binary_logloss: 0.134024\n",
            "[298]\tvalid_0's auc: 0.845055\tvalid_0's binary_logloss: 0.134011\n",
            "[299]\tvalid_0's auc: 0.845088\tvalid_0's binary_logloss: 0.133998\n",
            "[300]\tvalid_0's auc: 0.84513\tvalid_0's binary_logloss: 0.133983\n",
            "[301]\tvalid_0's auc: 0.8451\tvalid_0's binary_logloss: 0.133967\n",
            "[302]\tvalid_0's auc: 0.845123\tvalid_0's binary_logloss: 0.133961\n",
            "[303]\tvalid_0's auc: 0.845139\tvalid_0's binary_logloss: 0.133957\n",
            "[304]\tvalid_0's auc: 0.845111\tvalid_0's binary_logloss: 0.133955\n",
            "[305]\tvalid_0's auc: 0.845089\tvalid_0's binary_logloss: 0.133946\n",
            "[306]\tvalid_0's auc: 0.84504\tvalid_0's binary_logloss: 0.133938\n",
            "[307]\tvalid_0's auc: 0.845083\tvalid_0's binary_logloss: 0.133936\n",
            "[308]\tvalid_0's auc: 0.845078\tvalid_0's binary_logloss: 0.133929\n",
            "[309]\tvalid_0's auc: 0.845089\tvalid_0's binary_logloss: 0.133918\n",
            "[310]\tvalid_0's auc: 0.845073\tvalid_0's binary_logloss: 0.133909\n",
            "[311]\tvalid_0's auc: 0.845089\tvalid_0's binary_logloss: 0.133897\n",
            "[312]\tvalid_0's auc: 0.845052\tvalid_0's binary_logloss: 0.133887\n",
            "[313]\tvalid_0's auc: 0.845056\tvalid_0's binary_logloss: 0.133883\n",
            "[314]\tvalid_0's auc: 0.845077\tvalid_0's binary_logloss: 0.133873\n",
            "[315]\tvalid_0's auc: 0.845067\tvalid_0's binary_logloss: 0.133868\n",
            "[316]\tvalid_0's auc: 0.845151\tvalid_0's binary_logloss: 0.13385\n",
            "[317]\tvalid_0's auc: 0.845126\tvalid_0's binary_logloss: 0.133845\n",
            "[318]\tvalid_0's auc: 0.845077\tvalid_0's binary_logloss: 0.133847\n",
            "[319]\tvalid_0's auc: 0.845104\tvalid_0's binary_logloss: 0.13384\n",
            "[320]\tvalid_0's auc: 0.845124\tvalid_0's binary_logloss: 0.133824\n",
            "[321]\tvalid_0's auc: 0.845219\tvalid_0's binary_logloss: 0.133805\n",
            "[322]\tvalid_0's auc: 0.845251\tvalid_0's binary_logloss: 0.133798\n",
            "[323]\tvalid_0's auc: 0.845274\tvalid_0's binary_logloss: 0.133788\n",
            "[324]\tvalid_0's auc: 0.845296\tvalid_0's binary_logloss: 0.133783\n",
            "[325]\tvalid_0's auc: 0.845287\tvalid_0's binary_logloss: 0.133769\n",
            "[326]\tvalid_0's auc: 0.845309\tvalid_0's binary_logloss: 0.133764\n",
            "[327]\tvalid_0's auc: 0.845289\tvalid_0's binary_logloss: 0.133754\n",
            "[328]\tvalid_0's auc: 0.845296\tvalid_0's binary_logloss: 0.133751\n",
            "[329]\tvalid_0's auc: 0.845279\tvalid_0's binary_logloss: 0.133752\n",
            "[330]\tvalid_0's auc: 0.845273\tvalid_0's binary_logloss: 0.133744\n",
            "[331]\tvalid_0's auc: 0.845277\tvalid_0's binary_logloss: 0.133736\n",
            "[332]\tvalid_0's auc: 0.845296\tvalid_0's binary_logloss: 0.133728\n",
            "[333]\tvalid_0's auc: 0.845308\tvalid_0's binary_logloss: 0.133723\n",
            "[334]\tvalid_0's auc: 0.845298\tvalid_0's binary_logloss: 0.13372\n",
            "[335]\tvalid_0's auc: 0.845319\tvalid_0's binary_logloss: 0.133714\n",
            "[336]\tvalid_0's auc: 0.845364\tvalid_0's binary_logloss: 0.133702\n",
            "[337]\tvalid_0's auc: 0.845341\tvalid_0's binary_logloss: 0.133695\n",
            "[338]\tvalid_0's auc: 0.845344\tvalid_0's binary_logloss: 0.133689\n",
            "[339]\tvalid_0's auc: 0.84544\tvalid_0's binary_logloss: 0.133661\n",
            "[340]\tvalid_0's auc: 0.845478\tvalid_0's binary_logloss: 0.133651\n",
            "[341]\tvalid_0's auc: 0.845506\tvalid_0's binary_logloss: 0.133642\n",
            "[342]\tvalid_0's auc: 0.845508\tvalid_0's binary_logloss: 0.133638\n",
            "[343]\tvalid_0's auc: 0.845489\tvalid_0's binary_logloss: 0.133635\n",
            "[344]\tvalid_0's auc: 0.845504\tvalid_0's binary_logloss: 0.133622\n",
            "[345]\tvalid_0's auc: 0.845494\tvalid_0's binary_logloss: 0.133615\n",
            "[346]\tvalid_0's auc: 0.845514\tvalid_0's binary_logloss: 0.133604\n",
            "[347]\tvalid_0's auc: 0.845551\tvalid_0's binary_logloss: 0.133591\n",
            "[348]\tvalid_0's auc: 0.845518\tvalid_0's binary_logloss: 0.133587\n",
            "[349]\tvalid_0's auc: 0.845526\tvalid_0's binary_logloss: 0.133581\n",
            "[350]\tvalid_0's auc: 0.845513\tvalid_0's binary_logloss: 0.133575\n",
            "[351]\tvalid_0's auc: 0.845478\tvalid_0's binary_logloss: 0.133571\n",
            "[352]\tvalid_0's auc: 0.845552\tvalid_0's binary_logloss: 0.133566\n",
            "[353]\tvalid_0's auc: 0.845521\tvalid_0's binary_logloss: 0.133557\n",
            "[354]\tvalid_0's auc: 0.845552\tvalid_0's binary_logloss: 0.133547\n",
            "[355]\tvalid_0's auc: 0.845565\tvalid_0's binary_logloss: 0.133532\n",
            "[356]\tvalid_0's auc: 0.845576\tvalid_0's binary_logloss: 0.133523\n",
            "[357]\tvalid_0's auc: 0.845551\tvalid_0's binary_logloss: 0.133519\n",
            "[358]\tvalid_0's auc: 0.845564\tvalid_0's binary_logloss: 0.133512\n",
            "[359]\tvalid_0's auc: 0.845586\tvalid_0's binary_logloss: 0.133497\n",
            "[360]\tvalid_0's auc: 0.845548\tvalid_0's binary_logloss: 0.133494\n",
            "[361]\tvalid_0's auc: 0.845569\tvalid_0's binary_logloss: 0.133485\n",
            "[362]\tvalid_0's auc: 0.845551\tvalid_0's binary_logloss: 0.133481\n",
            "[363]\tvalid_0's auc: 0.845619\tvalid_0's binary_logloss: 0.133469\n",
            "[364]\tvalid_0's auc: 0.845631\tvalid_0's binary_logloss: 0.13346\n",
            "[365]\tvalid_0's auc: 0.845657\tvalid_0's binary_logloss: 0.133444\n",
            "[366]\tvalid_0's auc: 0.845652\tvalid_0's binary_logloss: 0.133439\n",
            "[367]\tvalid_0's auc: 0.845661\tvalid_0's binary_logloss: 0.133425\n",
            "[368]\tvalid_0's auc: 0.845668\tvalid_0's binary_logloss: 0.133423\n",
            "[369]\tvalid_0's auc: 0.845611\tvalid_0's binary_logloss: 0.133428\n",
            "[370]\tvalid_0's auc: 0.845599\tvalid_0's binary_logloss: 0.13342\n",
            "[371]\tvalid_0's auc: 0.845638\tvalid_0's binary_logloss: 0.133408\n",
            "[372]\tvalid_0's auc: 0.84564\tvalid_0's binary_logloss: 0.133403\n",
            "[373]\tvalid_0's auc: 0.845648\tvalid_0's binary_logloss: 0.133394\n",
            "[374]\tvalid_0's auc: 0.845619\tvalid_0's binary_logloss: 0.133393\n",
            "[375]\tvalid_0's auc: 0.845627\tvalid_0's binary_logloss: 0.133393\n",
            "[376]\tvalid_0's auc: 0.845646\tvalid_0's binary_logloss: 0.13338\n",
            "[377]\tvalid_0's auc: 0.845601\tvalid_0's binary_logloss: 0.133382\n",
            "[378]\tvalid_0's auc: 0.845587\tvalid_0's binary_logloss: 0.133375\n",
            "[379]\tvalid_0's auc: 0.845572\tvalid_0's binary_logloss: 0.133375\n",
            "[380]\tvalid_0's auc: 0.845554\tvalid_0's binary_logloss: 0.133375\n",
            "[381]\tvalid_0's auc: 0.845524\tvalid_0's binary_logloss: 0.133372\n",
            "[382]\tvalid_0's auc: 0.845554\tvalid_0's binary_logloss: 0.13336\n",
            "[383]\tvalid_0's auc: 0.84561\tvalid_0's binary_logloss: 0.133354\n",
            "[384]\tvalid_0's auc: 0.845619\tvalid_0's binary_logloss: 0.133346\n",
            "[385]\tvalid_0's auc: 0.845634\tvalid_0's binary_logloss: 0.133335\n",
            "[386]\tvalid_0's auc: 0.845621\tvalid_0's binary_logloss: 0.133332\n",
            "[387]\tvalid_0's auc: 0.845619\tvalid_0's binary_logloss: 0.133326\n",
            "[388]\tvalid_0's auc: 0.845653\tvalid_0's binary_logloss: 0.133317\n",
            "[389]\tvalid_0's auc: 0.845654\tvalid_0's binary_logloss: 0.133312\n",
            "[390]\tvalid_0's auc: 0.84571\tvalid_0's binary_logloss: 0.133298\n",
            "[391]\tvalid_0's auc: 0.845716\tvalid_0's binary_logloss: 0.133291\n",
            "[392]\tvalid_0's auc: 0.845722\tvalid_0's binary_logloss: 0.133287\n",
            "[393]\tvalid_0's auc: 0.845702\tvalid_0's binary_logloss: 0.133285\n",
            "[394]\tvalid_0's auc: 0.8457\tvalid_0's binary_logloss: 0.13328\n",
            "[395]\tvalid_0's auc: 0.845696\tvalid_0's binary_logloss: 0.133276\n",
            "[396]\tvalid_0's auc: 0.845714\tvalid_0's binary_logloss: 0.133263\n",
            "[397]\tvalid_0's auc: 0.845757\tvalid_0's binary_logloss: 0.133256\n",
            "[398]\tvalid_0's auc: 0.84575\tvalid_0's binary_logloss: 0.133248\n",
            "[399]\tvalid_0's auc: 0.845731\tvalid_0's binary_logloss: 0.133246\n",
            "[400]\tvalid_0's auc: 0.845731\tvalid_0's binary_logloss: 0.133245\n",
            "[401]\tvalid_0's auc: 0.845749\tvalid_0's binary_logloss: 0.133239\n",
            "[402]\tvalid_0's auc: 0.845732\tvalid_0's binary_logloss: 0.133234\n",
            "[403]\tvalid_0's auc: 0.845737\tvalid_0's binary_logloss: 0.133228\n",
            "[404]\tvalid_0's auc: 0.845746\tvalid_0's binary_logloss: 0.133222\n",
            "[405]\tvalid_0's auc: 0.845737\tvalid_0's binary_logloss: 0.133223\n",
            "[406]\tvalid_0's auc: 0.845731\tvalid_0's binary_logloss: 0.133217\n",
            "[407]\tvalid_0's auc: 0.845767\tvalid_0's binary_logloss: 0.133213\n",
            "[408]\tvalid_0's auc: 0.845787\tvalid_0's binary_logloss: 0.13321\n",
            "[409]\tvalid_0's auc: 0.845788\tvalid_0's binary_logloss: 0.1332\n",
            "[410]\tvalid_0's auc: 0.845792\tvalid_0's binary_logloss: 0.133196\n",
            "[411]\tvalid_0's auc: 0.845783\tvalid_0's binary_logloss: 0.133195\n",
            "[412]\tvalid_0's auc: 0.845822\tvalid_0's binary_logloss: 0.133183\n",
            "[413]\tvalid_0's auc: 0.845847\tvalid_0's binary_logloss: 0.13318\n",
            "[414]\tvalid_0's auc: 0.845864\tvalid_0's binary_logloss: 0.133176\n",
            "[415]\tvalid_0's auc: 0.845896\tvalid_0's binary_logloss: 0.133171\n",
            "[416]\tvalid_0's auc: 0.845901\tvalid_0's binary_logloss: 0.133162\n",
            "[417]\tvalid_0's auc: 0.8459\tvalid_0's binary_logloss: 0.133157\n",
            "[418]\tvalid_0's auc: 0.845902\tvalid_0's binary_logloss: 0.133159\n",
            "[419]\tvalid_0's auc: 0.845932\tvalid_0's binary_logloss: 0.133149\n",
            "[420]\tvalid_0's auc: 0.845956\tvalid_0's binary_logloss: 0.133145\n",
            "[421]\tvalid_0's auc: 0.845945\tvalid_0's binary_logloss: 0.13314\n",
            "[422]\tvalid_0's auc: 0.845971\tvalid_0's binary_logloss: 0.133133\n",
            "[423]\tvalid_0's auc: 0.845966\tvalid_0's binary_logloss: 0.133129\n",
            "[424]\tvalid_0's auc: 0.846011\tvalid_0's binary_logloss: 0.133117\n",
            "[425]\tvalid_0's auc: 0.84602\tvalid_0's binary_logloss: 0.133112\n",
            "[426]\tvalid_0's auc: 0.846007\tvalid_0's binary_logloss: 0.133112\n",
            "[427]\tvalid_0's auc: 0.846019\tvalid_0's binary_logloss: 0.133107\n",
            "[428]\tvalid_0's auc: 0.846024\tvalid_0's binary_logloss: 0.133104\n",
            "[429]\tvalid_0's auc: 0.846047\tvalid_0's binary_logloss: 0.1331\n",
            "[430]\tvalid_0's auc: 0.846022\tvalid_0's binary_logloss: 0.133104\n",
            "[431]\tvalid_0's auc: 0.846039\tvalid_0's binary_logloss: 0.133102\n",
            "[432]\tvalid_0's auc: 0.846025\tvalid_0's binary_logloss: 0.133105\n",
            "[433]\tvalid_0's auc: 0.846039\tvalid_0's binary_logloss: 0.133099\n",
            "[434]\tvalid_0's auc: 0.846026\tvalid_0's binary_logloss: 0.1331\n",
            "[435]\tvalid_0's auc: 0.846001\tvalid_0's binary_logloss: 0.133097\n",
            "[436]\tvalid_0's auc: 0.846002\tvalid_0's binary_logloss: 0.133097\n",
            "[437]\tvalid_0's auc: 0.846011\tvalid_0's binary_logloss: 0.133094\n",
            "[438]\tvalid_0's auc: 0.84602\tvalid_0's binary_logloss: 0.133091\n",
            "[439]\tvalid_0's auc: 0.846019\tvalid_0's binary_logloss: 0.133088\n",
            "[440]\tvalid_0's auc: 0.846\tvalid_0's binary_logloss: 0.133087\n",
            "[441]\tvalid_0's auc: 0.845986\tvalid_0's binary_logloss: 0.133086\n",
            "[442]\tvalid_0's auc: 0.846008\tvalid_0's binary_logloss: 0.133081\n",
            "[443]\tvalid_0's auc: 0.845971\tvalid_0's binary_logloss: 0.133085\n",
            "[444]\tvalid_0's auc: 0.846014\tvalid_0's binary_logloss: 0.133073\n",
            "[445]\tvalid_0's auc: 0.84602\tvalid_0's binary_logloss: 0.133068\n",
            "[446]\tvalid_0's auc: 0.84601\tvalid_0's binary_logloss: 0.13307\n",
            "[447]\tvalid_0's auc: 0.846042\tvalid_0's binary_logloss: 0.133063\n",
            "[448]\tvalid_0's auc: 0.846024\tvalid_0's binary_logloss: 0.133058\n",
            "[449]\tvalid_0's auc: 0.846074\tvalid_0's binary_logloss: 0.133049\n",
            "[450]\tvalid_0's auc: 0.846094\tvalid_0's binary_logloss: 0.13304\n",
            "[451]\tvalid_0's auc: 0.846113\tvalid_0's binary_logloss: 0.133037\n",
            "[452]\tvalid_0's auc: 0.846098\tvalid_0's binary_logloss: 0.133035\n",
            "[453]\tvalid_0's auc: 0.846051\tvalid_0's binary_logloss: 0.133045\n",
            "[454]\tvalid_0's auc: 0.84611\tvalid_0's binary_logloss: 0.133035\n",
            "[455]\tvalid_0's auc: 0.84611\tvalid_0's binary_logloss: 0.133034\n",
            "[456]\tvalid_0's auc: 0.846077\tvalid_0's binary_logloss: 0.133035\n",
            "[457]\tvalid_0's auc: 0.846119\tvalid_0's binary_logloss: 0.133033\n",
            "[458]\tvalid_0's auc: 0.846168\tvalid_0's binary_logloss: 0.133023\n",
            "[459]\tvalid_0's auc: 0.846173\tvalid_0's binary_logloss: 0.133019\n",
            "[460]\tvalid_0's auc: 0.846175\tvalid_0's binary_logloss: 0.133015\n",
            "[461]\tvalid_0's auc: 0.846168\tvalid_0's binary_logloss: 0.133012\n",
            "[462]\tvalid_0's auc: 0.846174\tvalid_0's binary_logloss: 0.133005\n",
            "[463]\tvalid_0's auc: 0.846229\tvalid_0's binary_logloss: 0.133\n",
            "[464]\tvalid_0's auc: 0.84622\tvalid_0's binary_logloss: 0.132996\n",
            "[465]\tvalid_0's auc: 0.846218\tvalid_0's binary_logloss: 0.132997\n",
            "[466]\tvalid_0's auc: 0.846219\tvalid_0's binary_logloss: 0.132997\n",
            "[467]\tvalid_0's auc: 0.846288\tvalid_0's binary_logloss: 0.132988\n",
            "[468]\tvalid_0's auc: 0.84627\tvalid_0's binary_logloss: 0.132986\n",
            "[469]\tvalid_0's auc: 0.84628\tvalid_0's binary_logloss: 0.132981\n",
            "[470]\tvalid_0's auc: 0.846323\tvalid_0's binary_logloss: 0.132978\n",
            "[471]\tvalid_0's auc: 0.846317\tvalid_0's binary_logloss: 0.132979\n",
            "[472]\tvalid_0's auc: 0.846295\tvalid_0's binary_logloss: 0.132978\n",
            "[473]\tvalid_0's auc: 0.846301\tvalid_0's binary_logloss: 0.132975\n",
            "[474]\tvalid_0's auc: 0.846293\tvalid_0's binary_logloss: 0.132973\n",
            "[475]\tvalid_0's auc: 0.846286\tvalid_0's binary_logloss: 0.132972\n",
            "[476]\tvalid_0's auc: 0.846329\tvalid_0's binary_logloss: 0.132966\n",
            "[477]\tvalid_0's auc: 0.846321\tvalid_0's binary_logloss: 0.132964\n",
            "[478]\tvalid_0's auc: 0.846303\tvalid_0's binary_logloss: 0.132961\n",
            "[479]\tvalid_0's auc: 0.846304\tvalid_0's binary_logloss: 0.132961\n",
            "[480]\tvalid_0's auc: 0.846296\tvalid_0's binary_logloss: 0.13296\n",
            "[481]\tvalid_0's auc: 0.8463\tvalid_0's binary_logloss: 0.132965\n",
            "[482]\tvalid_0's auc: 0.846277\tvalid_0's binary_logloss: 0.132962\n",
            "[483]\tvalid_0's auc: 0.846293\tvalid_0's binary_logloss: 0.132957\n",
            "[484]\tvalid_0's auc: 0.846308\tvalid_0's binary_logloss: 0.132956\n",
            "[485]\tvalid_0's auc: 0.846298\tvalid_0's binary_logloss: 0.132955\n",
            "[486]\tvalid_0's auc: 0.846311\tvalid_0's binary_logloss: 0.132949\n",
            "[487]\tvalid_0's auc: 0.84629\tvalid_0's binary_logloss: 0.132952\n",
            "[488]\tvalid_0's auc: 0.846284\tvalid_0's binary_logloss: 0.132948\n",
            "[489]\tvalid_0's auc: 0.846246\tvalid_0's binary_logloss: 0.132956\n",
            "[490]\tvalid_0's auc: 0.8463\tvalid_0's binary_logloss: 0.132947\n",
            "[491]\tvalid_0's auc: 0.846312\tvalid_0's binary_logloss: 0.132946\n",
            "[492]\tvalid_0's auc: 0.846343\tvalid_0's binary_logloss: 0.132938\n",
            "[493]\tvalid_0's auc: 0.846332\tvalid_0's binary_logloss: 0.132936\n",
            "[494]\tvalid_0's auc: 0.8463\tvalid_0's binary_logloss: 0.132938\n",
            "[495]\tvalid_0's auc: 0.846274\tvalid_0's binary_logloss: 0.132935\n",
            "[496]\tvalid_0's auc: 0.846248\tvalid_0's binary_logloss: 0.132935\n",
            "[497]\tvalid_0's auc: 0.846236\tvalid_0's binary_logloss: 0.132938\n",
            "[498]\tvalid_0's auc: 0.846266\tvalid_0's binary_logloss: 0.132931\n",
            "[499]\tvalid_0's auc: 0.84624\tvalid_0's binary_logloss: 0.132932\n",
            "[500]\tvalid_0's auc: 0.84625\tvalid_0's binary_logloss: 0.132929\n",
            "[501]\tvalid_0's auc: 0.846295\tvalid_0's binary_logloss: 0.132921\n",
            "[502]\tvalid_0's auc: 0.846277\tvalid_0's binary_logloss: 0.132918\n",
            "[503]\tvalid_0's auc: 0.846268\tvalid_0's binary_logloss: 0.132917\n",
            "[504]\tvalid_0's auc: 0.846245\tvalid_0's binary_logloss: 0.132916\n",
            "[505]\tvalid_0's auc: 0.84629\tvalid_0's binary_logloss: 0.132906\n",
            "[506]\tvalid_0's auc: 0.846279\tvalid_0's binary_logloss: 0.132909\n",
            "[507]\tvalid_0's auc: 0.846276\tvalid_0's binary_logloss: 0.132906\n",
            "[508]\tvalid_0's auc: 0.84629\tvalid_0's binary_logloss: 0.132902\n",
            "[509]\tvalid_0's auc: 0.846315\tvalid_0's binary_logloss: 0.132899\n",
            "[510]\tvalid_0's auc: 0.846309\tvalid_0's binary_logloss: 0.132902\n",
            "[511]\tvalid_0's auc: 0.846308\tvalid_0's binary_logloss: 0.132904\n",
            "[512]\tvalid_0's auc: 0.846297\tvalid_0's binary_logloss: 0.132901\n",
            "[513]\tvalid_0's auc: 0.846323\tvalid_0's binary_logloss: 0.132898\n",
            "[514]\tvalid_0's auc: 0.846301\tvalid_0's binary_logloss: 0.132897\n",
            "[515]\tvalid_0's auc: 0.846274\tvalid_0's binary_logloss: 0.132901\n",
            "[516]\tvalid_0's auc: 0.846289\tvalid_0's binary_logloss: 0.132904\n",
            "[517]\tvalid_0's auc: 0.846261\tvalid_0's binary_logloss: 0.132905\n",
            "[518]\tvalid_0's auc: 0.846268\tvalid_0's binary_logloss: 0.132907\n",
            "[519]\tvalid_0's auc: 0.846272\tvalid_0's binary_logloss: 0.132903\n",
            "[520]\tvalid_0's auc: 0.846265\tvalid_0's binary_logloss: 0.132901\n",
            "[521]\tvalid_0's auc: 0.846286\tvalid_0's binary_logloss: 0.132899\n",
            "[522]\tvalid_0's auc: 0.84629\tvalid_0's binary_logloss: 0.132899\n",
            "[523]\tvalid_0's auc: 0.846291\tvalid_0's binary_logloss: 0.132896\n",
            "[524]\tvalid_0's auc: 0.846291\tvalid_0's binary_logloss: 0.132894\n",
            "[525]\tvalid_0's auc: 0.846276\tvalid_0's binary_logloss: 0.132897\n",
            "[526]\tvalid_0's auc: 0.846283\tvalid_0's binary_logloss: 0.13289\n",
            "[527]\tvalid_0's auc: 0.846293\tvalid_0's binary_logloss: 0.132889\n",
            "[528]\tvalid_0's auc: 0.846319\tvalid_0's binary_logloss: 0.132884\n",
            "[529]\tvalid_0's auc: 0.846319\tvalid_0's binary_logloss: 0.132884\n",
            "[530]\tvalid_0's auc: 0.846302\tvalid_0's binary_logloss: 0.132883\n",
            "[531]\tvalid_0's auc: 0.846298\tvalid_0's binary_logloss: 0.13288\n",
            "[532]\tvalid_0's auc: 0.846326\tvalid_0's binary_logloss: 0.132875\n",
            "[533]\tvalid_0's auc: 0.846326\tvalid_0's binary_logloss: 0.13287\n",
            "[534]\tvalid_0's auc: 0.84634\tvalid_0's binary_logloss: 0.132865\n",
            "[535]\tvalid_0's auc: 0.846308\tvalid_0's binary_logloss: 0.132865\n",
            "[536]\tvalid_0's auc: 0.846318\tvalid_0's binary_logloss: 0.132861\n",
            "[537]\tvalid_0's auc: 0.846365\tvalid_0's binary_logloss: 0.132857\n",
            "[538]\tvalid_0's auc: 0.846373\tvalid_0's binary_logloss: 0.132855\n",
            "[539]\tvalid_0's auc: 0.846369\tvalid_0's binary_logloss: 0.132855\n",
            "[540]\tvalid_0's auc: 0.846347\tvalid_0's binary_logloss: 0.132856\n",
            "[541]\tvalid_0's auc: 0.846333\tvalid_0's binary_logloss: 0.132853\n",
            "[542]\tvalid_0's auc: 0.84631\tvalid_0's binary_logloss: 0.132856\n",
            "[543]\tvalid_0's auc: 0.846302\tvalid_0's binary_logloss: 0.132859\n",
            "[544]\tvalid_0's auc: 0.846279\tvalid_0's binary_logloss: 0.132856\n",
            "[545]\tvalid_0's auc: 0.846301\tvalid_0's binary_logloss: 0.13285\n",
            "[546]\tvalid_0's auc: 0.846297\tvalid_0's binary_logloss: 0.13285\n",
            "[547]\tvalid_0's auc: 0.846309\tvalid_0's binary_logloss: 0.132845\n",
            "[548]\tvalid_0's auc: 0.846338\tvalid_0's binary_logloss: 0.132842\n",
            "[549]\tvalid_0's auc: 0.846345\tvalid_0's binary_logloss: 0.132844\n",
            "[550]\tvalid_0's auc: 0.846382\tvalid_0's binary_logloss: 0.132841\n",
            "[551]\tvalid_0's auc: 0.846372\tvalid_0's binary_logloss: 0.132842\n",
            "[552]\tvalid_0's auc: 0.846389\tvalid_0's binary_logloss: 0.13284\n",
            "[553]\tvalid_0's auc: 0.846342\tvalid_0's binary_logloss: 0.132839\n",
            "[554]\tvalid_0's auc: 0.846367\tvalid_0's binary_logloss: 0.132836\n",
            "[555]\tvalid_0's auc: 0.846397\tvalid_0's binary_logloss: 0.132835\n",
            "[556]\tvalid_0's auc: 0.846364\tvalid_0's binary_logloss: 0.132833\n",
            "[557]\tvalid_0's auc: 0.846354\tvalid_0's binary_logloss: 0.132831\n",
            "[558]\tvalid_0's auc: 0.846332\tvalid_0's binary_logloss: 0.132833\n",
            "[559]\tvalid_0's auc: 0.846331\tvalid_0's binary_logloss: 0.132834\n",
            "[560]\tvalid_0's auc: 0.846321\tvalid_0's binary_logloss: 0.132831\n",
            "[561]\tvalid_0's auc: 0.846314\tvalid_0's binary_logloss: 0.132832\n",
            "[562]\tvalid_0's auc: 0.846308\tvalid_0's binary_logloss: 0.132832\n",
            "[563]\tvalid_0's auc: 0.846274\tvalid_0's binary_logloss: 0.13283\n",
            "[564]\tvalid_0's auc: 0.846335\tvalid_0's binary_logloss: 0.132821\n",
            "[565]\tvalid_0's auc: 0.84633\tvalid_0's binary_logloss: 0.132821\n",
            "[566]\tvalid_0's auc: 0.846302\tvalid_0's binary_logloss: 0.132823\n",
            "[567]\tvalid_0's auc: 0.846294\tvalid_0's binary_logloss: 0.132825\n",
            "[568]\tvalid_0's auc: 0.846287\tvalid_0's binary_logloss: 0.132823\n",
            "[569]\tvalid_0's auc: 0.846328\tvalid_0's binary_logloss: 0.132813\n",
            "[570]\tvalid_0's auc: 0.846329\tvalid_0's binary_logloss: 0.132812\n",
            "[571]\tvalid_0's auc: 0.846346\tvalid_0's binary_logloss: 0.132805\n",
            "[572]\tvalid_0's auc: 0.846333\tvalid_0's binary_logloss: 0.132807\n",
            "[573]\tvalid_0's auc: 0.846349\tvalid_0's binary_logloss: 0.1328\n",
            "[574]\tvalid_0's auc: 0.846327\tvalid_0's binary_logloss: 0.132801\n",
            "[575]\tvalid_0's auc: 0.846268\tvalid_0's binary_logloss: 0.132799\n",
            "[576]\tvalid_0's auc: 0.846308\tvalid_0's binary_logloss: 0.132798\n",
            "[577]\tvalid_0's auc: 0.846312\tvalid_0's binary_logloss: 0.132794\n",
            "[578]\tvalid_0's auc: 0.846334\tvalid_0's binary_logloss: 0.132792\n",
            "[579]\tvalid_0's auc: 0.846327\tvalid_0's binary_logloss: 0.132794\n",
            "[580]\tvalid_0's auc: 0.846318\tvalid_0's binary_logloss: 0.132796\n",
            "[581]\tvalid_0's auc: 0.846415\tvalid_0's binary_logloss: 0.132788\n",
            "[582]\tvalid_0's auc: 0.846394\tvalid_0's binary_logloss: 0.132793\n",
            "[583]\tvalid_0's auc: 0.846407\tvalid_0's binary_logloss: 0.132796\n",
            "[584]\tvalid_0's auc: 0.846352\tvalid_0's binary_logloss: 0.132799\n",
            "[585]\tvalid_0's auc: 0.846334\tvalid_0's binary_logloss: 0.132801\n",
            "[586]\tvalid_0's auc: 0.846374\tvalid_0's binary_logloss: 0.1328\n",
            "[587]\tvalid_0's auc: 0.846385\tvalid_0's binary_logloss: 0.1328\n",
            "[588]\tvalid_0's auc: 0.846348\tvalid_0's binary_logloss: 0.132807\n",
            "[589]\tvalid_0's auc: 0.846338\tvalid_0's binary_logloss: 0.132808\n",
            "[590]\tvalid_0's auc: 0.84635\tvalid_0's binary_logloss: 0.132805\n",
            "[591]\tvalid_0's auc: 0.846316\tvalid_0's binary_logloss: 0.132806\n",
            "[592]\tvalid_0's auc: 0.846293\tvalid_0's binary_logloss: 0.132809\n",
            "[593]\tvalid_0's auc: 0.846324\tvalid_0's binary_logloss: 0.132808\n",
            "[594]\tvalid_0's auc: 0.84626\tvalid_0's binary_logloss: 0.132812\n",
            "[595]\tvalid_0's auc: 0.846273\tvalid_0's binary_logloss: 0.132802\n",
            "[596]\tvalid_0's auc: 0.846275\tvalid_0's binary_logloss: 0.1328\n",
            "[597]\tvalid_0's auc: 0.846274\tvalid_0's binary_logloss: 0.132801\n",
            "[598]\tvalid_0's auc: 0.846291\tvalid_0's binary_logloss: 0.132797\n",
            "[599]\tvalid_0's auc: 0.84628\tvalid_0's binary_logloss: 0.132791\n",
            "[600]\tvalid_0's auc: 0.846263\tvalid_0's binary_logloss: 0.132794\n",
            "[601]\tvalid_0's auc: 0.846238\tvalid_0's binary_logloss: 0.132795\n",
            "[602]\tvalid_0's auc: 0.846268\tvalid_0's binary_logloss: 0.132792\n",
            "[603]\tvalid_0's auc: 0.846242\tvalid_0's binary_logloss: 0.132796\n",
            "[604]\tvalid_0's auc: 0.846272\tvalid_0's binary_logloss: 0.132791\n",
            "[605]\tvalid_0's auc: 0.84627\tvalid_0's binary_logloss: 0.132788\n",
            "[606]\tvalid_0's auc: 0.846287\tvalid_0's binary_logloss: 0.132789\n",
            "[607]\tvalid_0's auc: 0.846341\tvalid_0's binary_logloss: 0.132783\n",
            "[608]\tvalid_0's auc: 0.846326\tvalid_0's binary_logloss: 0.132782\n",
            "[609]\tvalid_0's auc: 0.846322\tvalid_0's binary_logloss: 0.132786\n",
            "[610]\tvalid_0's auc: 0.846285\tvalid_0's binary_logloss: 0.132786\n",
            "[611]\tvalid_0's auc: 0.846268\tvalid_0's binary_logloss: 0.132792\n",
            "[612]\tvalid_0's auc: 0.846258\tvalid_0's binary_logloss: 0.132791\n",
            "[613]\tvalid_0's auc: 0.846253\tvalid_0's binary_logloss: 0.132791\n",
            "[614]\tvalid_0's auc: 0.846263\tvalid_0's binary_logloss: 0.132789\n",
            "[615]\tvalid_0's auc: 0.846271\tvalid_0's binary_logloss: 0.132786\n",
            "[616]\tvalid_0's auc: 0.846245\tvalid_0's binary_logloss: 0.132787\n",
            "[617]\tvalid_0's auc: 0.846196\tvalid_0's binary_logloss: 0.132788\n",
            "[618]\tvalid_0's auc: 0.846177\tvalid_0's binary_logloss: 0.132786\n",
            "[619]\tvalid_0's auc: 0.846175\tvalid_0's binary_logloss: 0.132786\n",
            "[620]\tvalid_0's auc: 0.846149\tvalid_0's binary_logloss: 0.132786\n",
            "[621]\tvalid_0's auc: 0.846155\tvalid_0's binary_logloss: 0.132784\n",
            "[622]\tvalid_0's auc: 0.846176\tvalid_0's binary_logloss: 0.132781\n",
            "[623]\tvalid_0's auc: 0.846163\tvalid_0's binary_logloss: 0.132778\n",
            "[624]\tvalid_0's auc: 0.846168\tvalid_0's binary_logloss: 0.132779\n",
            "[625]\tvalid_0's auc: 0.84617\tvalid_0's binary_logloss: 0.132783\n",
            "[626]\tvalid_0's auc: 0.846164\tvalid_0's binary_logloss: 0.132784\n",
            "[627]\tvalid_0's auc: 0.846163\tvalid_0's binary_logloss: 0.132785\n",
            "[628]\tvalid_0's auc: 0.846159\tvalid_0's binary_logloss: 0.132785\n",
            "[629]\tvalid_0's auc: 0.846148\tvalid_0's binary_logloss: 0.132788\n",
            "[630]\tvalid_0's auc: 0.846144\tvalid_0's binary_logloss: 0.132786\n",
            "[631]\tvalid_0's auc: 0.846166\tvalid_0's binary_logloss: 0.132779\n",
            "[632]\tvalid_0's auc: 0.846179\tvalid_0's binary_logloss: 0.132777\n",
            "[633]\tvalid_0's auc: 0.846175\tvalid_0's binary_logloss: 0.132771\n",
            "[634]\tvalid_0's auc: 0.846154\tvalid_0's binary_logloss: 0.132768\n",
            "[635]\tvalid_0's auc: 0.846146\tvalid_0's binary_logloss: 0.132771\n",
            "[636]\tvalid_0's auc: 0.846128\tvalid_0's binary_logloss: 0.132775\n",
            "[637]\tvalid_0's auc: 0.846114\tvalid_0's binary_logloss: 0.132774\n",
            "[638]\tvalid_0's auc: 0.846132\tvalid_0's binary_logloss: 0.132772\n",
            "[639]\tvalid_0's auc: 0.846128\tvalid_0's binary_logloss: 0.132766\n",
            "[640]\tvalid_0's auc: 0.846149\tvalid_0's binary_logloss: 0.132756\n",
            "[641]\tvalid_0's auc: 0.846168\tvalid_0's binary_logloss: 0.132751\n",
            "[642]\tvalid_0's auc: 0.846149\tvalid_0's binary_logloss: 0.132754\n",
            "[643]\tvalid_0's auc: 0.846146\tvalid_0's binary_logloss: 0.132753\n",
            "[644]\tvalid_0's auc: 0.846188\tvalid_0's binary_logloss: 0.132745\n",
            "[645]\tvalid_0's auc: 0.846195\tvalid_0's binary_logloss: 0.132742\n",
            "[646]\tvalid_0's auc: 0.846208\tvalid_0's binary_logloss: 0.13274\n",
            "[647]\tvalid_0's auc: 0.846181\tvalid_0's binary_logloss: 0.132741\n",
            "[648]\tvalid_0's auc: 0.846158\tvalid_0's binary_logloss: 0.132744\n",
            "[649]\tvalid_0's auc: 0.846188\tvalid_0's binary_logloss: 0.13274\n",
            "[650]\tvalid_0's auc: 0.846169\tvalid_0's binary_logloss: 0.132745\n",
            "[651]\tvalid_0's auc: 0.846163\tvalid_0's binary_logloss: 0.132747\n",
            "[652]\tvalid_0's auc: 0.846163\tvalid_0's binary_logloss: 0.132741\n",
            "[653]\tvalid_0's auc: 0.846152\tvalid_0's binary_logloss: 0.132743\n",
            "[654]\tvalid_0's auc: 0.846235\tvalid_0's binary_logloss: 0.132736\n",
            "[655]\tvalid_0's auc: 0.84622\tvalid_0's binary_logloss: 0.132735\n",
            "[656]\tvalid_0's auc: 0.846276\tvalid_0's binary_logloss: 0.132729\n",
            "[657]\tvalid_0's auc: 0.84629\tvalid_0's binary_logloss: 0.132731\n",
            "[658]\tvalid_0's auc: 0.846312\tvalid_0's binary_logloss: 0.132729\n",
            "[659]\tvalid_0's auc: 0.846325\tvalid_0's binary_logloss: 0.132726\n",
            "[660]\tvalid_0's auc: 0.846313\tvalid_0's binary_logloss: 0.132722\n",
            "[661]\tvalid_0's auc: 0.846305\tvalid_0's binary_logloss: 0.132725\n",
            "[662]\tvalid_0's auc: 0.84629\tvalid_0's binary_logloss: 0.132723\n",
            "[663]\tvalid_0's auc: 0.846297\tvalid_0's binary_logloss: 0.132728\n",
            "[664]\tvalid_0's auc: 0.846293\tvalid_0's binary_logloss: 0.13273\n",
            "[665]\tvalid_0's auc: 0.8463\tvalid_0's binary_logloss: 0.13273\n",
            "[666]\tvalid_0's auc: 0.846292\tvalid_0's binary_logloss: 0.132733\n",
            "[667]\tvalid_0's auc: 0.846263\tvalid_0's binary_logloss: 0.132737\n",
            "[668]\tvalid_0's auc: 0.846258\tvalid_0's binary_logloss: 0.132736\n",
            "[669]\tvalid_0's auc: 0.846243\tvalid_0's binary_logloss: 0.132734\n",
            "[670]\tvalid_0's auc: 0.846281\tvalid_0's binary_logloss: 0.132732\n",
            "[671]\tvalid_0's auc: 0.84628\tvalid_0's binary_logloss: 0.132733\n",
            "[672]\tvalid_0's auc: 0.846272\tvalid_0's binary_logloss: 0.132731\n",
            "[673]\tvalid_0's auc: 0.84627\tvalid_0's binary_logloss: 0.13273\n",
            "[674]\tvalid_0's auc: 0.846251\tvalid_0's binary_logloss: 0.132733\n",
            "[675]\tvalid_0's auc: 0.846238\tvalid_0's binary_logloss: 0.132733\n",
            "[676]\tvalid_0's auc: 0.846196\tvalid_0's binary_logloss: 0.132735\n",
            "[677]\tvalid_0's auc: 0.846187\tvalid_0's binary_logloss: 0.132734\n",
            "[678]\tvalid_0's auc: 0.846153\tvalid_0's binary_logloss: 0.132734\n",
            "[679]\tvalid_0's auc: 0.846142\tvalid_0's binary_logloss: 0.132736\n",
            "[680]\tvalid_0's auc: 0.846118\tvalid_0's binary_logloss: 0.132742\n",
            "[681]\tvalid_0's auc: 0.846091\tvalid_0's binary_logloss: 0.132748\n",
            "[682]\tvalid_0's auc: 0.846095\tvalid_0's binary_logloss: 0.132745\n",
            "[683]\tvalid_0's auc: 0.846097\tvalid_0's binary_logloss: 0.132747\n",
            "[684]\tvalid_0's auc: 0.846078\tvalid_0's binary_logloss: 0.132745\n",
            "[685]\tvalid_0's auc: 0.846074\tvalid_0's binary_logloss: 0.132748\n",
            "[686]\tvalid_0's auc: 0.84607\tvalid_0's binary_logloss: 0.132748\n",
            "[687]\tvalid_0's auc: 0.846088\tvalid_0's binary_logloss: 0.132745\n",
            "[688]\tvalid_0's auc: 0.846111\tvalid_0's binary_logloss: 0.13274\n",
            "[689]\tvalid_0's auc: 0.846077\tvalid_0's binary_logloss: 0.132746\n",
            "[690]\tvalid_0's auc: 0.846077\tvalid_0's binary_logloss: 0.132748\n",
            "[691]\tvalid_0's auc: 0.84609\tvalid_0's binary_logloss: 0.132747\n",
            "[692]\tvalid_0's auc: 0.846113\tvalid_0's binary_logloss: 0.13274\n",
            "[693]\tvalid_0's auc: 0.846127\tvalid_0's binary_logloss: 0.132738\n",
            "[694]\tvalid_0's auc: 0.846125\tvalid_0's binary_logloss: 0.132739\n",
            "[695]\tvalid_0's auc: 0.846134\tvalid_0's binary_logloss: 0.132738\n",
            "[696]\tvalid_0's auc: 0.846119\tvalid_0's binary_logloss: 0.13274\n",
            "[697]\tvalid_0's auc: 0.846118\tvalid_0's binary_logloss: 0.132741\n",
            "[698]\tvalid_0's auc: 0.846116\tvalid_0's binary_logloss: 0.132741\n",
            "[699]\tvalid_0's auc: 0.846157\tvalid_0's binary_logloss: 0.132735\n",
            "[700]\tvalid_0's auc: 0.846138\tvalid_0's binary_logloss: 0.132737\n",
            "[701]\tvalid_0's auc: 0.846132\tvalid_0's binary_logloss: 0.132736\n",
            "[702]\tvalid_0's auc: 0.846122\tvalid_0's binary_logloss: 0.132738\n",
            "[703]\tvalid_0's auc: 0.846106\tvalid_0's binary_logloss: 0.132734\n",
            "[704]\tvalid_0's auc: 0.846099\tvalid_0's binary_logloss: 0.132735\n",
            "[705]\tvalid_0's auc: 0.846089\tvalid_0's binary_logloss: 0.132737\n",
            "[706]\tvalid_0's auc: 0.846066\tvalid_0's binary_logloss: 0.132738\n",
            "[707]\tvalid_0's auc: 0.846057\tvalid_0's binary_logloss: 0.13274\n",
            "[708]\tvalid_0's auc: 0.846067\tvalid_0's binary_logloss: 0.132739\n",
            "[709]\tvalid_0's auc: 0.846038\tvalid_0's binary_logloss: 0.132743\n",
            "[710]\tvalid_0's auc: 0.846044\tvalid_0's binary_logloss: 0.132744\n",
            "[711]\tvalid_0's auc: 0.846059\tvalid_0's binary_logloss: 0.132742\n",
            "[712]\tvalid_0's auc: 0.846059\tvalid_0's binary_logloss: 0.132743\n",
            "[713]\tvalid_0's auc: 0.846049\tvalid_0's binary_logloss: 0.132743\n",
            "[714]\tvalid_0's auc: 0.846037\tvalid_0's binary_logloss: 0.132741\n",
            "[715]\tvalid_0's auc: 0.846054\tvalid_0's binary_logloss: 0.132739\n",
            "[716]\tvalid_0's auc: 0.846004\tvalid_0's binary_logloss: 0.132737\n",
            "[717]\tvalid_0's auc: 0.846003\tvalid_0's binary_logloss: 0.132736\n",
            "[718]\tvalid_0's auc: 0.84598\tvalid_0's binary_logloss: 0.132737\n",
            "[719]\tvalid_0's auc: 0.845975\tvalid_0's binary_logloss: 0.132738\n",
            "[720]\tvalid_0's auc: 0.84601\tvalid_0's binary_logloss: 0.13273\n",
            "[721]\tvalid_0's auc: 0.845998\tvalid_0's binary_logloss: 0.132729\n",
            "[722]\tvalid_0's auc: 0.845995\tvalid_0's binary_logloss: 0.13273\n",
            "[723]\tvalid_0's auc: 0.845961\tvalid_0's binary_logloss: 0.132736\n",
            "[724]\tvalid_0's auc: 0.845971\tvalid_0's binary_logloss: 0.132733\n",
            "[725]\tvalid_0's auc: 0.845945\tvalid_0's binary_logloss: 0.132733\n",
            "[726]\tvalid_0's auc: 0.845942\tvalid_0's binary_logloss: 0.132733\n",
            "[727]\tvalid_0's auc: 0.845942\tvalid_0's binary_logloss: 0.13273\n",
            "[728]\tvalid_0's auc: 0.845921\tvalid_0's binary_logloss: 0.132735\n",
            "[729]\tvalid_0's auc: 0.845903\tvalid_0's binary_logloss: 0.132739\n",
            "[730]\tvalid_0's auc: 0.845902\tvalid_0's binary_logloss: 0.132736\n",
            "[731]\tvalid_0's auc: 0.845893\tvalid_0's binary_logloss: 0.132733\n",
            "[732]\tvalid_0's auc: 0.845857\tvalid_0's binary_logloss: 0.132733\n",
            "[733]\tvalid_0's auc: 0.845879\tvalid_0's binary_logloss: 0.132728\n",
            "[734]\tvalid_0's auc: 0.845852\tvalid_0's binary_logloss: 0.132733\n",
            "[735]\tvalid_0's auc: 0.845825\tvalid_0's binary_logloss: 0.132735\n",
            "[736]\tvalid_0's auc: 0.845836\tvalid_0's binary_logloss: 0.132734\n",
            "[737]\tvalid_0's auc: 0.845821\tvalid_0's binary_logloss: 0.132735\n",
            "[738]\tvalid_0's auc: 0.845815\tvalid_0's binary_logloss: 0.132738\n",
            "[739]\tvalid_0's auc: 0.8458\tvalid_0's binary_logloss: 0.132743\n",
            "[740]\tvalid_0's auc: 0.845833\tvalid_0's binary_logloss: 0.132741\n",
            "[741]\tvalid_0's auc: 0.845804\tvalid_0's binary_logloss: 0.132748\n",
            "[742]\tvalid_0's auc: 0.845814\tvalid_0's binary_logloss: 0.132744\n",
            "[743]\tvalid_0's auc: 0.845826\tvalid_0's binary_logloss: 0.132744\n",
            "[744]\tvalid_0's auc: 0.845859\tvalid_0's binary_logloss: 0.132741\n",
            "[745]\tvalid_0's auc: 0.845824\tvalid_0's binary_logloss: 0.132745\n",
            "[746]\tvalid_0's auc: 0.845815\tvalid_0's binary_logloss: 0.132746\n",
            "[747]\tvalid_0's auc: 0.845826\tvalid_0's binary_logloss: 0.132744\n",
            "[748]\tvalid_0's auc: 0.845809\tvalid_0's binary_logloss: 0.132746\n",
            "[749]\tvalid_0's auc: 0.845812\tvalid_0's binary_logloss: 0.132747\n",
            "[750]\tvalid_0's auc: 0.845777\tvalid_0's binary_logloss: 0.132753\n",
            "[751]\tvalid_0's auc: 0.845753\tvalid_0's binary_logloss: 0.132755\n",
            "[752]\tvalid_0's auc: 0.845767\tvalid_0's binary_logloss: 0.132756\n",
            "[753]\tvalid_0's auc: 0.845776\tvalid_0's binary_logloss: 0.132755\n",
            "[754]\tvalid_0's auc: 0.845752\tvalid_0's binary_logloss: 0.132759\n",
            "[755]\tvalid_0's auc: 0.845722\tvalid_0's binary_logloss: 0.132764\n",
            "[756]\tvalid_0's auc: 0.84568\tvalid_0's binary_logloss: 0.13277\n",
            "[757]\tvalid_0's auc: 0.845717\tvalid_0's binary_logloss: 0.132766\n",
            "[758]\tvalid_0's auc: 0.845705\tvalid_0's binary_logloss: 0.132767\n",
            "[759]\tvalid_0's auc: 0.84573\tvalid_0's binary_logloss: 0.132763\n",
            "[760]\tvalid_0's auc: 0.845748\tvalid_0's binary_logloss: 0.132762\n",
            "[761]\tvalid_0's auc: 0.845746\tvalid_0's binary_logloss: 0.132758\n",
            "[762]\tvalid_0's auc: 0.845724\tvalid_0's binary_logloss: 0.132759\n",
            "[763]\tvalid_0's auc: 0.845746\tvalid_0's binary_logloss: 0.13276\n",
            "[764]\tvalid_0's auc: 0.845749\tvalid_0's binary_logloss: 0.132759\n",
            "[765]\tvalid_0's auc: 0.845762\tvalid_0's binary_logloss: 0.132759\n",
            "[766]\tvalid_0's auc: 0.845755\tvalid_0's binary_logloss: 0.132762\n",
            "[767]\tvalid_0's auc: 0.845743\tvalid_0's binary_logloss: 0.132764\n",
            "[768]\tvalid_0's auc: 0.845776\tvalid_0's binary_logloss: 0.132755\n",
            "[769]\tvalid_0's auc: 0.845745\tvalid_0's binary_logloss: 0.132761\n",
            "[770]\tvalid_0's auc: 0.845734\tvalid_0's binary_logloss: 0.132765\n",
            "[771]\tvalid_0's auc: 0.845754\tvalid_0's binary_logloss: 0.132759\n",
            "[772]\tvalid_0's auc: 0.8457\tvalid_0's binary_logloss: 0.132765\n",
            "[773]\tvalid_0's auc: 0.845677\tvalid_0's binary_logloss: 0.132766\n",
            "[774]\tvalid_0's auc: 0.845685\tvalid_0's binary_logloss: 0.132765\n",
            "[775]\tvalid_0's auc: 0.845671\tvalid_0's binary_logloss: 0.132765\n",
            "[776]\tvalid_0's auc: 0.845659\tvalid_0's binary_logloss: 0.132764\n",
            "[777]\tvalid_0's auc: 0.845639\tvalid_0's binary_logloss: 0.132766\n",
            "[778]\tvalid_0's auc: 0.84562\tvalid_0's binary_logloss: 0.132769\n",
            "[779]\tvalid_0's auc: 0.845657\tvalid_0's binary_logloss: 0.132769\n",
            "[780]\tvalid_0's auc: 0.845669\tvalid_0's binary_logloss: 0.132768\n",
            "[781]\tvalid_0's auc: 0.845653\tvalid_0's binary_logloss: 0.132774\n",
            "[782]\tvalid_0's auc: 0.845644\tvalid_0's binary_logloss: 0.132776\n",
            "[783]\tvalid_0's auc: 0.845677\tvalid_0's binary_logloss: 0.132769\n",
            "[784]\tvalid_0's auc: 0.845667\tvalid_0's binary_logloss: 0.132768\n",
            "[785]\tvalid_0's auc: 0.845667\tvalid_0's binary_logloss: 0.132769\n",
            "[786]\tvalid_0's auc: 0.845669\tvalid_0's binary_logloss: 0.132771\n",
            "[787]\tvalid_0's auc: 0.845667\tvalid_0's binary_logloss: 0.132773\n",
            "[788]\tvalid_0's auc: 0.845626\tvalid_0's binary_logloss: 0.132773\n",
            "[789]\tvalid_0's auc: 0.845656\tvalid_0's binary_logloss: 0.132773\n",
            "[790]\tvalid_0's auc: 0.845636\tvalid_0's binary_logloss: 0.132773\n",
            "[791]\tvalid_0's auc: 0.845658\tvalid_0's binary_logloss: 0.132771\n",
            "[792]\tvalid_0's auc: 0.845647\tvalid_0's binary_logloss: 0.132772\n",
            "[793]\tvalid_0's auc: 0.845609\tvalid_0's binary_logloss: 0.132777\n",
            "[794]\tvalid_0's auc: 0.845602\tvalid_0's binary_logloss: 0.132776\n",
            "[795]\tvalid_0's auc: 0.845602\tvalid_0's binary_logloss: 0.132777\n",
            "[796]\tvalid_0's auc: 0.84564\tvalid_0's binary_logloss: 0.132776\n",
            "[797]\tvalid_0's auc: 0.845636\tvalid_0's binary_logloss: 0.132778\n",
            "[798]\tvalid_0's auc: 0.845632\tvalid_0's binary_logloss: 0.132775\n",
            "[799]\tvalid_0's auc: 0.845673\tvalid_0's binary_logloss: 0.132771\n",
            "[800]\tvalid_0's auc: 0.845664\tvalid_0's binary_logloss: 0.132766\n",
            "[801]\tvalid_0's auc: 0.845636\tvalid_0's binary_logloss: 0.132769\n",
            "[802]\tvalid_0's auc: 0.845616\tvalid_0's binary_logloss: 0.13277\n",
            "[803]\tvalid_0's auc: 0.845603\tvalid_0's binary_logloss: 0.132774\n",
            "[804]\tvalid_0's auc: 0.845595\tvalid_0's binary_logloss: 0.132778\n",
            "[805]\tvalid_0's auc: 0.845585\tvalid_0's binary_logloss: 0.13278\n",
            "[806]\tvalid_0's auc: 0.845586\tvalid_0's binary_logloss: 0.132779\n",
            "[807]\tvalid_0's auc: 0.845595\tvalid_0's binary_logloss: 0.132775\n",
            "[808]\tvalid_0's auc: 0.845604\tvalid_0's binary_logloss: 0.13277\n",
            "[809]\tvalid_0's auc: 0.845586\tvalid_0's binary_logloss: 0.132771\n",
            "[810]\tvalid_0's auc: 0.845576\tvalid_0's binary_logloss: 0.132773\n",
            "[811]\tvalid_0's auc: 0.845575\tvalid_0's binary_logloss: 0.132773\n",
            "[812]\tvalid_0's auc: 0.845567\tvalid_0's binary_logloss: 0.132775\n",
            "[813]\tvalid_0's auc: 0.845546\tvalid_0's binary_logloss: 0.132776\n",
            "[814]\tvalid_0's auc: 0.845504\tvalid_0's binary_logloss: 0.132778\n",
            "[815]\tvalid_0's auc: 0.845482\tvalid_0's binary_logloss: 0.132781\n",
            "[816]\tvalid_0's auc: 0.845506\tvalid_0's binary_logloss: 0.132776\n",
            "[817]\tvalid_0's auc: 0.845471\tvalid_0's binary_logloss: 0.132782\n",
            "[818]\tvalid_0's auc: 0.845463\tvalid_0's binary_logloss: 0.132783\n",
            "[819]\tvalid_0's auc: 0.845458\tvalid_0's binary_logloss: 0.132784\n",
            "[820]\tvalid_0's auc: 0.845467\tvalid_0's binary_logloss: 0.132782\n",
            "[821]\tvalid_0's auc: 0.845466\tvalid_0's binary_logloss: 0.132779\n",
            "[822]\tvalid_0's auc: 0.845498\tvalid_0's binary_logloss: 0.132778\n",
            "[823]\tvalid_0's auc: 0.845484\tvalid_0's binary_logloss: 0.132783\n",
            "[824]\tvalid_0's auc: 0.845511\tvalid_0's binary_logloss: 0.132773\n",
            "[825]\tvalid_0's auc: 0.84552\tvalid_0's binary_logloss: 0.13277\n",
            "[826]\tvalid_0's auc: 0.845532\tvalid_0's binary_logloss: 0.132769\n",
            "[827]\tvalid_0's auc: 0.845508\tvalid_0's binary_logloss: 0.132773\n",
            "[828]\tvalid_0's auc: 0.845505\tvalid_0's binary_logloss: 0.132772\n",
            "[829]\tvalid_0's auc: 0.845531\tvalid_0's binary_logloss: 0.132767\n",
            "[830]\tvalid_0's auc: 0.84553\tvalid_0's binary_logloss: 0.132767\n",
            "[831]\tvalid_0's auc: 0.8455\tvalid_0's binary_logloss: 0.132771\n",
            "[832]\tvalid_0's auc: 0.845476\tvalid_0's binary_logloss: 0.132779\n",
            "[833]\tvalid_0's auc: 0.845439\tvalid_0's binary_logloss: 0.132786\n",
            "[834]\tvalid_0's auc: 0.845446\tvalid_0's binary_logloss: 0.132781\n",
            "[835]\tvalid_0's auc: 0.84544\tvalid_0's binary_logloss: 0.13278\n",
            "[836]\tvalid_0's auc: 0.845458\tvalid_0's binary_logloss: 0.132776\n",
            "[837]\tvalid_0's auc: 0.845468\tvalid_0's binary_logloss: 0.13278\n",
            "[838]\tvalid_0's auc: 0.845447\tvalid_0's binary_logloss: 0.132782\n",
            "[839]\tvalid_0's auc: 0.845442\tvalid_0's binary_logloss: 0.132789\n",
            "[840]\tvalid_0's auc: 0.84544\tvalid_0's binary_logloss: 0.132787\n",
            "[841]\tvalid_0's auc: 0.845424\tvalid_0's binary_logloss: 0.132786\n",
            "[842]\tvalid_0's auc: 0.845429\tvalid_0's binary_logloss: 0.132786\n",
            "[843]\tvalid_0's auc: 0.84538\tvalid_0's binary_logloss: 0.13279\n",
            "[844]\tvalid_0's auc: 0.845365\tvalid_0's binary_logloss: 0.132794\n",
            "[845]\tvalid_0's auc: 0.845328\tvalid_0's binary_logloss: 0.132801\n",
            "[846]\tvalid_0's auc: 0.845292\tvalid_0's binary_logloss: 0.132805\n",
            "[847]\tvalid_0's auc: 0.845326\tvalid_0's binary_logloss: 0.132804\n",
            "[848]\tvalid_0's auc: 0.845328\tvalid_0's binary_logloss: 0.132799\n",
            "[849]\tvalid_0's auc: 0.845359\tvalid_0's binary_logloss: 0.132798\n",
            "[850]\tvalid_0's auc: 0.845364\tvalid_0's binary_logloss: 0.132797\n",
            "[851]\tvalid_0's auc: 0.845378\tvalid_0's binary_logloss: 0.132795\n",
            "[852]\tvalid_0's auc: 0.845334\tvalid_0's binary_logloss: 0.132801\n",
            "[853]\tvalid_0's auc: 0.845346\tvalid_0's binary_logloss: 0.132799\n",
            "[854]\tvalid_0's auc: 0.845347\tvalid_0's binary_logloss: 0.132799\n",
            "[855]\tvalid_0's auc: 0.845356\tvalid_0's binary_logloss: 0.132798\n",
            "[856]\tvalid_0's auc: 0.845325\tvalid_0's binary_logloss: 0.132801\n",
            "[857]\tvalid_0's auc: 0.845325\tvalid_0's binary_logloss: 0.1328\n",
            "[858]\tvalid_0's auc: 0.845351\tvalid_0's binary_logloss: 0.132792\n",
            "[859]\tvalid_0's auc: 0.845368\tvalid_0's binary_logloss: 0.132789\n",
            "[860]\tvalid_0's auc: 0.845374\tvalid_0's binary_logloss: 0.132793\n",
            "[861]\tvalid_0's auc: 0.845392\tvalid_0's binary_logloss: 0.132787\n",
            "[862]\tvalid_0's auc: 0.845405\tvalid_0's binary_logloss: 0.132785\n",
            "[863]\tvalid_0's auc: 0.845389\tvalid_0's binary_logloss: 0.132788\n",
            "[864]\tvalid_0's auc: 0.845472\tvalid_0's binary_logloss: 0.13278\n",
            "[865]\tvalid_0's auc: 0.845491\tvalid_0's binary_logloss: 0.132779\n",
            "[866]\tvalid_0's auc: 0.845461\tvalid_0's binary_logloss: 0.132782\n",
            "[867]\tvalid_0's auc: 0.845464\tvalid_0's binary_logloss: 0.132781\n",
            "[868]\tvalid_0's auc: 0.845491\tvalid_0's binary_logloss: 0.13278\n",
            "[869]\tvalid_0's auc: 0.845501\tvalid_0's binary_logloss: 0.132778\n",
            "[870]\tvalid_0's auc: 0.845542\tvalid_0's binary_logloss: 0.132774\n",
            "[871]\tvalid_0's auc: 0.845517\tvalid_0's binary_logloss: 0.132776\n",
            "[872]\tvalid_0's auc: 0.845508\tvalid_0's binary_logloss: 0.132777\n",
            "[873]\tvalid_0's auc: 0.845501\tvalid_0's binary_logloss: 0.132779\n",
            "[874]\tvalid_0's auc: 0.845494\tvalid_0's binary_logloss: 0.132781\n",
            "[875]\tvalid_0's auc: 0.845497\tvalid_0's binary_logloss: 0.132778\n",
            "[876]\tvalid_0's auc: 0.845497\tvalid_0's binary_logloss: 0.132776\n",
            "[877]\tvalid_0's auc: 0.845484\tvalid_0's binary_logloss: 0.132775\n",
            "[878]\tvalid_0's auc: 0.84547\tvalid_0's binary_logloss: 0.132779\n",
            "[879]\tvalid_0's auc: 0.845465\tvalid_0's binary_logloss: 0.132783\n",
            "[880]\tvalid_0's auc: 0.845441\tvalid_0's binary_logloss: 0.132785\n",
            "[881]\tvalid_0's auc: 0.845434\tvalid_0's binary_logloss: 0.132785\n",
            "[882]\tvalid_0's auc: 0.845463\tvalid_0's binary_logloss: 0.132779\n",
            "[883]\tvalid_0's auc: 0.845476\tvalid_0's binary_logloss: 0.132778\n",
            "[884]\tvalid_0's auc: 0.845484\tvalid_0's binary_logloss: 0.132774\n",
            "[885]\tvalid_0's auc: 0.84549\tvalid_0's binary_logloss: 0.132775\n",
            "[886]\tvalid_0's auc: 0.845505\tvalid_0's binary_logloss: 0.132772\n",
            "[887]\tvalid_0's auc: 0.845464\tvalid_0's binary_logloss: 0.132776\n",
            "[888]\tvalid_0's auc: 0.845447\tvalid_0's binary_logloss: 0.132775\n",
            "[889]\tvalid_0's auc: 0.845423\tvalid_0's binary_logloss: 0.132781\n",
            "[890]\tvalid_0's auc: 0.845395\tvalid_0's binary_logloss: 0.132787\n",
            "[891]\tvalid_0's auc: 0.845338\tvalid_0's binary_logloss: 0.132791\n",
            "[892]\tvalid_0's auc: 0.845298\tvalid_0's binary_logloss: 0.132795\n",
            "[893]\tvalid_0's auc: 0.845278\tvalid_0's binary_logloss: 0.132794\n",
            "[894]\tvalid_0's auc: 0.8453\tvalid_0's binary_logloss: 0.132796\n",
            "[895]\tvalid_0's auc: 0.845273\tvalid_0's binary_logloss: 0.132797\n",
            "[896]\tvalid_0's auc: 0.845243\tvalid_0's binary_logloss: 0.132799\n",
            "[897]\tvalid_0's auc: 0.845296\tvalid_0's binary_logloss: 0.132788\n",
            "[898]\tvalid_0's auc: 0.845303\tvalid_0's binary_logloss: 0.132789\n",
            "[899]\tvalid_0's auc: 0.845249\tvalid_0's binary_logloss: 0.132792\n",
            "[900]\tvalid_0's auc: 0.845282\tvalid_0's binary_logloss: 0.132789\n",
            "[901]\tvalid_0's auc: 0.845333\tvalid_0's binary_logloss: 0.132785\n",
            "[902]\tvalid_0's auc: 0.845304\tvalid_0's binary_logloss: 0.132791\n",
            "[903]\tvalid_0's auc: 0.845309\tvalid_0's binary_logloss: 0.13279\n",
            "[904]\tvalid_0's auc: 0.845318\tvalid_0's binary_logloss: 0.132787\n",
            "[905]\tvalid_0's auc: 0.845316\tvalid_0's binary_logloss: 0.132788\n",
            "[906]\tvalid_0's auc: 0.845318\tvalid_0's binary_logloss: 0.132791\n",
            "[907]\tvalid_0's auc: 0.845306\tvalid_0's binary_logloss: 0.132791\n",
            "[908]\tvalid_0's auc: 0.845315\tvalid_0's binary_logloss: 0.132789\n",
            "[909]\tvalid_0's auc: 0.845321\tvalid_0's binary_logloss: 0.132786\n",
            "[910]\tvalid_0's auc: 0.845312\tvalid_0's binary_logloss: 0.132788\n",
            "[911]\tvalid_0's auc: 0.845304\tvalid_0's binary_logloss: 0.132788\n",
            "[912]\tvalid_0's auc: 0.845316\tvalid_0's binary_logloss: 0.132789\n",
            "[913]\tvalid_0's auc: 0.845309\tvalid_0's binary_logloss: 0.132793\n",
            "[914]\tvalid_0's auc: 0.845285\tvalid_0's binary_logloss: 0.132798\n",
            "[915]\tvalid_0's auc: 0.845249\tvalid_0's binary_logloss: 0.132801\n",
            "[916]\tvalid_0's auc: 0.845259\tvalid_0's binary_logloss: 0.1328\n",
            "[917]\tvalid_0's auc: 0.845229\tvalid_0's binary_logloss: 0.132805\n",
            "[918]\tvalid_0's auc: 0.845261\tvalid_0's binary_logloss: 0.132798\n",
            "[919]\tvalid_0's auc: 0.84528\tvalid_0's binary_logloss: 0.132796\n",
            "[920]\tvalid_0's auc: 0.845272\tvalid_0's binary_logloss: 0.132799\n",
            "[921]\tvalid_0's auc: 0.845257\tvalid_0's binary_logloss: 0.132797\n",
            "[922]\tvalid_0's auc: 0.845242\tvalid_0's binary_logloss: 0.132801\n",
            "[923]\tvalid_0's auc: 0.845241\tvalid_0's binary_logloss: 0.132802\n",
            "[924]\tvalid_0's auc: 0.845222\tvalid_0's binary_logloss: 0.132807\n",
            "[925]\tvalid_0's auc: 0.845223\tvalid_0's binary_logloss: 0.132805\n",
            "[926]\tvalid_0's auc: 0.845233\tvalid_0's binary_logloss: 0.132805\n",
            "[927]\tvalid_0's auc: 0.845253\tvalid_0's binary_logloss: 0.132801\n",
            "[928]\tvalid_0's auc: 0.845306\tvalid_0's binary_logloss: 0.132796\n",
            "[929]\tvalid_0's auc: 0.845301\tvalid_0's binary_logloss: 0.132797\n",
            "[930]\tvalid_0's auc: 0.845286\tvalid_0's binary_logloss: 0.132801\n",
            "[931]\tvalid_0's auc: 0.845264\tvalid_0's binary_logloss: 0.132799\n",
            "[932]\tvalid_0's auc: 0.845244\tvalid_0's binary_logloss: 0.132802\n",
            "[933]\tvalid_0's auc: 0.845211\tvalid_0's binary_logloss: 0.132807\n",
            "[934]\tvalid_0's auc: 0.845223\tvalid_0's binary_logloss: 0.132808\n",
            "[935]\tvalid_0's auc: 0.845209\tvalid_0's binary_logloss: 0.13281\n",
            "[936]\tvalid_0's auc: 0.845213\tvalid_0's binary_logloss: 0.132809\n",
            "[937]\tvalid_0's auc: 0.845245\tvalid_0's binary_logloss: 0.132804\n",
            "[938]\tvalid_0's auc: 0.845244\tvalid_0's binary_logloss: 0.132804\n",
            "[939]\tvalid_0's auc: 0.845244\tvalid_0's binary_logloss: 0.132804\n",
            "[940]\tvalid_0's auc: 0.845284\tvalid_0's binary_logloss: 0.132797\n",
            "[941]\tvalid_0's auc: 0.845273\tvalid_0's binary_logloss: 0.132799\n",
            "[942]\tvalid_0's auc: 0.845276\tvalid_0's binary_logloss: 0.1328\n",
            "[943]\tvalid_0's auc: 0.845261\tvalid_0's binary_logloss: 0.132803\n",
            "[944]\tvalid_0's auc: 0.845257\tvalid_0's binary_logloss: 0.132807\n",
            "[945]\tvalid_0's auc: 0.845237\tvalid_0's binary_logloss: 0.13281\n",
            "[946]\tvalid_0's auc: 0.845236\tvalid_0's binary_logloss: 0.132804\n",
            "[947]\tvalid_0's auc: 0.845211\tvalid_0's binary_logloss: 0.132813\n",
            "[948]\tvalid_0's auc: 0.845183\tvalid_0's binary_logloss: 0.132814\n",
            "[949]\tvalid_0's auc: 0.8452\tvalid_0's binary_logloss: 0.13281\n",
            "[950]\tvalid_0's auc: 0.845171\tvalid_0's binary_logloss: 0.132813\n",
            "[951]\tvalid_0's auc: 0.845174\tvalid_0's binary_logloss: 0.132812\n",
            "[952]\tvalid_0's auc: 0.845175\tvalid_0's binary_logloss: 0.132812\n",
            "[953]\tvalid_0's auc: 0.84517\tvalid_0's binary_logloss: 0.132813\n",
            "[954]\tvalid_0's auc: 0.845165\tvalid_0's binary_logloss: 0.132813\n",
            "[955]\tvalid_0's auc: 0.845149\tvalid_0's binary_logloss: 0.132816\n",
            "[956]\tvalid_0's auc: 0.845128\tvalid_0's binary_logloss: 0.132822\n",
            "[957]\tvalid_0's auc: 0.845173\tvalid_0's binary_logloss: 0.13282\n",
            "[958]\tvalid_0's auc: 0.845156\tvalid_0's binary_logloss: 0.132823\n",
            "[959]\tvalid_0's auc: 0.845156\tvalid_0's binary_logloss: 0.132823\n",
            "[960]\tvalid_0's auc: 0.845152\tvalid_0's binary_logloss: 0.132823\n",
            "[961]\tvalid_0's auc: 0.845162\tvalid_0's binary_logloss: 0.132825\n",
            "[962]\tvalid_0's auc: 0.845154\tvalid_0's binary_logloss: 0.132828\n",
            "[963]\tvalid_0's auc: 0.845137\tvalid_0's binary_logloss: 0.132829\n",
            "[964]\tvalid_0's auc: 0.845136\tvalid_0's binary_logloss: 0.132828\n",
            "[965]\tvalid_0's auc: 0.845133\tvalid_0's binary_logloss: 0.13283\n",
            "[966]\tvalid_0's auc: 0.8451\tvalid_0's binary_logloss: 0.132833\n",
            "[967]\tvalid_0's auc: 0.845089\tvalid_0's binary_logloss: 0.132835\n",
            "[968]\tvalid_0's auc: 0.845097\tvalid_0's binary_logloss: 0.132834\n",
            "[969]\tvalid_0's auc: 0.845079\tvalid_0's binary_logloss: 0.132835\n",
            "[970]\tvalid_0's auc: 0.845073\tvalid_0's binary_logloss: 0.132839\n",
            "[971]\tvalid_0's auc: 0.845075\tvalid_0's binary_logloss: 0.132834\n",
            "[972]\tvalid_0's auc: 0.845039\tvalid_0's binary_logloss: 0.132838\n",
            "[973]\tvalid_0's auc: 0.844976\tvalid_0's binary_logloss: 0.132848\n",
            "[974]\tvalid_0's auc: 0.844976\tvalid_0's binary_logloss: 0.132852\n",
            "[975]\tvalid_0's auc: 0.844993\tvalid_0's binary_logloss: 0.132843\n",
            "[976]\tvalid_0's auc: 0.845061\tvalid_0's binary_logloss: 0.132837\n",
            "[977]\tvalid_0's auc: 0.845056\tvalid_0's binary_logloss: 0.132838\n",
            "[978]\tvalid_0's auc: 0.845079\tvalid_0's binary_logloss: 0.132836\n",
            "[979]\tvalid_0's auc: 0.845068\tvalid_0's binary_logloss: 0.13284\n",
            "[980]\tvalid_0's auc: 0.845047\tvalid_0's binary_logloss: 0.132847\n",
            "[981]\tvalid_0's auc: 0.845072\tvalid_0's binary_logloss: 0.132842\n",
            "[982]\tvalid_0's auc: 0.845055\tvalid_0's binary_logloss: 0.132843\n",
            "[983]\tvalid_0's auc: 0.845047\tvalid_0's binary_logloss: 0.132844\n",
            "[984]\tvalid_0's auc: 0.845064\tvalid_0's binary_logloss: 0.132842\n",
            "[985]\tvalid_0's auc: 0.845038\tvalid_0's binary_logloss: 0.132847\n",
            "[986]\tvalid_0's auc: 0.845028\tvalid_0's binary_logloss: 0.132852\n",
            "[987]\tvalid_0's auc: 0.845034\tvalid_0's binary_logloss: 0.132852\n",
            "[988]\tvalid_0's auc: 0.845058\tvalid_0's binary_logloss: 0.132851\n",
            "[989]\tvalid_0's auc: 0.845052\tvalid_0's binary_logloss: 0.132853\n",
            "[990]\tvalid_0's auc: 0.845039\tvalid_0's binary_logloss: 0.132855\n",
            "[991]\tvalid_0's auc: 0.845031\tvalid_0's binary_logloss: 0.132858\n",
            "[992]\tvalid_0's auc: 0.845034\tvalid_0's binary_logloss: 0.132855\n",
            "[993]\tvalid_0's auc: 0.845038\tvalid_0's binary_logloss: 0.132853\n",
            "[994]\tvalid_0's auc: 0.845038\tvalid_0's binary_logloss: 0.132852\n",
            "[995]\tvalid_0's auc: 0.845013\tvalid_0's binary_logloss: 0.132857\n",
            "[996]\tvalid_0's auc: 0.844989\tvalid_0's binary_logloss: 0.132858\n",
            "[997]\tvalid_0's auc: 0.844967\tvalid_0's binary_logloss: 0.132862\n",
            "[998]\tvalid_0's auc: 0.844994\tvalid_0's binary_logloss: 0.132854\n",
            "[999]\tvalid_0's auc: 0.844991\tvalid_0's binary_logloss: 0.13286\n",
            "[1000]\tvalid_0's auc: 0.844955\tvalid_0's binary_logloss: 0.132863\n",
            "CPU times: user 1min 53s, sys: 482 ms, total: 1min 53s\n",
            "Wall time: 59.5 s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LGBMClassifier(boosting_type='gbdt', class_weight=None, colsample_bytree=0.5,\n",
              "               importance_type='split', learning_rate=0.01, max_depth=7,\n",
              "               min_child_samples=20, min_child_weight=0.001, min_split_gain=0.0,\n",
              "               n_estimators=1000, n_jobs=-1, nthread=-1, num_leaves=31,\n",
              "               objective='binary', random_state=42, reg_alpha=0.3,\n",
              "               reg_lambda=0.0, scale_pos_weight=1.0, silent=True, subsample=0.8,\n",
              "               subsample_for_bin=200000, subsample_freq=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 86
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "k_CEadVqLzvy",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "2f340968-0980-4e72-cd87-695f0e18160b"
      },
      "source": [
        "y_pred = model.predict_proba(X_val)[:,1]\n",
        "label=\"LightGBM Classifer\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "i+=1\n",
        "result = plot_auc(y_val,y_pred,label)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t LightGBM Classifer on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.133 and 0.845 for LightGBM Classifer on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "BF22prvpLzv1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "a4eecf99-d0ad-4df9-f5ce-0a531119ebe7"
      },
      "source": [
        "result_df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137490</td>\n",
              "      <td>0.832512</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>XGBoost Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.133102</td>\n",
              "      <td>0.845507</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>LightGBM Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.132863</td>\n",
              "      <td>0.844955</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                      Model Dataset  log_loss       auc\n",
              "0        Logisic Regression  Normal  0.542129  0.812101\n",
              "1  Decision Tree Classifier  Normal  0.148082  0.779135\n",
              "2  Random Forest Classifier  Normal  0.137490  0.832512\n",
              "3         XGBoost Classifer  Normal  0.133102  0.845507\n",
              "4        LightGBM Classifer  Normal  0.132863  0.844955"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "LECmbx42-V2m"
      },
      "source": [
        "# Random Forest (top 250)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CUgv74lV-V2n",
        "colab": {}
      },
      "source": [
        "model = ensemble.RandomForestClassifier(class_weight='balanced')\n",
        "params = { 'n_estimators':[500,1000],\n",
        "          'max_depth':[250,500], \n",
        "         'min_samples_split':[100,500],\n",
        "          'min_samples_leaf':[5,7],\n",
        "          'max_leaf_nodes':[100,250] \n",
        "          }\n"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "eCt8L0AV-V2q",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "outputId": "21a20044-fae8-4b6b-ad41-50295a03ed7f"
      },
      "source": [
        "%%time\n",
        "find_best_params(model,params,cv=3,X_train=X_train_200)"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 3 folds for each of 10 candidates, totalling 30 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done  30 out of  30 | elapsed: 32.4min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "The best auc score was 0.818\n",
            "The best params were: {'n_estimators': 1000, 'min_samples_split': 500, 'min_samples_leaf': 5, 'max_leaf_nodes': 250, 'max_depth': 500}\n",
            "CPU times: user 2min 54s, sys: 530 ms, total: 2min 54s\n",
            "Wall time: 35min 20s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight='balanced',\n",
              "                       criterion='gini', max_depth=500, max_features='auto',\n",
              "                       max_leaf_nodes=250, max_samples=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=5, min_samples_split=500,\n",
              "                       min_weight_fraction_leaf=0.0, n_estimators=1000,\n",
              "                       n_jobs=None, oob_score=False, random_state=None,\n",
              "                       verbose=0, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "DVdFANOy-V2u",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        },
        "outputId": "d7081b60-a1ba-42ab-a2c7-734c367886fd"
      },
      "source": [
        "model = ensemble.RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight='balanced',\n",
        "                       criterion='gini', max_depth=500, max_features='auto',\n",
        "                       max_leaf_nodes=250, max_samples=None,\n",
        "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
        "                       min_samples_leaf=5, min_samples_split=500,\n",
        "                       min_weight_fraction_leaf=0.0, n_estimators=1000,\n",
        "                       n_jobs=-1, oob_score=False,random_state=42,\n",
        "                       verbose=0, warm_start=False)\n",
        "model.fit(X_train_200,y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight='balanced',\n",
              "                       criterion='gini', max_depth=500, max_features='auto',\n",
              "                       max_leaf_nodes=250, max_samples=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=5, min_samples_split=500,\n",
              "                       min_weight_fraction_leaf=0.0, n_estimators=1000,\n",
              "                       n_jobs=-1, oob_score=False, random_state=42, verbose=0,\n",
              "                       warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 91
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "okg4Xksh-V2y",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "91539041-2c24-4551-fb57-549faac39351"
      },
      "source": [
        "cc_model = calibration.CalibratedClassifierCV(model,cv='prefit')\n",
        "cc_model.fit(X_train_200,y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "CalibratedClassifierCV(base_estimator=RandomForestClassifier(bootstrap=True,\n",
              "                                                             ccp_alpha=0.0,\n",
              "                                                             class_weight='balanced',\n",
              "                                                             criterion='gini',\n",
              "                                                             max_depth=500,\n",
              "                                                             max_features='auto',\n",
              "                                                             max_leaf_nodes=250,\n",
              "                                                             max_samples=None,\n",
              "                                                             min_impurity_decrease=0.0,\n",
              "                                                             min_impurity_split=None,\n",
              "                                                             min_samples_leaf=5,\n",
              "                                                             min_samples_split=500,\n",
              "                                                             min_weight_fraction_leaf=0.0,\n",
              "                                                             n_estimators=1000,\n",
              "                                                             n_jobs=-1,\n",
              "                                                             oob_score=False,\n",
              "                                                             random_state=42,\n",
              "                                                             verbose=0,\n",
              "                                                             warm_start=False),\n",
              "                       cv='prefit', method='sigmoid')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "5eSsmef9-V21",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "64325714-3f66-4e86-de8b-b8ea4fec74e6"
      },
      "source": [
        "i+=1\n",
        "y_pred = cc_model.predict_proba(X_val_200)[:,1]\n",
        "label=\"Random Forest Classifier (top 250)\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "result = plot_auc(y_val,y_pred,label,dataset)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t Random Forest Classifier (top 250) on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.137 and 0.834 for Random Forest Classifier (top 250) on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "8Mo702Rt-V23",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "c79b049b-21a3-41e9-a45c-13aba3e3559a"
      },
      "source": [
        "result_df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137490</td>\n",
              "      <td>0.832512</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>XGBoost Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.133102</td>\n",
              "      <td>0.845507</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>LightGBM Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.132863</td>\n",
              "      <td>0.844955</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Random Forest Classifier (top 250)</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137054</td>\n",
              "      <td>0.834184</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                Model Dataset  log_loss       auc\n",
              "0                  Logisic Regression  Normal  0.542129  0.812101\n",
              "1            Decision Tree Classifier  Normal  0.148082  0.779135\n",
              "2            Random Forest Classifier  Normal  0.137490  0.832512\n",
              "3                   XGBoost Classifer  Normal  0.133102  0.845507\n",
              "4                  LightGBM Classifer  Normal  0.132863  0.844955\n",
              "5  Random Forest Classifier (top 250)  Normal  0.137054  0.834184"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 94
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "i-y4IONn-V26"
      },
      "source": [
        "# XGB Classifier (top 250)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Gjg5rhGa-V27",
        "colab": {}
      },
      "source": [
        "model = xgb.XGBClassifier(n_jobs=-1,\n",
        "                      nthread=-1,\n",
        "                      scale_pos_weight=1,\n",
        "                      learning_rate=0.01,  \n",
        "                      colsample_bytree = 0.8, \n",
        "                      subsample = 0.7,\n",
        "                      objective='binary:logistic', \n",
        "                      n_estimators=801,\n",
        "                      reg_alpha = 0.3,\n",
        "                      max_depth=6, \n",
        "                      gamma=0,random_state=42\n",
        "                     )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "8Q0dR9Kt-V2-",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "dc1ed2bc-8f52-4a6d-c41b-ec10c1587c84"
      },
      "source": [
        "#https://towardsdatascience.com/fine-tuning-xgboost-in-python-like-a-boss-b4543ed8b1e\n",
        "\"\"\"\n",
        "manually changed each parameters 1 by 1 and compared the validation auc\n",
        "and selected the best parameter value\n",
        "\"\"\"\n",
        "eval_metric = ['auc','error']\n",
        "%time model.fit(X_train_200,y_train,eval_set=[(X_val_200,y_val)],eval_metric=eval_metric,verbose=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0]\tvalidation_0-auc:0.787164\tvalidation_0-error:0.040954\n",
            "[1]\tvalidation_0-auc:0.790325\tvalidation_0-error:0.039902\n",
            "[2]\tvalidation_0-auc:0.793572\tvalidation_0-error:0.039902\n",
            "[3]\tvalidation_0-auc:0.810991\tvalidation_0-error:0.039551\n",
            "[4]\tvalidation_0-auc:0.815335\tvalidation_0-error:0.039551\n",
            "[5]\tvalidation_0-auc:0.820607\tvalidation_0-error:0.039551\n",
            "[6]\tvalidation_0-auc:0.822459\tvalidation_0-error:0.039551\n",
            "[7]\tvalidation_0-auc:0.825145\tvalidation_0-error:0.039551\n",
            "[8]\tvalidation_0-auc:0.825992\tvalidation_0-error:0.039551\n",
            "[9]\tvalidation_0-auc:0.826397\tvalidation_0-error:0.039551\n",
            "[10]\tvalidation_0-auc:0.826164\tvalidation_0-error:0.039551\n",
            "[11]\tvalidation_0-auc:0.824959\tvalidation_0-error:0.039551\n",
            "[12]\tvalidation_0-auc:0.826101\tvalidation_0-error:0.039551\n",
            "[13]\tvalidation_0-auc:0.824791\tvalidation_0-error:0.039551\n",
            "[14]\tvalidation_0-auc:0.824845\tvalidation_0-error:0.039551\n",
            "[15]\tvalidation_0-auc:0.826077\tvalidation_0-error:0.039551\n",
            "[16]\tvalidation_0-auc:0.826134\tvalidation_0-error:0.039551\n",
            "[17]\tvalidation_0-auc:0.825663\tvalidation_0-error:0.039551\n",
            "[18]\tvalidation_0-auc:0.826208\tvalidation_0-error:0.039551\n",
            "[19]\tvalidation_0-auc:0.826516\tvalidation_0-error:0.039551\n",
            "[20]\tvalidation_0-auc:0.827243\tvalidation_0-error:0.039551\n",
            "[21]\tvalidation_0-auc:0.827706\tvalidation_0-error:0.039551\n",
            "[22]\tvalidation_0-auc:0.826961\tvalidation_0-error:0.039551\n",
            "[23]\tvalidation_0-auc:0.827754\tvalidation_0-error:0.039551\n",
            "[24]\tvalidation_0-auc:0.827824\tvalidation_0-error:0.039551\n",
            "[25]\tvalidation_0-auc:0.828045\tvalidation_0-error:0.039551\n",
            "[26]\tvalidation_0-auc:0.828004\tvalidation_0-error:0.039551\n",
            "[27]\tvalidation_0-auc:0.828039\tvalidation_0-error:0.039551\n",
            "[28]\tvalidation_0-auc:0.828172\tvalidation_0-error:0.039551\n",
            "[29]\tvalidation_0-auc:0.827919\tvalidation_0-error:0.039551\n",
            "[30]\tvalidation_0-auc:0.828211\tvalidation_0-error:0.039551\n",
            "[31]\tvalidation_0-auc:0.828369\tvalidation_0-error:0.039551\n",
            "[32]\tvalidation_0-auc:0.828329\tvalidation_0-error:0.039551\n",
            "[33]\tvalidation_0-auc:0.828576\tvalidation_0-error:0.039551\n",
            "[34]\tvalidation_0-auc:0.828248\tvalidation_0-error:0.039551\n",
            "[35]\tvalidation_0-auc:0.828495\tvalidation_0-error:0.039551\n",
            "[36]\tvalidation_0-auc:0.828561\tvalidation_0-error:0.039551\n",
            "[37]\tvalidation_0-auc:0.828863\tvalidation_0-error:0.039551\n",
            "[38]\tvalidation_0-auc:0.828933\tvalidation_0-error:0.039551\n",
            "[39]\tvalidation_0-auc:0.828831\tvalidation_0-error:0.039551\n",
            "[40]\tvalidation_0-auc:0.829087\tvalidation_0-error:0.039551\n",
            "[41]\tvalidation_0-auc:0.829247\tvalidation_0-error:0.039551\n",
            "[42]\tvalidation_0-auc:0.830801\tvalidation_0-error:0.039551\n",
            "[43]\tvalidation_0-auc:0.830772\tvalidation_0-error:0.039551\n",
            "[44]\tvalidation_0-auc:0.830809\tvalidation_0-error:0.039551\n",
            "[45]\tvalidation_0-auc:0.829925\tvalidation_0-error:0.039551\n",
            "[46]\tvalidation_0-auc:0.829717\tvalidation_0-error:0.039551\n",
            "[47]\tvalidation_0-auc:0.829704\tvalidation_0-error:0.039551\n",
            "[48]\tvalidation_0-auc:0.829969\tvalidation_0-error:0.039551\n",
            "[49]\tvalidation_0-auc:0.830106\tvalidation_0-error:0.039551\n",
            "[50]\tvalidation_0-auc:0.831253\tvalidation_0-error:0.039551\n",
            "[51]\tvalidation_0-auc:0.831002\tvalidation_0-error:0.039551\n",
            "[52]\tvalidation_0-auc:0.831249\tvalidation_0-error:0.039551\n",
            "[53]\tvalidation_0-auc:0.83154\tvalidation_0-error:0.039551\n",
            "[54]\tvalidation_0-auc:0.831511\tvalidation_0-error:0.039551\n",
            "[55]\tvalidation_0-auc:0.831838\tvalidation_0-error:0.039551\n",
            "[56]\tvalidation_0-auc:0.831517\tvalidation_0-error:0.039551\n",
            "[57]\tvalidation_0-auc:0.831344\tvalidation_0-error:0.039551\n",
            "[58]\tvalidation_0-auc:0.831375\tvalidation_0-error:0.039551\n",
            "[59]\tvalidation_0-auc:0.831459\tvalidation_0-error:0.039551\n",
            "[60]\tvalidation_0-auc:0.831136\tvalidation_0-error:0.039551\n",
            "[61]\tvalidation_0-auc:0.831095\tvalidation_0-error:0.039551\n",
            "[62]\tvalidation_0-auc:0.831105\tvalidation_0-error:0.039551\n",
            "[63]\tvalidation_0-auc:0.831256\tvalidation_0-error:0.039551\n",
            "[64]\tvalidation_0-auc:0.831447\tvalidation_0-error:0.039551\n",
            "[65]\tvalidation_0-auc:0.831363\tvalidation_0-error:0.039551\n",
            "[66]\tvalidation_0-auc:0.831269\tvalidation_0-error:0.039551\n",
            "[67]\tvalidation_0-auc:0.832113\tvalidation_0-error:0.039551\n",
            "[68]\tvalidation_0-auc:0.832062\tvalidation_0-error:0.039551\n",
            "[69]\tvalidation_0-auc:0.832078\tvalidation_0-error:0.039551\n",
            "[70]\tvalidation_0-auc:0.832014\tvalidation_0-error:0.039551\n",
            "[71]\tvalidation_0-auc:0.832314\tvalidation_0-error:0.039551\n",
            "[72]\tvalidation_0-auc:0.832369\tvalidation_0-error:0.039551\n",
            "[73]\tvalidation_0-auc:0.832447\tvalidation_0-error:0.039551\n",
            "[74]\tvalidation_0-auc:0.832657\tvalidation_0-error:0.039551\n",
            "[75]\tvalidation_0-auc:0.832597\tvalidation_0-error:0.039551\n",
            "[76]\tvalidation_0-auc:0.832673\tvalidation_0-error:0.039551\n",
            "[77]\tvalidation_0-auc:0.832729\tvalidation_0-error:0.039551\n",
            "[78]\tvalidation_0-auc:0.832589\tvalidation_0-error:0.039551\n",
            "[79]\tvalidation_0-auc:0.832608\tvalidation_0-error:0.039551\n",
            "[80]\tvalidation_0-auc:0.832875\tvalidation_0-error:0.039551\n",
            "[81]\tvalidation_0-auc:0.833004\tvalidation_0-error:0.039551\n",
            "[82]\tvalidation_0-auc:0.83307\tvalidation_0-error:0.039551\n",
            "[83]\tvalidation_0-auc:0.83308\tvalidation_0-error:0.039551\n",
            "[84]\tvalidation_0-auc:0.833018\tvalidation_0-error:0.039551\n",
            "[85]\tvalidation_0-auc:0.833041\tvalidation_0-error:0.039551\n",
            "[86]\tvalidation_0-auc:0.833204\tvalidation_0-error:0.039551\n",
            "[87]\tvalidation_0-auc:0.833167\tvalidation_0-error:0.039551\n",
            "[88]\tvalidation_0-auc:0.833323\tvalidation_0-error:0.039551\n",
            "[89]\tvalidation_0-auc:0.833261\tvalidation_0-error:0.039551\n",
            "[90]\tvalidation_0-auc:0.833395\tvalidation_0-error:0.039551\n",
            "[91]\tvalidation_0-auc:0.83361\tvalidation_0-error:0.039551\n",
            "[92]\tvalidation_0-auc:0.833744\tvalidation_0-error:0.039551\n",
            "[93]\tvalidation_0-auc:0.833927\tvalidation_0-error:0.039551\n",
            "[94]\tvalidation_0-auc:0.833814\tvalidation_0-error:0.039551\n",
            "[95]\tvalidation_0-auc:0.833974\tvalidation_0-error:0.039551\n",
            "[96]\tvalidation_0-auc:0.834038\tvalidation_0-error:0.039551\n",
            "[97]\tvalidation_0-auc:0.834196\tvalidation_0-error:0.039551\n",
            "[98]\tvalidation_0-auc:0.834358\tvalidation_0-error:0.039551\n",
            "[99]\tvalidation_0-auc:0.834321\tvalidation_0-error:0.039551\n",
            "[100]\tvalidation_0-auc:0.834465\tvalidation_0-error:0.039551\n",
            "[101]\tvalidation_0-auc:0.834464\tvalidation_0-error:0.039551\n",
            "[102]\tvalidation_0-auc:0.834479\tvalidation_0-error:0.039551\n",
            "[103]\tvalidation_0-auc:0.834526\tvalidation_0-error:0.039551\n",
            "[104]\tvalidation_0-auc:0.834495\tvalidation_0-error:0.039551\n",
            "[105]\tvalidation_0-auc:0.834712\tvalidation_0-error:0.039551\n",
            "[106]\tvalidation_0-auc:0.834791\tvalidation_0-error:0.039551\n",
            "[107]\tvalidation_0-auc:0.834939\tvalidation_0-error:0.039551\n",
            "[108]\tvalidation_0-auc:0.834832\tvalidation_0-error:0.039551\n",
            "[109]\tvalidation_0-auc:0.834829\tvalidation_0-error:0.039551\n",
            "[110]\tvalidation_0-auc:0.834791\tvalidation_0-error:0.039551\n",
            "[111]\tvalidation_0-auc:0.834756\tvalidation_0-error:0.039551\n",
            "[112]\tvalidation_0-auc:0.834777\tvalidation_0-error:0.039551\n",
            "[113]\tvalidation_0-auc:0.834853\tvalidation_0-error:0.039551\n",
            "[114]\tvalidation_0-auc:0.83494\tvalidation_0-error:0.039551\n",
            "[115]\tvalidation_0-auc:0.834963\tvalidation_0-error:0.039551\n",
            "[116]\tvalidation_0-auc:0.834918\tvalidation_0-error:0.039551\n",
            "[117]\tvalidation_0-auc:0.834973\tvalidation_0-error:0.039551\n",
            "[118]\tvalidation_0-auc:0.835033\tvalidation_0-error:0.039551\n",
            "[119]\tvalidation_0-auc:0.834937\tvalidation_0-error:0.039551\n",
            "[120]\tvalidation_0-auc:0.835151\tvalidation_0-error:0.039551\n",
            "[121]\tvalidation_0-auc:0.835141\tvalidation_0-error:0.039551\n",
            "[122]\tvalidation_0-auc:0.835143\tvalidation_0-error:0.039551\n",
            "[123]\tvalidation_0-auc:0.835109\tvalidation_0-error:0.039551\n",
            "[124]\tvalidation_0-auc:0.835176\tvalidation_0-error:0.039551\n",
            "[125]\tvalidation_0-auc:0.835188\tvalidation_0-error:0.039551\n",
            "[126]\tvalidation_0-auc:0.835209\tvalidation_0-error:0.039551\n",
            "[127]\tvalidation_0-auc:0.835366\tvalidation_0-error:0.039551\n",
            "[128]\tvalidation_0-auc:0.835347\tvalidation_0-error:0.039551\n",
            "[129]\tvalidation_0-auc:0.835457\tvalidation_0-error:0.039551\n",
            "[130]\tvalidation_0-auc:0.835473\tvalidation_0-error:0.039551\n",
            "[131]\tvalidation_0-auc:0.835608\tvalidation_0-error:0.039551\n",
            "[132]\tvalidation_0-auc:0.835651\tvalidation_0-error:0.039551\n",
            "[133]\tvalidation_0-auc:0.835678\tvalidation_0-error:0.039551\n",
            "[134]\tvalidation_0-auc:0.835751\tvalidation_0-error:0.039551\n",
            "[135]\tvalidation_0-auc:0.835746\tvalidation_0-error:0.039551\n",
            "[136]\tvalidation_0-auc:0.835763\tvalidation_0-error:0.039551\n",
            "[137]\tvalidation_0-auc:0.83596\tvalidation_0-error:0.039551\n",
            "[138]\tvalidation_0-auc:0.835902\tvalidation_0-error:0.039551\n",
            "[139]\tvalidation_0-auc:0.835817\tvalidation_0-error:0.039551\n",
            "[140]\tvalidation_0-auc:0.835794\tvalidation_0-error:0.039551\n",
            "[141]\tvalidation_0-auc:0.835925\tvalidation_0-error:0.039551\n",
            "[142]\tvalidation_0-auc:0.835949\tvalidation_0-error:0.039551\n",
            "[143]\tvalidation_0-auc:0.835943\tvalidation_0-error:0.039551\n",
            "[144]\tvalidation_0-auc:0.836135\tvalidation_0-error:0.039551\n",
            "[145]\tvalidation_0-auc:0.836029\tvalidation_0-error:0.039551\n",
            "[146]\tvalidation_0-auc:0.836309\tvalidation_0-error:0.039551\n",
            "[147]\tvalidation_0-auc:0.836328\tvalidation_0-error:0.039551\n",
            "[148]\tvalidation_0-auc:0.836469\tvalidation_0-error:0.039551\n",
            "[149]\tvalidation_0-auc:0.83663\tvalidation_0-error:0.039551\n",
            "[150]\tvalidation_0-auc:0.836699\tvalidation_0-error:0.039551\n",
            "[151]\tvalidation_0-auc:0.836801\tvalidation_0-error:0.039551\n",
            "[152]\tvalidation_0-auc:0.836826\tvalidation_0-error:0.039551\n",
            "[153]\tvalidation_0-auc:0.836751\tvalidation_0-error:0.039551\n",
            "[154]\tvalidation_0-auc:0.836689\tvalidation_0-error:0.039551\n",
            "[155]\tvalidation_0-auc:0.836585\tvalidation_0-error:0.039551\n",
            "[156]\tvalidation_0-auc:0.836667\tvalidation_0-error:0.039551\n",
            "[157]\tvalidation_0-auc:0.836897\tvalidation_0-error:0.039551\n",
            "[158]\tvalidation_0-auc:0.836748\tvalidation_0-error:0.039551\n",
            "[159]\tvalidation_0-auc:0.836506\tvalidation_0-error:0.039551\n",
            "[160]\tvalidation_0-auc:0.83688\tvalidation_0-error:0.039551\n",
            "[161]\tvalidation_0-auc:0.836847\tvalidation_0-error:0.039551\n",
            "[162]\tvalidation_0-auc:0.836848\tvalidation_0-error:0.039551\n",
            "[163]\tvalidation_0-auc:0.836911\tvalidation_0-error:0.039551\n",
            "[164]\tvalidation_0-auc:0.837034\tvalidation_0-error:0.039551\n",
            "[165]\tvalidation_0-auc:0.837256\tvalidation_0-error:0.039551\n",
            "[166]\tvalidation_0-auc:0.837276\tvalidation_0-error:0.039551\n",
            "[167]\tvalidation_0-auc:0.837623\tvalidation_0-error:0.039551\n",
            "[168]\tvalidation_0-auc:0.837673\tvalidation_0-error:0.039551\n",
            "[169]\tvalidation_0-auc:0.837689\tvalidation_0-error:0.039551\n",
            "[170]\tvalidation_0-auc:0.837696\tvalidation_0-error:0.039551\n",
            "[171]\tvalidation_0-auc:0.8377\tvalidation_0-error:0.039551\n",
            "[172]\tvalidation_0-auc:0.837875\tvalidation_0-error:0.039551\n",
            "[173]\tvalidation_0-auc:0.837845\tvalidation_0-error:0.039551\n",
            "[174]\tvalidation_0-auc:0.837705\tvalidation_0-error:0.039551\n",
            "[175]\tvalidation_0-auc:0.837922\tvalidation_0-error:0.039551\n",
            "[176]\tvalidation_0-auc:0.838033\tvalidation_0-error:0.039551\n",
            "[177]\tvalidation_0-auc:0.838104\tvalidation_0-error:0.039551\n",
            "[178]\tvalidation_0-auc:0.838463\tvalidation_0-error:0.039551\n",
            "[179]\tvalidation_0-auc:0.838726\tvalidation_0-error:0.039551\n",
            "[180]\tvalidation_0-auc:0.83884\tvalidation_0-error:0.039551\n",
            "[181]\tvalidation_0-auc:0.838952\tvalidation_0-error:0.039551\n",
            "[182]\tvalidation_0-auc:0.839058\tvalidation_0-error:0.039551\n",
            "[183]\tvalidation_0-auc:0.838949\tvalidation_0-error:0.039551\n",
            "[184]\tvalidation_0-auc:0.839077\tvalidation_0-error:0.039551\n",
            "[185]\tvalidation_0-auc:0.83906\tvalidation_0-error:0.039551\n",
            "[186]\tvalidation_0-auc:0.839179\tvalidation_0-error:0.039551\n",
            "[187]\tvalidation_0-auc:0.839232\tvalidation_0-error:0.039551\n",
            "[188]\tvalidation_0-auc:0.839257\tvalidation_0-error:0.039551\n",
            "[189]\tvalidation_0-auc:0.839302\tvalidation_0-error:0.039551\n",
            "[190]\tvalidation_0-auc:0.83941\tvalidation_0-error:0.039551\n",
            "[191]\tvalidation_0-auc:0.839698\tvalidation_0-error:0.039551\n",
            "[192]\tvalidation_0-auc:0.83987\tvalidation_0-error:0.039551\n",
            "[193]\tvalidation_0-auc:0.839898\tvalidation_0-error:0.039551\n",
            "[194]\tvalidation_0-auc:0.839924\tvalidation_0-error:0.039551\n",
            "[195]\tvalidation_0-auc:0.840059\tvalidation_0-error:0.039551\n",
            "[196]\tvalidation_0-auc:0.84014\tvalidation_0-error:0.039551\n",
            "[197]\tvalidation_0-auc:0.840104\tvalidation_0-error:0.039551\n",
            "[198]\tvalidation_0-auc:0.840046\tvalidation_0-error:0.039551\n",
            "[199]\tvalidation_0-auc:0.840103\tvalidation_0-error:0.039551\n",
            "[200]\tvalidation_0-auc:0.84013\tvalidation_0-error:0.039551\n",
            "[201]\tvalidation_0-auc:0.840183\tvalidation_0-error:0.039551\n",
            "[202]\tvalidation_0-auc:0.840278\tvalidation_0-error:0.039551\n",
            "[203]\tvalidation_0-auc:0.840265\tvalidation_0-error:0.039551\n",
            "[204]\tvalidation_0-auc:0.840361\tvalidation_0-error:0.039551\n",
            "[205]\tvalidation_0-auc:0.840392\tvalidation_0-error:0.039551\n",
            "[206]\tvalidation_0-auc:0.840433\tvalidation_0-error:0.039551\n",
            "[207]\tvalidation_0-auc:0.840502\tvalidation_0-error:0.039551\n",
            "[208]\tvalidation_0-auc:0.84042\tvalidation_0-error:0.039551\n",
            "[209]\tvalidation_0-auc:0.840434\tvalidation_0-error:0.039551\n",
            "[210]\tvalidation_0-auc:0.840483\tvalidation_0-error:0.039551\n",
            "[211]\tvalidation_0-auc:0.840445\tvalidation_0-error:0.039551\n",
            "[212]\tvalidation_0-auc:0.84057\tvalidation_0-error:0.039551\n",
            "[213]\tvalidation_0-auc:0.840538\tvalidation_0-error:0.039551\n",
            "[214]\tvalidation_0-auc:0.840576\tvalidation_0-error:0.039551\n",
            "[215]\tvalidation_0-auc:0.840673\tvalidation_0-error:0.039551\n",
            "[216]\tvalidation_0-auc:0.840652\tvalidation_0-error:0.039551\n",
            "[217]\tvalidation_0-auc:0.840641\tvalidation_0-error:0.039551\n",
            "[218]\tvalidation_0-auc:0.840535\tvalidation_0-error:0.039551\n",
            "[219]\tvalidation_0-auc:0.840529\tvalidation_0-error:0.039551\n",
            "[220]\tvalidation_0-auc:0.840553\tvalidation_0-error:0.039551\n",
            "[221]\tvalidation_0-auc:0.840627\tvalidation_0-error:0.039551\n",
            "[222]\tvalidation_0-auc:0.840604\tvalidation_0-error:0.039551\n",
            "[223]\tvalidation_0-auc:0.84058\tvalidation_0-error:0.039551\n",
            "[224]\tvalidation_0-auc:0.840583\tvalidation_0-error:0.039551\n",
            "[225]\tvalidation_0-auc:0.840621\tvalidation_0-error:0.039551\n",
            "[226]\tvalidation_0-auc:0.840523\tvalidation_0-error:0.039551\n",
            "[227]\tvalidation_0-auc:0.840537\tvalidation_0-error:0.039551\n",
            "[228]\tvalidation_0-auc:0.840547\tvalidation_0-error:0.039551\n",
            "[229]\tvalidation_0-auc:0.840492\tvalidation_0-error:0.039551\n",
            "[230]\tvalidation_0-auc:0.840465\tvalidation_0-error:0.039551\n",
            "[231]\tvalidation_0-auc:0.840473\tvalidation_0-error:0.039551\n",
            "[232]\tvalidation_0-auc:0.840668\tvalidation_0-error:0.039551\n",
            "[233]\tvalidation_0-auc:0.840612\tvalidation_0-error:0.039551\n",
            "[234]\tvalidation_0-auc:0.840593\tvalidation_0-error:0.039551\n",
            "[235]\tvalidation_0-auc:0.840582\tvalidation_0-error:0.039551\n",
            "[236]\tvalidation_0-auc:0.840455\tvalidation_0-error:0.039551\n",
            "[237]\tvalidation_0-auc:0.840424\tvalidation_0-error:0.039551\n",
            "[238]\tvalidation_0-auc:0.840386\tvalidation_0-error:0.039551\n",
            "[239]\tvalidation_0-auc:0.840398\tvalidation_0-error:0.039551\n",
            "[240]\tvalidation_0-auc:0.840522\tvalidation_0-error:0.039551\n",
            "[241]\tvalidation_0-auc:0.840673\tvalidation_0-error:0.039551\n",
            "[242]\tvalidation_0-auc:0.840738\tvalidation_0-error:0.039551\n",
            "[243]\tvalidation_0-auc:0.840713\tvalidation_0-error:0.039551\n",
            "[244]\tvalidation_0-auc:0.840686\tvalidation_0-error:0.039551\n",
            "[245]\tvalidation_0-auc:0.840618\tvalidation_0-error:0.039551\n",
            "[246]\tvalidation_0-auc:0.840542\tvalidation_0-error:0.039551\n",
            "[247]\tvalidation_0-auc:0.840608\tvalidation_0-error:0.039551\n",
            "[248]\tvalidation_0-auc:0.840513\tvalidation_0-error:0.039551\n",
            "[249]\tvalidation_0-auc:0.840566\tvalidation_0-error:0.039551\n",
            "[250]\tvalidation_0-auc:0.840388\tvalidation_0-error:0.039551\n",
            "[251]\tvalidation_0-auc:0.84037\tvalidation_0-error:0.039551\n",
            "[252]\tvalidation_0-auc:0.840284\tvalidation_0-error:0.039551\n",
            "[253]\tvalidation_0-auc:0.840349\tvalidation_0-error:0.039551\n",
            "[254]\tvalidation_0-auc:0.840288\tvalidation_0-error:0.039551\n",
            "[255]\tvalidation_0-auc:0.840557\tvalidation_0-error:0.039551\n",
            "[256]\tvalidation_0-auc:0.840623\tvalidation_0-error:0.039551\n",
            "[257]\tvalidation_0-auc:0.840714\tvalidation_0-error:0.039551\n",
            "[258]\tvalidation_0-auc:0.84071\tvalidation_0-error:0.039551\n",
            "[259]\tvalidation_0-auc:0.840694\tvalidation_0-error:0.039551\n",
            "[260]\tvalidation_0-auc:0.840803\tvalidation_0-error:0.039551\n",
            "[261]\tvalidation_0-auc:0.840726\tvalidation_0-error:0.039551\n",
            "[262]\tvalidation_0-auc:0.840868\tvalidation_0-error:0.039551\n",
            "[263]\tvalidation_0-auc:0.840832\tvalidation_0-error:0.039551\n",
            "[264]\tvalidation_0-auc:0.840836\tvalidation_0-error:0.039551\n",
            "[265]\tvalidation_0-auc:0.840829\tvalidation_0-error:0.039551\n",
            "[266]\tvalidation_0-auc:0.840919\tvalidation_0-error:0.039551\n",
            "[267]\tvalidation_0-auc:0.841108\tvalidation_0-error:0.039551\n",
            "[268]\tvalidation_0-auc:0.841306\tvalidation_0-error:0.039551\n",
            "[269]\tvalidation_0-auc:0.841473\tvalidation_0-error:0.039551\n",
            "[270]\tvalidation_0-auc:0.841615\tvalidation_0-error:0.039551\n",
            "[271]\tvalidation_0-auc:0.841483\tvalidation_0-error:0.039551\n",
            "[272]\tvalidation_0-auc:0.841447\tvalidation_0-error:0.039551\n",
            "[273]\tvalidation_0-auc:0.84155\tvalidation_0-error:0.039551\n",
            "[274]\tvalidation_0-auc:0.841613\tvalidation_0-error:0.039551\n",
            "[275]\tvalidation_0-auc:0.841555\tvalidation_0-error:0.039551\n",
            "[276]\tvalidation_0-auc:0.841588\tvalidation_0-error:0.039551\n",
            "[277]\tvalidation_0-auc:0.841633\tvalidation_0-error:0.039551\n",
            "[278]\tvalidation_0-auc:0.841525\tvalidation_0-error:0.039551\n",
            "[279]\tvalidation_0-auc:0.841605\tvalidation_0-error:0.039551\n",
            "[280]\tvalidation_0-auc:0.841615\tvalidation_0-error:0.039551\n",
            "[281]\tvalidation_0-auc:0.841551\tvalidation_0-error:0.039551\n",
            "[282]\tvalidation_0-auc:0.841629\tvalidation_0-error:0.039551\n",
            "[283]\tvalidation_0-auc:0.841685\tvalidation_0-error:0.039551\n",
            "[284]\tvalidation_0-auc:0.841765\tvalidation_0-error:0.039551\n",
            "[285]\tvalidation_0-auc:0.841734\tvalidation_0-error:0.039551\n",
            "[286]\tvalidation_0-auc:0.84172\tvalidation_0-error:0.039551\n",
            "[287]\tvalidation_0-auc:0.841544\tvalidation_0-error:0.039551\n",
            "[288]\tvalidation_0-auc:0.841532\tvalidation_0-error:0.039551\n",
            "[289]\tvalidation_0-auc:0.841511\tvalidation_0-error:0.039551\n",
            "[290]\tvalidation_0-auc:0.841624\tvalidation_0-error:0.039551\n",
            "[291]\tvalidation_0-auc:0.841632\tvalidation_0-error:0.039551\n",
            "[292]\tvalidation_0-auc:0.841703\tvalidation_0-error:0.039551\n",
            "[293]\tvalidation_0-auc:0.841702\tvalidation_0-error:0.039551\n",
            "[294]\tvalidation_0-auc:0.841833\tvalidation_0-error:0.039551\n",
            "[295]\tvalidation_0-auc:0.841854\tvalidation_0-error:0.039551\n",
            "[296]\tvalidation_0-auc:0.841979\tvalidation_0-error:0.039551\n",
            "[297]\tvalidation_0-auc:0.841913\tvalidation_0-error:0.039551\n",
            "[298]\tvalidation_0-auc:0.8419\tvalidation_0-error:0.039551\n",
            "[299]\tvalidation_0-auc:0.841975\tvalidation_0-error:0.039551\n",
            "[300]\tvalidation_0-auc:0.842115\tvalidation_0-error:0.039551\n",
            "[301]\tvalidation_0-auc:0.842178\tvalidation_0-error:0.039551\n",
            "[302]\tvalidation_0-auc:0.842145\tvalidation_0-error:0.039551\n",
            "[303]\tvalidation_0-auc:0.842225\tvalidation_0-error:0.039551\n",
            "[304]\tvalidation_0-auc:0.842214\tvalidation_0-error:0.039551\n",
            "[305]\tvalidation_0-auc:0.842198\tvalidation_0-error:0.039551\n",
            "[306]\tvalidation_0-auc:0.842099\tvalidation_0-error:0.039551\n",
            "[307]\tvalidation_0-auc:0.842133\tvalidation_0-error:0.039551\n",
            "[308]\tvalidation_0-auc:0.84211\tvalidation_0-error:0.039551\n",
            "[309]\tvalidation_0-auc:0.842214\tvalidation_0-error:0.039551\n",
            "[310]\tvalidation_0-auc:0.84221\tvalidation_0-error:0.039551\n",
            "[311]\tvalidation_0-auc:0.842207\tvalidation_0-error:0.039551\n",
            "[312]\tvalidation_0-auc:0.842188\tvalidation_0-error:0.039551\n",
            "[313]\tvalidation_0-auc:0.842184\tvalidation_0-error:0.039551\n",
            "[314]\tvalidation_0-auc:0.842284\tvalidation_0-error:0.039551\n",
            "[315]\tvalidation_0-auc:0.842281\tvalidation_0-error:0.039551\n",
            "[316]\tvalidation_0-auc:0.842262\tvalidation_0-error:0.039551\n",
            "[317]\tvalidation_0-auc:0.842289\tvalidation_0-error:0.039551\n",
            "[318]\tvalidation_0-auc:0.842284\tvalidation_0-error:0.039551\n",
            "[319]\tvalidation_0-auc:0.84238\tvalidation_0-error:0.039551\n",
            "[320]\tvalidation_0-auc:0.842481\tvalidation_0-error:0.039551\n",
            "[321]\tvalidation_0-auc:0.842473\tvalidation_0-error:0.039551\n",
            "[322]\tvalidation_0-auc:0.8425\tvalidation_0-error:0.039551\n",
            "[323]\tvalidation_0-auc:0.842553\tvalidation_0-error:0.039551\n",
            "[324]\tvalidation_0-auc:0.84259\tvalidation_0-error:0.039551\n",
            "[325]\tvalidation_0-auc:0.842572\tvalidation_0-error:0.039551\n",
            "[326]\tvalidation_0-auc:0.842601\tvalidation_0-error:0.039551\n",
            "[327]\tvalidation_0-auc:0.842559\tvalidation_0-error:0.039551\n",
            "[328]\tvalidation_0-auc:0.842522\tvalidation_0-error:0.039551\n",
            "[329]\tvalidation_0-auc:0.842612\tvalidation_0-error:0.039551\n",
            "[330]\tvalidation_0-auc:0.842735\tvalidation_0-error:0.039551\n",
            "[331]\tvalidation_0-auc:0.842814\tvalidation_0-error:0.039551\n",
            "[332]\tvalidation_0-auc:0.842844\tvalidation_0-error:0.039551\n",
            "[333]\tvalidation_0-auc:0.842865\tvalidation_0-error:0.039551\n",
            "[334]\tvalidation_0-auc:0.842793\tvalidation_0-error:0.039551\n",
            "[335]\tvalidation_0-auc:0.842875\tvalidation_0-error:0.039551\n",
            "[336]\tvalidation_0-auc:0.84291\tvalidation_0-error:0.039551\n",
            "[337]\tvalidation_0-auc:0.842924\tvalidation_0-error:0.039551\n",
            "[338]\tvalidation_0-auc:0.842936\tvalidation_0-error:0.039551\n",
            "[339]\tvalidation_0-auc:0.843\tvalidation_0-error:0.039551\n",
            "[340]\tvalidation_0-auc:0.843126\tvalidation_0-error:0.039551\n",
            "[341]\tvalidation_0-auc:0.843235\tvalidation_0-error:0.039551\n",
            "[342]\tvalidation_0-auc:0.843303\tvalidation_0-error:0.039551\n",
            "[343]\tvalidation_0-auc:0.843369\tvalidation_0-error:0.039551\n",
            "[344]\tvalidation_0-auc:0.843361\tvalidation_0-error:0.039551\n",
            "[345]\tvalidation_0-auc:0.843427\tvalidation_0-error:0.039551\n",
            "[346]\tvalidation_0-auc:0.843475\tvalidation_0-error:0.039551\n",
            "[347]\tvalidation_0-auc:0.843564\tvalidation_0-error:0.039551\n",
            "[348]\tvalidation_0-auc:0.843583\tvalidation_0-error:0.039551\n",
            "[349]\tvalidation_0-auc:0.843592\tvalidation_0-error:0.039551\n",
            "[350]\tvalidation_0-auc:0.843613\tvalidation_0-error:0.039551\n",
            "[351]\tvalidation_0-auc:0.843584\tvalidation_0-error:0.039551\n",
            "[352]\tvalidation_0-auc:0.843657\tvalidation_0-error:0.039551\n",
            "[353]\tvalidation_0-auc:0.843708\tvalidation_0-error:0.039551\n",
            "[354]\tvalidation_0-auc:0.843705\tvalidation_0-error:0.039551\n",
            "[355]\tvalidation_0-auc:0.84374\tvalidation_0-error:0.039551\n",
            "[356]\tvalidation_0-auc:0.843737\tvalidation_0-error:0.039551\n",
            "[357]\tvalidation_0-auc:0.843796\tvalidation_0-error:0.039551\n",
            "[358]\tvalidation_0-auc:0.843803\tvalidation_0-error:0.039551\n",
            "[359]\tvalidation_0-auc:0.843838\tvalidation_0-error:0.039551\n",
            "[360]\tvalidation_0-auc:0.843865\tvalidation_0-error:0.039551\n",
            "[361]\tvalidation_0-auc:0.843861\tvalidation_0-error:0.039551\n",
            "[362]\tvalidation_0-auc:0.843878\tvalidation_0-error:0.039551\n",
            "[363]\tvalidation_0-auc:0.843948\tvalidation_0-error:0.039551\n",
            "[364]\tvalidation_0-auc:0.843993\tvalidation_0-error:0.039551\n",
            "[365]\tvalidation_0-auc:0.844076\tvalidation_0-error:0.039551\n",
            "[366]\tvalidation_0-auc:0.84414\tvalidation_0-error:0.039551\n",
            "[367]\tvalidation_0-auc:0.844151\tvalidation_0-error:0.039551\n",
            "[368]\tvalidation_0-auc:0.844163\tvalidation_0-error:0.039551\n",
            "[369]\tvalidation_0-auc:0.844202\tvalidation_0-error:0.039551\n",
            "[370]\tvalidation_0-auc:0.844253\tvalidation_0-error:0.039551\n",
            "[371]\tvalidation_0-auc:0.844296\tvalidation_0-error:0.039551\n",
            "[372]\tvalidation_0-auc:0.844292\tvalidation_0-error:0.039551\n",
            "[373]\tvalidation_0-auc:0.844278\tvalidation_0-error:0.039551\n",
            "[374]\tvalidation_0-auc:0.844353\tvalidation_0-error:0.039551\n",
            "[375]\tvalidation_0-auc:0.844403\tvalidation_0-error:0.039551\n",
            "[376]\tvalidation_0-auc:0.844474\tvalidation_0-error:0.039551\n",
            "[377]\tvalidation_0-auc:0.844461\tvalidation_0-error:0.039551\n",
            "[378]\tvalidation_0-auc:0.844523\tvalidation_0-error:0.039551\n",
            "[379]\tvalidation_0-auc:0.844543\tvalidation_0-error:0.039551\n",
            "[380]\tvalidation_0-auc:0.844566\tvalidation_0-error:0.039551\n",
            "[381]\tvalidation_0-auc:0.844649\tvalidation_0-error:0.039551\n",
            "[382]\tvalidation_0-auc:0.844604\tvalidation_0-error:0.039551\n",
            "[383]\tvalidation_0-auc:0.844586\tvalidation_0-error:0.039551\n",
            "[384]\tvalidation_0-auc:0.844555\tvalidation_0-error:0.039551\n",
            "[385]\tvalidation_0-auc:0.844523\tvalidation_0-error:0.039551\n",
            "[386]\tvalidation_0-auc:0.844519\tvalidation_0-error:0.039551\n",
            "[387]\tvalidation_0-auc:0.844598\tvalidation_0-error:0.039551\n",
            "[388]\tvalidation_0-auc:0.844608\tvalidation_0-error:0.039551\n",
            "[389]\tvalidation_0-auc:0.844561\tvalidation_0-error:0.039551\n",
            "[390]\tvalidation_0-auc:0.844574\tvalidation_0-error:0.039551\n",
            "[391]\tvalidation_0-auc:0.844599\tvalidation_0-error:0.039551\n",
            "[392]\tvalidation_0-auc:0.844598\tvalidation_0-error:0.039551\n",
            "[393]\tvalidation_0-auc:0.844604\tvalidation_0-error:0.039551\n",
            "[394]\tvalidation_0-auc:0.844594\tvalidation_0-error:0.039551\n",
            "[395]\tvalidation_0-auc:0.844547\tvalidation_0-error:0.039551\n",
            "[396]\tvalidation_0-auc:0.844568\tvalidation_0-error:0.039551\n",
            "[397]\tvalidation_0-auc:0.844557\tvalidation_0-error:0.039551\n",
            "[398]\tvalidation_0-auc:0.844494\tvalidation_0-error:0.039551\n",
            "[399]\tvalidation_0-auc:0.844494\tvalidation_0-error:0.039551\n",
            "[400]\tvalidation_0-auc:0.844512\tvalidation_0-error:0.039551\n",
            "[401]\tvalidation_0-auc:0.844592\tvalidation_0-error:0.039551\n",
            "[402]\tvalidation_0-auc:0.844553\tvalidation_0-error:0.039551\n",
            "[403]\tvalidation_0-auc:0.844577\tvalidation_0-error:0.039551\n",
            "[404]\tvalidation_0-auc:0.844545\tvalidation_0-error:0.039551\n",
            "[405]\tvalidation_0-auc:0.844556\tvalidation_0-error:0.039551\n",
            "[406]\tvalidation_0-auc:0.844604\tvalidation_0-error:0.039551\n",
            "[407]\tvalidation_0-auc:0.844652\tvalidation_0-error:0.039551\n",
            "[408]\tvalidation_0-auc:0.844717\tvalidation_0-error:0.039551\n",
            "[409]\tvalidation_0-auc:0.844796\tvalidation_0-error:0.039551\n",
            "[410]\tvalidation_0-auc:0.844833\tvalidation_0-error:0.039551\n",
            "[411]\tvalidation_0-auc:0.844831\tvalidation_0-error:0.039551\n",
            "[412]\tvalidation_0-auc:0.844861\tvalidation_0-error:0.039551\n",
            "[413]\tvalidation_0-auc:0.844879\tvalidation_0-error:0.039551\n",
            "[414]\tvalidation_0-auc:0.844917\tvalidation_0-error:0.039551\n",
            "[415]\tvalidation_0-auc:0.845065\tvalidation_0-error:0.039551\n",
            "[416]\tvalidation_0-auc:0.845101\tvalidation_0-error:0.039551\n",
            "[417]\tvalidation_0-auc:0.845047\tvalidation_0-error:0.039551\n",
            "[418]\tvalidation_0-auc:0.845007\tvalidation_0-error:0.039551\n",
            "[419]\tvalidation_0-auc:0.845012\tvalidation_0-error:0.039551\n",
            "[420]\tvalidation_0-auc:0.844992\tvalidation_0-error:0.039551\n",
            "[421]\tvalidation_0-auc:0.845014\tvalidation_0-error:0.039551\n",
            "[422]\tvalidation_0-auc:0.844997\tvalidation_0-error:0.039551\n",
            "[423]\tvalidation_0-auc:0.844977\tvalidation_0-error:0.039551\n",
            "[424]\tvalidation_0-auc:0.84493\tvalidation_0-error:0.039551\n",
            "[425]\tvalidation_0-auc:0.844979\tvalidation_0-error:0.039551\n",
            "[426]\tvalidation_0-auc:0.84498\tvalidation_0-error:0.039551\n",
            "[427]\tvalidation_0-auc:0.845006\tvalidation_0-error:0.039551\n",
            "[428]\tvalidation_0-auc:0.844972\tvalidation_0-error:0.039551\n",
            "[429]\tvalidation_0-auc:0.844975\tvalidation_0-error:0.039551\n",
            "[430]\tvalidation_0-auc:0.84501\tvalidation_0-error:0.039551\n",
            "[431]\tvalidation_0-auc:0.845069\tvalidation_0-error:0.039551\n",
            "[432]\tvalidation_0-auc:0.845089\tvalidation_0-error:0.039551\n",
            "[433]\tvalidation_0-auc:0.845105\tvalidation_0-error:0.039551\n",
            "[434]\tvalidation_0-auc:0.845197\tvalidation_0-error:0.039551\n",
            "[435]\tvalidation_0-auc:0.845187\tvalidation_0-error:0.039551\n",
            "[436]\tvalidation_0-auc:0.845153\tvalidation_0-error:0.039551\n",
            "[437]\tvalidation_0-auc:0.845241\tvalidation_0-error:0.039551\n",
            "[438]\tvalidation_0-auc:0.845207\tvalidation_0-error:0.039551\n",
            "[439]\tvalidation_0-auc:0.845299\tvalidation_0-error:0.039551\n",
            "[440]\tvalidation_0-auc:0.845277\tvalidation_0-error:0.039551\n",
            "[441]\tvalidation_0-auc:0.845325\tvalidation_0-error:0.039551\n",
            "[442]\tvalidation_0-auc:0.845366\tvalidation_0-error:0.039551\n",
            "[443]\tvalidation_0-auc:0.845378\tvalidation_0-error:0.039551\n",
            "[444]\tvalidation_0-auc:0.845425\tvalidation_0-error:0.039551\n",
            "[445]\tvalidation_0-auc:0.845412\tvalidation_0-error:0.039551\n",
            "[446]\tvalidation_0-auc:0.845487\tvalidation_0-error:0.039551\n",
            "[447]\tvalidation_0-auc:0.8455\tvalidation_0-error:0.039551\n",
            "[448]\tvalidation_0-auc:0.845545\tvalidation_0-error:0.039551\n",
            "[449]\tvalidation_0-auc:0.845461\tvalidation_0-error:0.039551\n",
            "[450]\tvalidation_0-auc:0.845468\tvalidation_0-error:0.039551\n",
            "[451]\tvalidation_0-auc:0.845472\tvalidation_0-error:0.039551\n",
            "[452]\tvalidation_0-auc:0.845518\tvalidation_0-error:0.039551\n",
            "[453]\tvalidation_0-auc:0.845533\tvalidation_0-error:0.039551\n",
            "[454]\tvalidation_0-auc:0.845597\tvalidation_0-error:0.039551\n",
            "[455]\tvalidation_0-auc:0.845625\tvalidation_0-error:0.039551\n",
            "[456]\tvalidation_0-auc:0.845616\tvalidation_0-error:0.039551\n",
            "[457]\tvalidation_0-auc:0.845581\tvalidation_0-error:0.039551\n",
            "[458]\tvalidation_0-auc:0.845541\tvalidation_0-error:0.039551\n",
            "[459]\tvalidation_0-auc:0.845531\tvalidation_0-error:0.039551\n",
            "[460]\tvalidation_0-auc:0.845544\tvalidation_0-error:0.039551\n",
            "[461]\tvalidation_0-auc:0.84552\tvalidation_0-error:0.039551\n",
            "[462]\tvalidation_0-auc:0.845503\tvalidation_0-error:0.039551\n",
            "[463]\tvalidation_0-auc:0.845581\tvalidation_0-error:0.039551\n",
            "[464]\tvalidation_0-auc:0.845627\tvalidation_0-error:0.039551\n",
            "[465]\tvalidation_0-auc:0.845649\tvalidation_0-error:0.039551\n",
            "[466]\tvalidation_0-auc:0.845598\tvalidation_0-error:0.039551\n",
            "[467]\tvalidation_0-auc:0.845625\tvalidation_0-error:0.039551\n",
            "[468]\tvalidation_0-auc:0.845617\tvalidation_0-error:0.039551\n",
            "[469]\tvalidation_0-auc:0.845647\tvalidation_0-error:0.039551\n",
            "[470]\tvalidation_0-auc:0.845703\tvalidation_0-error:0.039551\n",
            "[471]\tvalidation_0-auc:0.845696\tvalidation_0-error:0.039551\n",
            "[472]\tvalidation_0-auc:0.845663\tvalidation_0-error:0.039551\n",
            "[473]\tvalidation_0-auc:0.845675\tvalidation_0-error:0.039551\n",
            "[474]\tvalidation_0-auc:0.845637\tvalidation_0-error:0.039551\n",
            "[475]\tvalidation_0-auc:0.845631\tvalidation_0-error:0.039551\n",
            "[476]\tvalidation_0-auc:0.845618\tvalidation_0-error:0.039551\n",
            "[477]\tvalidation_0-auc:0.845618\tvalidation_0-error:0.039551\n",
            "[478]\tvalidation_0-auc:0.845633\tvalidation_0-error:0.039551\n",
            "[479]\tvalidation_0-auc:0.845603\tvalidation_0-error:0.039551\n",
            "[480]\tvalidation_0-auc:0.845594\tvalidation_0-error:0.039551\n",
            "[481]\tvalidation_0-auc:0.8456\tvalidation_0-error:0.039551\n",
            "[482]\tvalidation_0-auc:0.845564\tvalidation_0-error:0.039551\n",
            "[483]\tvalidation_0-auc:0.845543\tvalidation_0-error:0.039551\n",
            "[484]\tvalidation_0-auc:0.845625\tvalidation_0-error:0.039551\n",
            "[485]\tvalidation_0-auc:0.845618\tvalidation_0-error:0.039551\n",
            "[486]\tvalidation_0-auc:0.845668\tvalidation_0-error:0.039551\n",
            "[487]\tvalidation_0-auc:0.845676\tvalidation_0-error:0.039551\n",
            "[488]\tvalidation_0-auc:0.845665\tvalidation_0-error:0.039551\n",
            "[489]\tvalidation_0-auc:0.845691\tvalidation_0-error:0.039551\n",
            "[490]\tvalidation_0-auc:0.845684\tvalidation_0-error:0.039551\n",
            "[491]\tvalidation_0-auc:0.845658\tvalidation_0-error:0.039551\n",
            "[492]\tvalidation_0-auc:0.845685\tvalidation_0-error:0.039551\n",
            "[493]\tvalidation_0-auc:0.845707\tvalidation_0-error:0.039551\n",
            "[494]\tvalidation_0-auc:0.845684\tvalidation_0-error:0.039551\n",
            "[495]\tvalidation_0-auc:0.845681\tvalidation_0-error:0.039551\n",
            "[496]\tvalidation_0-auc:0.845667\tvalidation_0-error:0.039551\n",
            "[497]\tvalidation_0-auc:0.84569\tvalidation_0-error:0.039551\n",
            "[498]\tvalidation_0-auc:0.845673\tvalidation_0-error:0.039551\n",
            "[499]\tvalidation_0-auc:0.845664\tvalidation_0-error:0.039551\n",
            "[500]\tvalidation_0-auc:0.845681\tvalidation_0-error:0.039551\n",
            "[501]\tvalidation_0-auc:0.845684\tvalidation_0-error:0.039551\n",
            "[502]\tvalidation_0-auc:0.845689\tvalidation_0-error:0.039551\n",
            "[503]\tvalidation_0-auc:0.845703\tvalidation_0-error:0.039551\n",
            "[504]\tvalidation_0-auc:0.845694\tvalidation_0-error:0.039551\n",
            "[505]\tvalidation_0-auc:0.845717\tvalidation_0-error:0.039551\n",
            "[506]\tvalidation_0-auc:0.845693\tvalidation_0-error:0.039551\n",
            "[507]\tvalidation_0-auc:0.845644\tvalidation_0-error:0.039551\n",
            "[508]\tvalidation_0-auc:0.845635\tvalidation_0-error:0.039551\n",
            "[509]\tvalidation_0-auc:0.845641\tvalidation_0-error:0.039551\n",
            "[510]\tvalidation_0-auc:0.84559\tvalidation_0-error:0.039551\n",
            "[511]\tvalidation_0-auc:0.845631\tvalidation_0-error:0.039551\n",
            "[512]\tvalidation_0-auc:0.845618\tvalidation_0-error:0.039551\n",
            "[513]\tvalidation_0-auc:0.845639\tvalidation_0-error:0.039551\n",
            "[514]\tvalidation_0-auc:0.845632\tvalidation_0-error:0.039551\n",
            "[515]\tvalidation_0-auc:0.845679\tvalidation_0-error:0.039551\n",
            "[516]\tvalidation_0-auc:0.8457\tvalidation_0-error:0.039551\n",
            "[517]\tvalidation_0-auc:0.845691\tvalidation_0-error:0.039551\n",
            "[518]\tvalidation_0-auc:0.845684\tvalidation_0-error:0.039551\n",
            "[519]\tvalidation_0-auc:0.84568\tvalidation_0-error:0.039551\n",
            "[520]\tvalidation_0-auc:0.84565\tvalidation_0-error:0.039551\n",
            "[521]\tvalidation_0-auc:0.845683\tvalidation_0-error:0.039551\n",
            "[522]\tvalidation_0-auc:0.845681\tvalidation_0-error:0.039551\n",
            "[523]\tvalidation_0-auc:0.845657\tvalidation_0-error:0.039551\n",
            "[524]\tvalidation_0-auc:0.845685\tvalidation_0-error:0.039551\n",
            "[525]\tvalidation_0-auc:0.845712\tvalidation_0-error:0.039551\n",
            "[526]\tvalidation_0-auc:0.845738\tvalidation_0-error:0.039551\n",
            "[527]\tvalidation_0-auc:0.845746\tvalidation_0-error:0.039551\n",
            "[528]\tvalidation_0-auc:0.845729\tvalidation_0-error:0.039551\n",
            "[529]\tvalidation_0-auc:0.845712\tvalidation_0-error:0.039551\n",
            "[530]\tvalidation_0-auc:0.845762\tvalidation_0-error:0.039551\n",
            "[531]\tvalidation_0-auc:0.845794\tvalidation_0-error:0.039551\n",
            "[532]\tvalidation_0-auc:0.845811\tvalidation_0-error:0.039551\n",
            "[533]\tvalidation_0-auc:0.84589\tvalidation_0-error:0.039551\n",
            "[534]\tvalidation_0-auc:0.845916\tvalidation_0-error:0.039551\n",
            "[535]\tvalidation_0-auc:0.845946\tvalidation_0-error:0.039551\n",
            "[536]\tvalidation_0-auc:0.845911\tvalidation_0-error:0.039551\n",
            "[537]\tvalidation_0-auc:0.84588\tvalidation_0-error:0.039551\n",
            "[538]\tvalidation_0-auc:0.845861\tvalidation_0-error:0.039551\n",
            "[539]\tvalidation_0-auc:0.845825\tvalidation_0-error:0.039551\n",
            "[540]\tvalidation_0-auc:0.845874\tvalidation_0-error:0.039551\n",
            "[541]\tvalidation_0-auc:0.845837\tvalidation_0-error:0.039551\n",
            "[542]\tvalidation_0-auc:0.845841\tvalidation_0-error:0.039551\n",
            "[543]\tvalidation_0-auc:0.845848\tvalidation_0-error:0.039551\n",
            "[544]\tvalidation_0-auc:0.845868\tvalidation_0-error:0.039551\n",
            "[545]\tvalidation_0-auc:0.845906\tvalidation_0-error:0.039551\n",
            "[546]\tvalidation_0-auc:0.845969\tvalidation_0-error:0.039551\n",
            "[547]\tvalidation_0-auc:0.845949\tvalidation_0-error:0.039551\n",
            "[548]\tvalidation_0-auc:0.845995\tvalidation_0-error:0.039551\n",
            "[549]\tvalidation_0-auc:0.845978\tvalidation_0-error:0.039551\n",
            "[550]\tvalidation_0-auc:0.845957\tvalidation_0-error:0.039551\n",
            "[551]\tvalidation_0-auc:0.845934\tvalidation_0-error:0.039551\n",
            "[552]\tvalidation_0-auc:0.845974\tvalidation_0-error:0.039551\n",
            "[553]\tvalidation_0-auc:0.845927\tvalidation_0-error:0.039551\n",
            "[554]\tvalidation_0-auc:0.845923\tvalidation_0-error:0.039551\n",
            "[555]\tvalidation_0-auc:0.845953\tvalidation_0-error:0.039551\n",
            "[556]\tvalidation_0-auc:0.845986\tvalidation_0-error:0.039551\n",
            "[557]\tvalidation_0-auc:0.846017\tvalidation_0-error:0.039551\n",
            "[558]\tvalidation_0-auc:0.846031\tvalidation_0-error:0.039551\n",
            "[559]\tvalidation_0-auc:0.846019\tvalidation_0-error:0.039551\n",
            "[560]\tvalidation_0-auc:0.846026\tvalidation_0-error:0.039551\n",
            "[561]\tvalidation_0-auc:0.846064\tvalidation_0-error:0.039551\n",
            "[562]\tvalidation_0-auc:0.846028\tvalidation_0-error:0.039551\n",
            "[563]\tvalidation_0-auc:0.846048\tvalidation_0-error:0.039551\n",
            "[564]\tvalidation_0-auc:0.846007\tvalidation_0-error:0.039551\n",
            "[565]\tvalidation_0-auc:0.846001\tvalidation_0-error:0.039551\n",
            "[566]\tvalidation_0-auc:0.846018\tvalidation_0-error:0.039551\n",
            "[567]\tvalidation_0-auc:0.846038\tvalidation_0-error:0.039551\n",
            "[568]\tvalidation_0-auc:0.846047\tvalidation_0-error:0.039551\n",
            "[569]\tvalidation_0-auc:0.846045\tvalidation_0-error:0.039551\n",
            "[570]\tvalidation_0-auc:0.846027\tvalidation_0-error:0.039551\n",
            "[571]\tvalidation_0-auc:0.845992\tvalidation_0-error:0.039551\n",
            "[572]\tvalidation_0-auc:0.845999\tvalidation_0-error:0.039551\n",
            "[573]\tvalidation_0-auc:0.846063\tvalidation_0-error:0.039551\n",
            "[574]\tvalidation_0-auc:0.846023\tvalidation_0-error:0.039551\n",
            "[575]\tvalidation_0-auc:0.846065\tvalidation_0-error:0.039551\n",
            "[576]\tvalidation_0-auc:0.84604\tvalidation_0-error:0.039551\n",
            "[577]\tvalidation_0-auc:0.846003\tvalidation_0-error:0.039551\n",
            "[578]\tvalidation_0-auc:0.846025\tvalidation_0-error:0.039551\n",
            "[579]\tvalidation_0-auc:0.846044\tvalidation_0-error:0.039551\n",
            "[580]\tvalidation_0-auc:0.846061\tvalidation_0-error:0.039551\n",
            "[581]\tvalidation_0-auc:0.846059\tvalidation_0-error:0.039551\n",
            "[582]\tvalidation_0-auc:0.846058\tvalidation_0-error:0.039551\n",
            "[583]\tvalidation_0-auc:0.846071\tvalidation_0-error:0.039551\n",
            "[584]\tvalidation_0-auc:0.846059\tvalidation_0-error:0.039551\n",
            "[585]\tvalidation_0-auc:0.846048\tvalidation_0-error:0.039551\n",
            "[586]\tvalidation_0-auc:0.846074\tvalidation_0-error:0.039551\n",
            "[587]\tvalidation_0-auc:0.846109\tvalidation_0-error:0.039551\n",
            "[588]\tvalidation_0-auc:0.84611\tvalidation_0-error:0.039551\n",
            "[589]\tvalidation_0-auc:0.846108\tvalidation_0-error:0.039551\n",
            "[590]\tvalidation_0-auc:0.846103\tvalidation_0-error:0.039551\n",
            "[591]\tvalidation_0-auc:0.846102\tvalidation_0-error:0.039551\n",
            "[592]\tvalidation_0-auc:0.846109\tvalidation_0-error:0.039551\n",
            "[593]\tvalidation_0-auc:0.846169\tvalidation_0-error:0.039551\n",
            "[594]\tvalidation_0-auc:0.846149\tvalidation_0-error:0.039551\n",
            "[595]\tvalidation_0-auc:0.846164\tvalidation_0-error:0.039551\n",
            "[596]\tvalidation_0-auc:0.846165\tvalidation_0-error:0.039551\n",
            "[597]\tvalidation_0-auc:0.846217\tvalidation_0-error:0.039551\n",
            "[598]\tvalidation_0-auc:0.84624\tvalidation_0-error:0.039551\n",
            "[599]\tvalidation_0-auc:0.846228\tvalidation_0-error:0.039551\n",
            "[600]\tvalidation_0-auc:0.846253\tvalidation_0-error:0.039551\n",
            "[601]\tvalidation_0-auc:0.846274\tvalidation_0-error:0.039551\n",
            "[602]\tvalidation_0-auc:0.846298\tvalidation_0-error:0.039551\n",
            "[603]\tvalidation_0-auc:0.846289\tvalidation_0-error:0.039551\n",
            "[604]\tvalidation_0-auc:0.846299\tvalidation_0-error:0.039551\n",
            "[605]\tvalidation_0-auc:0.846255\tvalidation_0-error:0.039551\n",
            "[606]\tvalidation_0-auc:0.846279\tvalidation_0-error:0.039551\n",
            "[607]\tvalidation_0-auc:0.846275\tvalidation_0-error:0.039551\n",
            "[608]\tvalidation_0-auc:0.846267\tvalidation_0-error:0.039551\n",
            "[609]\tvalidation_0-auc:0.84628\tvalidation_0-error:0.039551\n",
            "[610]\tvalidation_0-auc:0.846267\tvalidation_0-error:0.039551\n",
            "[611]\tvalidation_0-auc:0.846237\tvalidation_0-error:0.039551\n",
            "[612]\tvalidation_0-auc:0.846241\tvalidation_0-error:0.039551\n",
            "[613]\tvalidation_0-auc:0.846247\tvalidation_0-error:0.039551\n",
            "[614]\tvalidation_0-auc:0.846269\tvalidation_0-error:0.039551\n",
            "[615]\tvalidation_0-auc:0.846241\tvalidation_0-error:0.039551\n",
            "[616]\tvalidation_0-auc:0.846286\tvalidation_0-error:0.039551\n",
            "[617]\tvalidation_0-auc:0.846303\tvalidation_0-error:0.039551\n",
            "[618]\tvalidation_0-auc:0.846331\tvalidation_0-error:0.039551\n",
            "[619]\tvalidation_0-auc:0.8463\tvalidation_0-error:0.039551\n",
            "[620]\tvalidation_0-auc:0.846324\tvalidation_0-error:0.039551\n",
            "[621]\tvalidation_0-auc:0.846329\tvalidation_0-error:0.039551\n",
            "[622]\tvalidation_0-auc:0.846354\tvalidation_0-error:0.039551\n",
            "[623]\tvalidation_0-auc:0.846352\tvalidation_0-error:0.039551\n",
            "[624]\tvalidation_0-auc:0.846379\tvalidation_0-error:0.039551\n",
            "[625]\tvalidation_0-auc:0.846489\tvalidation_0-error:0.039551\n",
            "[626]\tvalidation_0-auc:0.846496\tvalidation_0-error:0.039551\n",
            "[627]\tvalidation_0-auc:0.846527\tvalidation_0-error:0.039551\n",
            "[628]\tvalidation_0-auc:0.846539\tvalidation_0-error:0.039551\n",
            "[629]\tvalidation_0-auc:0.846552\tvalidation_0-error:0.039551\n",
            "[630]\tvalidation_0-auc:0.846544\tvalidation_0-error:0.039551\n",
            "[631]\tvalidation_0-auc:0.846526\tvalidation_0-error:0.039551\n",
            "[632]\tvalidation_0-auc:0.846499\tvalidation_0-error:0.039551\n",
            "[633]\tvalidation_0-auc:0.846477\tvalidation_0-error:0.039551\n",
            "[634]\tvalidation_0-auc:0.846499\tvalidation_0-error:0.039551\n",
            "[635]\tvalidation_0-auc:0.846484\tvalidation_0-error:0.039551\n",
            "[636]\tvalidation_0-auc:0.846488\tvalidation_0-error:0.039551\n",
            "[637]\tvalidation_0-auc:0.846494\tvalidation_0-error:0.039551\n",
            "[638]\tvalidation_0-auc:0.846523\tvalidation_0-error:0.039551\n",
            "[639]\tvalidation_0-auc:0.8465\tvalidation_0-error:0.039551\n",
            "[640]\tvalidation_0-auc:0.846522\tvalidation_0-error:0.039551\n",
            "[641]\tvalidation_0-auc:0.846523\tvalidation_0-error:0.039551\n",
            "[642]\tvalidation_0-auc:0.846531\tvalidation_0-error:0.039551\n",
            "[643]\tvalidation_0-auc:0.846561\tvalidation_0-error:0.039551\n",
            "[644]\tvalidation_0-auc:0.846594\tvalidation_0-error:0.039551\n",
            "[645]\tvalidation_0-auc:0.846571\tvalidation_0-error:0.039551\n",
            "[646]\tvalidation_0-auc:0.846556\tvalidation_0-error:0.039551\n",
            "[647]\tvalidation_0-auc:0.846582\tvalidation_0-error:0.039551\n",
            "[648]\tvalidation_0-auc:0.846576\tvalidation_0-error:0.039551\n",
            "[649]\tvalidation_0-auc:0.846598\tvalidation_0-error:0.039551\n",
            "[650]\tvalidation_0-auc:0.846592\tvalidation_0-error:0.039551\n",
            "[651]\tvalidation_0-auc:0.846609\tvalidation_0-error:0.039551\n",
            "[652]\tvalidation_0-auc:0.846611\tvalidation_0-error:0.039551\n",
            "[653]\tvalidation_0-auc:0.846624\tvalidation_0-error:0.039551\n",
            "[654]\tvalidation_0-auc:0.846613\tvalidation_0-error:0.039551\n",
            "[655]\tvalidation_0-auc:0.846664\tvalidation_0-error:0.039551\n",
            "[656]\tvalidation_0-auc:0.846694\tvalidation_0-error:0.039551\n",
            "[657]\tvalidation_0-auc:0.846699\tvalidation_0-error:0.039551\n",
            "[658]\tvalidation_0-auc:0.846667\tvalidation_0-error:0.039551\n",
            "[659]\tvalidation_0-auc:0.846686\tvalidation_0-error:0.039551\n",
            "[660]\tvalidation_0-auc:0.846662\tvalidation_0-error:0.039551\n",
            "[661]\tvalidation_0-auc:0.846652\tvalidation_0-error:0.039551\n",
            "[662]\tvalidation_0-auc:0.846658\tvalidation_0-error:0.039551\n",
            "[663]\tvalidation_0-auc:0.846673\tvalidation_0-error:0.039551\n",
            "[664]\tvalidation_0-auc:0.846685\tvalidation_0-error:0.039551\n",
            "[665]\tvalidation_0-auc:0.846637\tvalidation_0-error:0.039551\n",
            "[666]\tvalidation_0-auc:0.846643\tvalidation_0-error:0.039551\n",
            "[667]\tvalidation_0-auc:0.846638\tvalidation_0-error:0.039551\n",
            "[668]\tvalidation_0-auc:0.846656\tvalidation_0-error:0.039551\n",
            "[669]\tvalidation_0-auc:0.846653\tvalidation_0-error:0.039551\n",
            "[670]\tvalidation_0-auc:0.846664\tvalidation_0-error:0.039551\n",
            "[671]\tvalidation_0-auc:0.84667\tvalidation_0-error:0.039551\n",
            "[672]\tvalidation_0-auc:0.84668\tvalidation_0-error:0.039551\n",
            "[673]\tvalidation_0-auc:0.846679\tvalidation_0-error:0.039551\n",
            "[674]\tvalidation_0-auc:0.846753\tvalidation_0-error:0.039551\n",
            "[675]\tvalidation_0-auc:0.846751\tvalidation_0-error:0.039551\n",
            "[676]\tvalidation_0-auc:0.846718\tvalidation_0-error:0.039551\n",
            "[677]\tvalidation_0-auc:0.846722\tvalidation_0-error:0.039551\n",
            "[678]\tvalidation_0-auc:0.846714\tvalidation_0-error:0.039551\n",
            "[679]\tvalidation_0-auc:0.846731\tvalidation_0-error:0.039551\n",
            "[680]\tvalidation_0-auc:0.846753\tvalidation_0-error:0.039551\n",
            "[681]\tvalidation_0-auc:0.846761\tvalidation_0-error:0.039551\n",
            "[682]\tvalidation_0-auc:0.846749\tvalidation_0-error:0.039551\n",
            "[683]\tvalidation_0-auc:0.846724\tvalidation_0-error:0.039551\n",
            "[684]\tvalidation_0-auc:0.846709\tvalidation_0-error:0.039551\n",
            "[685]\tvalidation_0-auc:0.846682\tvalidation_0-error:0.039551\n",
            "[686]\tvalidation_0-auc:0.846721\tvalidation_0-error:0.039551\n",
            "[687]\tvalidation_0-auc:0.846776\tvalidation_0-error:0.039551\n",
            "[688]\tvalidation_0-auc:0.84677\tvalidation_0-error:0.039551\n",
            "[689]\tvalidation_0-auc:0.846769\tvalidation_0-error:0.039551\n",
            "[690]\tvalidation_0-auc:0.846741\tvalidation_0-error:0.039551\n",
            "[691]\tvalidation_0-auc:0.846763\tvalidation_0-error:0.039551\n",
            "[692]\tvalidation_0-auc:0.846738\tvalidation_0-error:0.039551\n",
            "[693]\tvalidation_0-auc:0.846712\tvalidation_0-error:0.039551\n",
            "[694]\tvalidation_0-auc:0.846707\tvalidation_0-error:0.039551\n",
            "[695]\tvalidation_0-auc:0.846732\tvalidation_0-error:0.039551\n",
            "[696]\tvalidation_0-auc:0.846693\tvalidation_0-error:0.039551\n",
            "[697]\tvalidation_0-auc:0.846692\tvalidation_0-error:0.039551\n",
            "[698]\tvalidation_0-auc:0.84669\tvalidation_0-error:0.039551\n",
            "[699]\tvalidation_0-auc:0.846742\tvalidation_0-error:0.039551\n",
            "[700]\tvalidation_0-auc:0.846727\tvalidation_0-error:0.039551\n",
            "[701]\tvalidation_0-auc:0.846718\tvalidation_0-error:0.039551\n",
            "[702]\tvalidation_0-auc:0.846765\tvalidation_0-error:0.039551\n",
            "[703]\tvalidation_0-auc:0.846751\tvalidation_0-error:0.039551\n",
            "[704]\tvalidation_0-auc:0.8467\tvalidation_0-error:0.039551\n",
            "[705]\tvalidation_0-auc:0.846705\tvalidation_0-error:0.039551\n",
            "[706]\tvalidation_0-auc:0.846713\tvalidation_0-error:0.039551\n",
            "[707]\tvalidation_0-auc:0.846737\tvalidation_0-error:0.039551\n",
            "[708]\tvalidation_0-auc:0.846729\tvalidation_0-error:0.039551\n",
            "[709]\tvalidation_0-auc:0.84671\tvalidation_0-error:0.039551\n",
            "[710]\tvalidation_0-auc:0.84673\tvalidation_0-error:0.039551\n",
            "[711]\tvalidation_0-auc:0.846774\tvalidation_0-error:0.039551\n",
            "[712]\tvalidation_0-auc:0.846807\tvalidation_0-error:0.039551\n",
            "[713]\tvalidation_0-auc:0.846778\tvalidation_0-error:0.039551\n",
            "[714]\tvalidation_0-auc:0.846775\tvalidation_0-error:0.039551\n",
            "[715]\tvalidation_0-auc:0.846766\tvalidation_0-error:0.039551\n",
            "[716]\tvalidation_0-auc:0.846736\tvalidation_0-error:0.039551\n",
            "[717]\tvalidation_0-auc:0.846737\tvalidation_0-error:0.039551\n",
            "[718]\tvalidation_0-auc:0.846772\tvalidation_0-error:0.039551\n",
            "[719]\tvalidation_0-auc:0.846834\tvalidation_0-error:0.039551\n",
            "[720]\tvalidation_0-auc:0.84685\tvalidation_0-error:0.039551\n",
            "[721]\tvalidation_0-auc:0.846822\tvalidation_0-error:0.039551\n",
            "[722]\tvalidation_0-auc:0.84682\tvalidation_0-error:0.039551\n",
            "[723]\tvalidation_0-auc:0.8468\tvalidation_0-error:0.039551\n",
            "[724]\tvalidation_0-auc:0.846789\tvalidation_0-error:0.039551\n",
            "[725]\tvalidation_0-auc:0.846794\tvalidation_0-error:0.039551\n",
            "[726]\tvalidation_0-auc:0.846777\tvalidation_0-error:0.039551\n",
            "[727]\tvalidation_0-auc:0.846802\tvalidation_0-error:0.039551\n",
            "[728]\tvalidation_0-auc:0.846774\tvalidation_0-error:0.039551\n",
            "[729]\tvalidation_0-auc:0.846771\tvalidation_0-error:0.039551\n",
            "[730]\tvalidation_0-auc:0.846749\tvalidation_0-error:0.039551\n",
            "[731]\tvalidation_0-auc:0.846741\tvalidation_0-error:0.039551\n",
            "[732]\tvalidation_0-auc:0.84677\tvalidation_0-error:0.039551\n",
            "[733]\tvalidation_0-auc:0.846746\tvalidation_0-error:0.039551\n",
            "[734]\tvalidation_0-auc:0.84675\tvalidation_0-error:0.039551\n",
            "[735]\tvalidation_0-auc:0.846743\tvalidation_0-error:0.039551\n",
            "[736]\tvalidation_0-auc:0.846698\tvalidation_0-error:0.039551\n",
            "[737]\tvalidation_0-auc:0.84672\tvalidation_0-error:0.039551\n",
            "[738]\tvalidation_0-auc:0.846731\tvalidation_0-error:0.039551\n",
            "[739]\tvalidation_0-auc:0.846719\tvalidation_0-error:0.039551\n",
            "[740]\tvalidation_0-auc:0.846726\tvalidation_0-error:0.039551\n",
            "[741]\tvalidation_0-auc:0.846716\tvalidation_0-error:0.039551\n",
            "[742]\tvalidation_0-auc:0.84677\tvalidation_0-error:0.039551\n",
            "[743]\tvalidation_0-auc:0.846747\tvalidation_0-error:0.039551\n",
            "[744]\tvalidation_0-auc:0.846727\tvalidation_0-error:0.039551\n",
            "[745]\tvalidation_0-auc:0.846707\tvalidation_0-error:0.039551\n",
            "[746]\tvalidation_0-auc:0.846732\tvalidation_0-error:0.039551\n",
            "[747]\tvalidation_0-auc:0.846763\tvalidation_0-error:0.039551\n",
            "[748]\tvalidation_0-auc:0.846783\tvalidation_0-error:0.039551\n",
            "[749]\tvalidation_0-auc:0.846777\tvalidation_0-error:0.039551\n",
            "[750]\tvalidation_0-auc:0.846761\tvalidation_0-error:0.039551\n",
            "[751]\tvalidation_0-auc:0.84679\tvalidation_0-error:0.039551\n",
            "[752]\tvalidation_0-auc:0.846777\tvalidation_0-error:0.039551\n",
            "[753]\tvalidation_0-auc:0.846769\tvalidation_0-error:0.039551\n",
            "[754]\tvalidation_0-auc:0.846784\tvalidation_0-error:0.039551\n",
            "[755]\tvalidation_0-auc:0.846705\tvalidation_0-error:0.039551\n",
            "[756]\tvalidation_0-auc:0.846764\tvalidation_0-error:0.039551\n",
            "[757]\tvalidation_0-auc:0.846772\tvalidation_0-error:0.039551\n",
            "[758]\tvalidation_0-auc:0.846767\tvalidation_0-error:0.039551\n",
            "[759]\tvalidation_0-auc:0.846786\tvalidation_0-error:0.039551\n",
            "[760]\tvalidation_0-auc:0.846784\tvalidation_0-error:0.039551\n",
            "[761]\tvalidation_0-auc:0.846794\tvalidation_0-error:0.039551\n",
            "[762]\tvalidation_0-auc:0.846785\tvalidation_0-error:0.039551\n",
            "[763]\tvalidation_0-auc:0.846777\tvalidation_0-error:0.039551\n",
            "[764]\tvalidation_0-auc:0.846836\tvalidation_0-error:0.039551\n",
            "[765]\tvalidation_0-auc:0.846809\tvalidation_0-error:0.039551\n",
            "[766]\tvalidation_0-auc:0.846783\tvalidation_0-error:0.039551\n",
            "[767]\tvalidation_0-auc:0.84682\tvalidation_0-error:0.039551\n",
            "[768]\tvalidation_0-auc:0.846788\tvalidation_0-error:0.039551\n",
            "[769]\tvalidation_0-auc:0.846756\tvalidation_0-error:0.039551\n",
            "[770]\tvalidation_0-auc:0.846744\tvalidation_0-error:0.039551\n",
            "[771]\tvalidation_0-auc:0.846748\tvalidation_0-error:0.039551\n",
            "[772]\tvalidation_0-auc:0.846751\tvalidation_0-error:0.039551\n",
            "[773]\tvalidation_0-auc:0.84676\tvalidation_0-error:0.039551\n",
            "[774]\tvalidation_0-auc:0.846759\tvalidation_0-error:0.039551\n",
            "[775]\tvalidation_0-auc:0.846779\tvalidation_0-error:0.039551\n",
            "[776]\tvalidation_0-auc:0.84679\tvalidation_0-error:0.039551\n",
            "[777]\tvalidation_0-auc:0.846795\tvalidation_0-error:0.039551\n",
            "[778]\tvalidation_0-auc:0.846758\tvalidation_0-error:0.039551\n",
            "[779]\tvalidation_0-auc:0.846783\tvalidation_0-error:0.039551\n",
            "[780]\tvalidation_0-auc:0.846746\tvalidation_0-error:0.039551\n",
            "[781]\tvalidation_0-auc:0.846739\tvalidation_0-error:0.039551\n",
            "[782]\tvalidation_0-auc:0.846721\tvalidation_0-error:0.039551\n",
            "[783]\tvalidation_0-auc:0.846731\tvalidation_0-error:0.039551\n",
            "[784]\tvalidation_0-auc:0.84672\tvalidation_0-error:0.039551\n",
            "[785]\tvalidation_0-auc:0.846743\tvalidation_0-error:0.039551\n",
            "[786]\tvalidation_0-auc:0.846748\tvalidation_0-error:0.039551\n",
            "[787]\tvalidation_0-auc:0.846739\tvalidation_0-error:0.039551\n",
            "[788]\tvalidation_0-auc:0.846748\tvalidation_0-error:0.039551\n",
            "[789]\tvalidation_0-auc:0.846758\tvalidation_0-error:0.039551\n",
            "[790]\tvalidation_0-auc:0.846753\tvalidation_0-error:0.039551\n",
            "[791]\tvalidation_0-auc:0.84672\tvalidation_0-error:0.039551\n",
            "[792]\tvalidation_0-auc:0.846727\tvalidation_0-error:0.039551\n",
            "[793]\tvalidation_0-auc:0.846701\tvalidation_0-error:0.039551\n",
            "[794]\tvalidation_0-auc:0.846687\tvalidation_0-error:0.039551\n",
            "[795]\tvalidation_0-auc:0.84668\tvalidation_0-error:0.039551\n",
            "[796]\tvalidation_0-auc:0.846671\tvalidation_0-error:0.039551\n",
            "[797]\tvalidation_0-auc:0.846682\tvalidation_0-error:0.039551\n",
            "[798]\tvalidation_0-auc:0.846672\tvalidation_0-error:0.039551\n",
            "[799]\tvalidation_0-auc:0.846696\tvalidation_0-error:0.039551\n",
            "[800]\tvalidation_0-auc:0.846664\tvalidation_0-error:0.039551\n",
            "CPU times: user 14min 27s, sys: 1.18 s, total: 14min 28s\n",
            "Wall time: 7min 24s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
              "              colsample_bynode=1, colsample_bytree=0.8, gamma=0,\n",
              "              learning_rate=0.01, max_delta_step=0, max_depth=6,\n",
              "              min_child_weight=1, missing=None, n_estimators=801, n_jobs=-1,\n",
              "              nthread=-1, objective='binary:logistic', random_state=42,\n",
              "              reg_alpha=0.3, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
              "              silent=None, subsample=0.7, verbosity=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "9ZI4N9TM-V3D",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "0a72b93f-6e4b-4a46-b805-8521c9aa09a6"
      },
      "source": [
        "y_pred = model.predict_proba(X_val_200)[:,1]\n",
        "label=\"XGBoost Classifer(top 250)\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "i+=1\n",
        "result = plot_auc(y_val,y_pred,label)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t XGBoost Classifer(top 250) on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.133 and 0.847 for XGBoost Classifer(top 250) on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "LN2FW-s2-V3G",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "f7ebbe01-43a3-4ce4-de3d-b1acaea67e9b"
      },
      "source": [
        "result_df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137490</td>\n",
              "      <td>0.832512</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>XGBoost Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.133102</td>\n",
              "      <td>0.845507</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>LightGBM Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.132863</td>\n",
              "      <td>0.844955</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Random Forest Classifier (top 250)</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137054</td>\n",
              "      <td>0.834184</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>XGBoost Classifer(top 250)</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.132907</td>\n",
              "      <td>0.846664</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                Model Dataset  log_loss       auc\n",
              "0                  Logisic Regression  Normal  0.542129  0.812101\n",
              "1            Decision Tree Classifier  Normal  0.148082  0.779135\n",
              "2            Random Forest Classifier  Normal  0.137490  0.832512\n",
              "3                   XGBoost Classifer  Normal  0.133102  0.845507\n",
              "4                  LightGBM Classifer  Normal  0.132863  0.844955\n",
              "5  Random Forest Classifier (top 250)  Normal  0.137054  0.834184\n",
              "6          XGBoost Classifer(top 250)  Normal  0.132907  0.846664"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "mYW7_x88-V3J"
      },
      "source": [
        "# LightGBM Classifier (top 250)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "EYaevQf8-V3J",
        "colab": {}
      },
      "source": [
        "model = lgb.LGBMClassifier(n_jobs=-1,\n",
        "                      nthread=-1,\n",
        "                      scale_pos_weight=1,\n",
        "                      learning_rate=0.01,  \n",
        "                      colsample_bytree = 0.5,\n",
        "                      subsample = 0.8,\n",
        "                      objective='binary', \n",
        "                      n_estimators=681,\n",
        "                      reg_alpha = 1,\n",
        "                      max_depth=9, random_state=42\n",
        "                     )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "JhKg3-Lv-V3M",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "d41c8348-9c1b-4409-c370-ad5edad4c4f8"
      },
      "source": [
        "#https://towardsdatascience.com/fine-tuning-xgboost-in-python-like-a-boss-b4543ed8b1e\n",
        "\"\"\"\n",
        "manually changed each parameters 1 by 1 and compared the validation auc\n",
        "and selected the best parameter value\n",
        "\"\"\"\n",
        "eval_metric = ['auc','error']\n",
        "%time model.fit(X_train_200,y_train,eval_set=[(X_val_200,y_val)],eval_metric=eval_metric,verbose=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1]\tvalid_0's auc: 0.788085\tvalid_0's binary_logloss: 0.165817\n",
            "[2]\tvalid_0's auc: 0.810009\tvalid_0's binary_logloss: 0.165169\n",
            "[3]\tvalid_0's auc: 0.81412\tvalid_0's binary_logloss: 0.164621\n",
            "[4]\tvalid_0's auc: 0.822918\tvalid_0's binary_logloss: 0.163985\n",
            "[5]\tvalid_0's auc: 0.81925\tvalid_0's binary_logloss: 0.163402\n",
            "[6]\tvalid_0's auc: 0.824432\tvalid_0's binary_logloss: 0.162789\n",
            "[7]\tvalid_0's auc: 0.826867\tvalid_0's binary_logloss: 0.162199\n",
            "[8]\tvalid_0's auc: 0.825315\tvalid_0's binary_logloss: 0.161682\n",
            "[9]\tvalid_0's auc: 0.824712\tvalid_0's binary_logloss: 0.161165\n",
            "[10]\tvalid_0's auc: 0.825902\tvalid_0's binary_logloss: 0.160634\n",
            "[11]\tvalid_0's auc: 0.828092\tvalid_0's binary_logloss: 0.160092\n",
            "[12]\tvalid_0's auc: 0.830035\tvalid_0's binary_logloss: 0.159538\n",
            "[13]\tvalid_0's auc: 0.829749\tvalid_0's binary_logloss: 0.159093\n",
            "[14]\tvalid_0's auc: 0.829009\tvalid_0's binary_logloss: 0.158673\n",
            "[15]\tvalid_0's auc: 0.828538\tvalid_0's binary_logloss: 0.158255\n",
            "[16]\tvalid_0's auc: 0.829204\tvalid_0's binary_logloss: 0.157817\n",
            "[17]\tvalid_0's auc: 0.829008\tvalid_0's binary_logloss: 0.157421\n",
            "[18]\tvalid_0's auc: 0.828938\tvalid_0's binary_logloss: 0.15702\n",
            "[19]\tvalid_0's auc: 0.829481\tvalid_0's binary_logloss: 0.156577\n",
            "[20]\tvalid_0's auc: 0.828918\tvalid_0's binary_logloss: 0.156209\n",
            "[21]\tvalid_0's auc: 0.829746\tvalid_0's binary_logloss: 0.15582\n",
            "[22]\tvalid_0's auc: 0.830144\tvalid_0's binary_logloss: 0.15543\n",
            "[23]\tvalid_0's auc: 0.832163\tvalid_0's binary_logloss: 0.155028\n",
            "[24]\tvalid_0's auc: 0.831446\tvalid_0's binary_logloss: 0.154728\n",
            "[25]\tvalid_0's auc: 0.830601\tvalid_0's binary_logloss: 0.154453\n",
            "[26]\tvalid_0's auc: 0.831588\tvalid_0's binary_logloss: 0.154069\n",
            "[27]\tvalid_0's auc: 0.831613\tvalid_0's binary_logloss: 0.153782\n",
            "[28]\tvalid_0's auc: 0.832288\tvalid_0's binary_logloss: 0.153411\n",
            "[29]\tvalid_0's auc: 0.832927\tvalid_0's binary_logloss: 0.153065\n",
            "[30]\tvalid_0's auc: 0.832789\tvalid_0's binary_logloss: 0.152774\n",
            "[31]\tvalid_0's auc: 0.832075\tvalid_0's binary_logloss: 0.152511\n",
            "[32]\tvalid_0's auc: 0.832096\tvalid_0's binary_logloss: 0.152243\n",
            "[33]\tvalid_0's auc: 0.832758\tvalid_0's binary_logloss: 0.151943\n",
            "[34]\tvalid_0's auc: 0.832511\tvalid_0's binary_logloss: 0.151688\n",
            "[35]\tvalid_0's auc: 0.83329\tvalid_0's binary_logloss: 0.151377\n",
            "[36]\tvalid_0's auc: 0.833576\tvalid_0's binary_logloss: 0.151075\n",
            "[37]\tvalid_0's auc: 0.833581\tvalid_0's binary_logloss: 0.150814\n",
            "[38]\tvalid_0's auc: 0.833776\tvalid_0's binary_logloss: 0.150514\n",
            "[39]\tvalid_0's auc: 0.83381\tvalid_0's binary_logloss: 0.150287\n",
            "[40]\tvalid_0's auc: 0.833801\tvalid_0's binary_logloss: 0.150091\n",
            "[41]\tvalid_0's auc: 0.834155\tvalid_0's binary_logloss: 0.149821\n",
            "[42]\tvalid_0's auc: 0.834319\tvalid_0's binary_logloss: 0.1496\n",
            "[43]\tvalid_0's auc: 0.834603\tvalid_0's binary_logloss: 0.149335\n",
            "[44]\tvalid_0's auc: 0.834313\tvalid_0's binary_logloss: 0.149143\n",
            "[45]\tvalid_0's auc: 0.834533\tvalid_0's binary_logloss: 0.148907\n",
            "[46]\tvalid_0's auc: 0.83495\tvalid_0's binary_logloss: 0.148658\n",
            "[47]\tvalid_0's auc: 0.83514\tvalid_0's binary_logloss: 0.148413\n",
            "[48]\tvalid_0's auc: 0.834879\tvalid_0's binary_logloss: 0.14823\n",
            "[49]\tvalid_0's auc: 0.834884\tvalid_0's binary_logloss: 0.148033\n",
            "[50]\tvalid_0's auc: 0.834808\tvalid_0's binary_logloss: 0.14784\n",
            "[51]\tvalid_0's auc: 0.835\tvalid_0's binary_logloss: 0.147617\n",
            "[52]\tvalid_0's auc: 0.83473\tvalid_0's binary_logloss: 0.147442\n",
            "[53]\tvalid_0's auc: 0.83485\tvalid_0's binary_logloss: 0.147262\n",
            "[54]\tvalid_0's auc: 0.83511\tvalid_0's binary_logloss: 0.147047\n",
            "[55]\tvalid_0's auc: 0.834937\tvalid_0's binary_logloss: 0.146903\n",
            "[56]\tvalid_0's auc: 0.834589\tvalid_0's binary_logloss: 0.146763\n",
            "[57]\tvalid_0's auc: 0.834419\tvalid_0's binary_logloss: 0.146619\n",
            "[58]\tvalid_0's auc: 0.834539\tvalid_0's binary_logloss: 0.146402\n",
            "[59]\tvalid_0's auc: 0.83476\tvalid_0's binary_logloss: 0.146245\n",
            "[60]\tvalid_0's auc: 0.835153\tvalid_0's binary_logloss: 0.146042\n",
            "[61]\tvalid_0's auc: 0.83567\tvalid_0's binary_logloss: 0.145832\n",
            "[62]\tvalid_0's auc: 0.83599\tvalid_0's binary_logloss: 0.145641\n",
            "[63]\tvalid_0's auc: 0.836226\tvalid_0's binary_logloss: 0.145467\n",
            "[64]\tvalid_0's auc: 0.836278\tvalid_0's binary_logloss: 0.145306\n",
            "[65]\tvalid_0's auc: 0.83649\tvalid_0's binary_logloss: 0.145134\n",
            "[66]\tvalid_0's auc: 0.836605\tvalid_0's binary_logloss: 0.145\n",
            "[67]\tvalid_0's auc: 0.836746\tvalid_0's binary_logloss: 0.144843\n",
            "[68]\tvalid_0's auc: 0.836701\tvalid_0's binary_logloss: 0.144694\n",
            "[69]\tvalid_0's auc: 0.836806\tvalid_0's binary_logloss: 0.144564\n",
            "[70]\tvalid_0's auc: 0.836701\tvalid_0's binary_logloss: 0.144439\n",
            "[71]\tvalid_0's auc: 0.836731\tvalid_0's binary_logloss: 0.144287\n",
            "[72]\tvalid_0's auc: 0.837086\tvalid_0's binary_logloss: 0.144139\n",
            "[73]\tvalid_0's auc: 0.837197\tvalid_0's binary_logloss: 0.143993\n",
            "[74]\tvalid_0's auc: 0.837201\tvalid_0's binary_logloss: 0.143854\n",
            "[75]\tvalid_0's auc: 0.837417\tvalid_0's binary_logloss: 0.1437\n",
            "[76]\tvalid_0's auc: 0.837663\tvalid_0's binary_logloss: 0.143538\n",
            "[77]\tvalid_0's auc: 0.837717\tvalid_0's binary_logloss: 0.143433\n",
            "[78]\tvalid_0's auc: 0.837777\tvalid_0's binary_logloss: 0.1433\n",
            "[79]\tvalid_0's auc: 0.837647\tvalid_0's binary_logloss: 0.143157\n",
            "[80]\tvalid_0's auc: 0.837718\tvalid_0's binary_logloss: 0.143035\n",
            "[81]\tvalid_0's auc: 0.837805\tvalid_0's binary_logloss: 0.142923\n",
            "[82]\tvalid_0's auc: 0.837658\tvalid_0's binary_logloss: 0.14283\n",
            "[83]\tvalid_0's auc: 0.837576\tvalid_0's binary_logloss: 0.142743\n",
            "[84]\tvalid_0's auc: 0.837524\tvalid_0's binary_logloss: 0.142623\n",
            "[85]\tvalid_0's auc: 0.83763\tvalid_0's binary_logloss: 0.142495\n",
            "[86]\tvalid_0's auc: 0.837832\tvalid_0's binary_logloss: 0.142364\n",
            "[87]\tvalid_0's auc: 0.838042\tvalid_0's binary_logloss: 0.142221\n",
            "[88]\tvalid_0's auc: 0.838221\tvalid_0's binary_logloss: 0.142101\n",
            "[89]\tvalid_0's auc: 0.838228\tvalid_0's binary_logloss: 0.141986\n",
            "[90]\tvalid_0's auc: 0.838395\tvalid_0's binary_logloss: 0.141869\n",
            "[91]\tvalid_0's auc: 0.838352\tvalid_0's binary_logloss: 0.14178\n",
            "[92]\tvalid_0's auc: 0.838316\tvalid_0's binary_logloss: 0.141685\n",
            "[93]\tvalid_0's auc: 0.838587\tvalid_0's binary_logloss: 0.14157\n",
            "[94]\tvalid_0's auc: 0.838512\tvalid_0's binary_logloss: 0.141495\n",
            "[95]\tvalid_0's auc: 0.838468\tvalid_0's binary_logloss: 0.141394\n",
            "[96]\tvalid_0's auc: 0.838631\tvalid_0's binary_logloss: 0.141277\n",
            "[97]\tvalid_0's auc: 0.838526\tvalid_0's binary_logloss: 0.141207\n",
            "[98]\tvalid_0's auc: 0.838579\tvalid_0's binary_logloss: 0.141107\n",
            "[99]\tvalid_0's auc: 0.838762\tvalid_0's binary_logloss: 0.141\n",
            "[100]\tvalid_0's auc: 0.838792\tvalid_0's binary_logloss: 0.140894\n",
            "[101]\tvalid_0's auc: 0.839\tvalid_0's binary_logloss: 0.140779\n",
            "[102]\tvalid_0's auc: 0.838938\tvalid_0's binary_logloss: 0.140711\n",
            "[103]\tvalid_0's auc: 0.839227\tvalid_0's binary_logloss: 0.140616\n",
            "[104]\tvalid_0's auc: 0.839228\tvalid_0's binary_logloss: 0.140543\n",
            "[105]\tvalid_0's auc: 0.83925\tvalid_0's binary_logloss: 0.140455\n",
            "[106]\tvalid_0's auc: 0.839388\tvalid_0's binary_logloss: 0.14036\n",
            "[107]\tvalid_0's auc: 0.839511\tvalid_0's binary_logloss: 0.140285\n",
            "[108]\tvalid_0's auc: 0.839722\tvalid_0's binary_logloss: 0.140178\n",
            "[109]\tvalid_0's auc: 0.839726\tvalid_0's binary_logloss: 0.140105\n",
            "[110]\tvalid_0's auc: 0.839718\tvalid_0's binary_logloss: 0.140046\n",
            "[111]\tvalid_0's auc: 0.839688\tvalid_0's binary_logloss: 0.139991\n",
            "[112]\tvalid_0's auc: 0.839529\tvalid_0's binary_logloss: 0.139936\n",
            "[113]\tvalid_0's auc: 0.839512\tvalid_0's binary_logloss: 0.139882\n",
            "[114]\tvalid_0's auc: 0.83965\tvalid_0's binary_logloss: 0.139791\n",
            "[115]\tvalid_0's auc: 0.839546\tvalid_0's binary_logloss: 0.139732\n",
            "[116]\tvalid_0's auc: 0.839471\tvalid_0's binary_logloss: 0.139669\n",
            "[117]\tvalid_0's auc: 0.839461\tvalid_0's binary_logloss: 0.139613\n",
            "[118]\tvalid_0's auc: 0.839675\tvalid_0's binary_logloss: 0.139523\n",
            "[119]\tvalid_0's auc: 0.839776\tvalid_0's binary_logloss: 0.139456\n",
            "[120]\tvalid_0's auc: 0.839781\tvalid_0's binary_logloss: 0.139381\n",
            "[121]\tvalid_0's auc: 0.83974\tvalid_0's binary_logloss: 0.139328\n",
            "[122]\tvalid_0's auc: 0.839862\tvalid_0's binary_logloss: 0.139246\n",
            "[123]\tvalid_0's auc: 0.840021\tvalid_0's binary_logloss: 0.139159\n",
            "[124]\tvalid_0's auc: 0.840033\tvalid_0's binary_logloss: 0.139095\n",
            "[125]\tvalid_0's auc: 0.840281\tvalid_0's binary_logloss: 0.139015\n",
            "[126]\tvalid_0's auc: 0.840413\tvalid_0's binary_logloss: 0.138936\n",
            "[127]\tvalid_0's auc: 0.840411\tvalid_0's binary_logloss: 0.138872\n",
            "[128]\tvalid_0's auc: 0.840496\tvalid_0's binary_logloss: 0.138806\n",
            "[129]\tvalid_0's auc: 0.840567\tvalid_0's binary_logloss: 0.138747\n",
            "[130]\tvalid_0's auc: 0.840685\tvalid_0's binary_logloss: 0.138687\n",
            "[131]\tvalid_0's auc: 0.84072\tvalid_0's binary_logloss: 0.138632\n",
            "[132]\tvalid_0's auc: 0.840792\tvalid_0's binary_logloss: 0.138583\n",
            "[133]\tvalid_0's auc: 0.84087\tvalid_0's binary_logloss: 0.138509\n",
            "[134]\tvalid_0's auc: 0.840937\tvalid_0's binary_logloss: 0.138441\n",
            "[135]\tvalid_0's auc: 0.840951\tvalid_0's binary_logloss: 0.138397\n",
            "[136]\tvalid_0's auc: 0.840983\tvalid_0's binary_logloss: 0.138345\n",
            "[137]\tvalid_0's auc: 0.841042\tvalid_0's binary_logloss: 0.138275\n",
            "[138]\tvalid_0's auc: 0.841068\tvalid_0's binary_logloss: 0.138227\n",
            "[139]\tvalid_0's auc: 0.841047\tvalid_0's binary_logloss: 0.138182\n",
            "[140]\tvalid_0's auc: 0.841032\tvalid_0's binary_logloss: 0.138143\n",
            "[141]\tvalid_0's auc: 0.840981\tvalid_0's binary_logloss: 0.138102\n",
            "[142]\tvalid_0's auc: 0.840969\tvalid_0's binary_logloss: 0.138044\n",
            "[143]\tvalid_0's auc: 0.840988\tvalid_0's binary_logloss: 0.137979\n",
            "[144]\tvalid_0's auc: 0.841055\tvalid_0's binary_logloss: 0.13793\n",
            "[145]\tvalid_0's auc: 0.841037\tvalid_0's binary_logloss: 0.13787\n",
            "[146]\tvalid_0's auc: 0.841119\tvalid_0's binary_logloss: 0.13781\n",
            "[147]\tvalid_0's auc: 0.841169\tvalid_0's binary_logloss: 0.137755\n",
            "[148]\tvalid_0's auc: 0.841171\tvalid_0's binary_logloss: 0.137717\n",
            "[149]\tvalid_0's auc: 0.84125\tvalid_0's binary_logloss: 0.137657\n",
            "[150]\tvalid_0's auc: 0.841302\tvalid_0's binary_logloss: 0.137595\n",
            "[151]\tvalid_0's auc: 0.841359\tvalid_0's binary_logloss: 0.137542\n",
            "[152]\tvalid_0's auc: 0.841405\tvalid_0's binary_logloss: 0.137504\n",
            "[153]\tvalid_0's auc: 0.8414\tvalid_0's binary_logloss: 0.137472\n",
            "[154]\tvalid_0's auc: 0.841428\tvalid_0's binary_logloss: 0.137422\n",
            "[155]\tvalid_0's auc: 0.841425\tvalid_0's binary_logloss: 0.137386\n",
            "[156]\tvalid_0's auc: 0.841367\tvalid_0's binary_logloss: 0.137351\n",
            "[157]\tvalid_0's auc: 0.841368\tvalid_0's binary_logloss: 0.137295\n",
            "[158]\tvalid_0's auc: 0.841411\tvalid_0's binary_logloss: 0.137263\n",
            "[159]\tvalid_0's auc: 0.841416\tvalid_0's binary_logloss: 0.137213\n",
            "[160]\tvalid_0's auc: 0.84141\tvalid_0's binary_logloss: 0.137171\n",
            "[161]\tvalid_0's auc: 0.841488\tvalid_0's binary_logloss: 0.13713\n",
            "[162]\tvalid_0's auc: 0.841478\tvalid_0's binary_logloss: 0.137094\n",
            "[163]\tvalid_0's auc: 0.841501\tvalid_0's binary_logloss: 0.137041\n",
            "[164]\tvalid_0's auc: 0.841511\tvalid_0's binary_logloss: 0.137005\n",
            "[165]\tvalid_0's auc: 0.84156\tvalid_0's binary_logloss: 0.13697\n",
            "[166]\tvalid_0's auc: 0.841579\tvalid_0's binary_logloss: 0.136926\n",
            "[167]\tvalid_0's auc: 0.841725\tvalid_0's binary_logloss: 0.13687\n",
            "[168]\tvalid_0's auc: 0.841721\tvalid_0's binary_logloss: 0.136845\n",
            "[169]\tvalid_0's auc: 0.841765\tvalid_0's binary_logloss: 0.1368\n",
            "[170]\tvalid_0's auc: 0.841825\tvalid_0's binary_logloss: 0.136751\n",
            "[171]\tvalid_0's auc: 0.841849\tvalid_0's binary_logloss: 0.136704\n",
            "[172]\tvalid_0's auc: 0.841859\tvalid_0's binary_logloss: 0.136664\n",
            "[173]\tvalid_0's auc: 0.841846\tvalid_0's binary_logloss: 0.136623\n",
            "[174]\tvalid_0's auc: 0.841843\tvalid_0's binary_logloss: 0.136593\n",
            "[175]\tvalid_0's auc: 0.841834\tvalid_0's binary_logloss: 0.136553\n",
            "[176]\tvalid_0's auc: 0.841769\tvalid_0's binary_logloss: 0.136515\n",
            "[177]\tvalid_0's auc: 0.841776\tvalid_0's binary_logloss: 0.136476\n",
            "[178]\tvalid_0's auc: 0.841816\tvalid_0's binary_logloss: 0.136454\n",
            "[179]\tvalid_0's auc: 0.84168\tvalid_0's binary_logloss: 0.136433\n",
            "[180]\tvalid_0's auc: 0.841737\tvalid_0's binary_logloss: 0.136398\n",
            "[181]\tvalid_0's auc: 0.841801\tvalid_0's binary_logloss: 0.13637\n",
            "[182]\tvalid_0's auc: 0.84184\tvalid_0's binary_logloss: 0.136336\n",
            "[183]\tvalid_0's auc: 0.841856\tvalid_0's binary_logloss: 0.136316\n",
            "[184]\tvalid_0's auc: 0.841892\tvalid_0's binary_logloss: 0.136287\n",
            "[185]\tvalid_0's auc: 0.841937\tvalid_0's binary_logloss: 0.136262\n",
            "[186]\tvalid_0's auc: 0.841911\tvalid_0's binary_logloss: 0.136237\n",
            "[187]\tvalid_0's auc: 0.841835\tvalid_0's binary_logloss: 0.136211\n",
            "[188]\tvalid_0's auc: 0.841901\tvalid_0's binary_logloss: 0.136173\n",
            "[189]\tvalid_0's auc: 0.841884\tvalid_0's binary_logloss: 0.136158\n",
            "[190]\tvalid_0's auc: 0.841871\tvalid_0's binary_logloss: 0.136139\n",
            "[191]\tvalid_0's auc: 0.841904\tvalid_0's binary_logloss: 0.136101\n",
            "[192]\tvalid_0's auc: 0.841966\tvalid_0's binary_logloss: 0.136085\n",
            "[193]\tvalid_0's auc: 0.841998\tvalid_0's binary_logloss: 0.136063\n",
            "[194]\tvalid_0's auc: 0.842304\tvalid_0's binary_logloss: 0.136017\n",
            "[195]\tvalid_0's auc: 0.842369\tvalid_0's binary_logloss: 0.135981\n",
            "[196]\tvalid_0's auc: 0.84236\tvalid_0's binary_logloss: 0.135967\n",
            "[197]\tvalid_0's auc: 0.842356\tvalid_0's binary_logloss: 0.135956\n",
            "[198]\tvalid_0's auc: 0.842379\tvalid_0's binary_logloss: 0.135927\n",
            "[199]\tvalid_0's auc: 0.842414\tvalid_0's binary_logloss: 0.1359\n",
            "[200]\tvalid_0's auc: 0.842356\tvalid_0's binary_logloss: 0.13588\n",
            "[201]\tvalid_0's auc: 0.842451\tvalid_0's binary_logloss: 0.135857\n",
            "[202]\tvalid_0's auc: 0.842396\tvalid_0's binary_logloss: 0.135841\n",
            "[203]\tvalid_0's auc: 0.84247\tvalid_0's binary_logloss: 0.13581\n",
            "[204]\tvalid_0's auc: 0.842511\tvalid_0's binary_logloss: 0.135784\n",
            "[205]\tvalid_0's auc: 0.842397\tvalid_0's binary_logloss: 0.135767\n",
            "[206]\tvalid_0's auc: 0.84243\tvalid_0's binary_logloss: 0.135745\n",
            "[207]\tvalid_0's auc: 0.842459\tvalid_0's binary_logloss: 0.135726\n",
            "[208]\tvalid_0's auc: 0.842497\tvalid_0's binary_logloss: 0.135692\n",
            "[209]\tvalid_0's auc: 0.84244\tvalid_0's binary_logloss: 0.13568\n",
            "[210]\tvalid_0's auc: 0.842541\tvalid_0's binary_logloss: 0.13564\n",
            "[211]\tvalid_0's auc: 0.842555\tvalid_0's binary_logloss: 0.135612\n",
            "[212]\tvalid_0's auc: 0.842548\tvalid_0's binary_logloss: 0.135596\n",
            "[213]\tvalid_0's auc: 0.842584\tvalid_0's binary_logloss: 0.13556\n",
            "[214]\tvalid_0's auc: 0.8426\tvalid_0's binary_logloss: 0.135532\n",
            "[215]\tvalid_0's auc: 0.842605\tvalid_0's binary_logloss: 0.135503\n",
            "[216]\tvalid_0's auc: 0.842612\tvalid_0's binary_logloss: 0.135482\n",
            "[217]\tvalid_0's auc: 0.842567\tvalid_0's binary_logloss: 0.135467\n",
            "[218]\tvalid_0's auc: 0.842581\tvalid_0's binary_logloss: 0.135436\n",
            "[219]\tvalid_0's auc: 0.842645\tvalid_0's binary_logloss: 0.135418\n",
            "[220]\tvalid_0's auc: 0.842614\tvalid_0's binary_logloss: 0.135399\n",
            "[221]\tvalid_0's auc: 0.842615\tvalid_0's binary_logloss: 0.135383\n",
            "[222]\tvalid_0's auc: 0.842681\tvalid_0's binary_logloss: 0.135354\n",
            "[223]\tvalid_0's auc: 0.842714\tvalid_0's binary_logloss: 0.135327\n",
            "[224]\tvalid_0's auc: 0.842753\tvalid_0's binary_logloss: 0.135306\n",
            "[225]\tvalid_0's auc: 0.842743\tvalid_0's binary_logloss: 0.135281\n",
            "[226]\tvalid_0's auc: 0.842742\tvalid_0's binary_logloss: 0.135261\n",
            "[227]\tvalid_0's auc: 0.842728\tvalid_0's binary_logloss: 0.135247\n",
            "[228]\tvalid_0's auc: 0.842789\tvalid_0's binary_logloss: 0.135235\n",
            "[229]\tvalid_0's auc: 0.842839\tvalid_0's binary_logloss: 0.135209\n",
            "[230]\tvalid_0's auc: 0.842858\tvalid_0's binary_logloss: 0.135184\n",
            "[231]\tvalid_0's auc: 0.842947\tvalid_0's binary_logloss: 0.135152\n",
            "[232]\tvalid_0's auc: 0.842959\tvalid_0's binary_logloss: 0.135124\n",
            "[233]\tvalid_0's auc: 0.842995\tvalid_0's binary_logloss: 0.135099\n",
            "[234]\tvalid_0's auc: 0.842984\tvalid_0's binary_logloss: 0.13508\n",
            "[235]\tvalid_0's auc: 0.84302\tvalid_0's binary_logloss: 0.135055\n",
            "[236]\tvalid_0's auc: 0.843043\tvalid_0's binary_logloss: 0.135032\n",
            "[237]\tvalid_0's auc: 0.843062\tvalid_0's binary_logloss: 0.135015\n",
            "[238]\tvalid_0's auc: 0.84319\tvalid_0's binary_logloss: 0.134983\n",
            "[239]\tvalid_0's auc: 0.843216\tvalid_0's binary_logloss: 0.134964\n",
            "[240]\tvalid_0's auc: 0.843261\tvalid_0's binary_logloss: 0.134939\n",
            "[241]\tvalid_0's auc: 0.843262\tvalid_0's binary_logloss: 0.134929\n",
            "[242]\tvalid_0's auc: 0.843253\tvalid_0's binary_logloss: 0.134908\n",
            "[243]\tvalid_0's auc: 0.843177\tvalid_0's binary_logloss: 0.134907\n",
            "[244]\tvalid_0's auc: 0.843222\tvalid_0's binary_logloss: 0.134894\n",
            "[245]\tvalid_0's auc: 0.843229\tvalid_0's binary_logloss: 0.134881\n",
            "[246]\tvalid_0's auc: 0.843188\tvalid_0's binary_logloss: 0.13487\n",
            "[247]\tvalid_0's auc: 0.843222\tvalid_0's binary_logloss: 0.134862\n",
            "[248]\tvalid_0's auc: 0.843209\tvalid_0's binary_logloss: 0.13485\n",
            "[249]\tvalid_0's auc: 0.84325\tvalid_0's binary_logloss: 0.134829\n",
            "[250]\tvalid_0's auc: 0.84323\tvalid_0's binary_logloss: 0.134825\n",
            "[251]\tvalid_0's auc: 0.84322\tvalid_0's binary_logloss: 0.134807\n",
            "[252]\tvalid_0's auc: 0.843266\tvalid_0's binary_logloss: 0.134781\n",
            "[253]\tvalid_0's auc: 0.84328\tvalid_0's binary_logloss: 0.134753\n",
            "[254]\tvalid_0's auc: 0.843262\tvalid_0's binary_logloss: 0.134743\n",
            "[255]\tvalid_0's auc: 0.843351\tvalid_0's binary_logloss: 0.134723\n",
            "[256]\tvalid_0's auc: 0.843323\tvalid_0's binary_logloss: 0.134713\n",
            "[257]\tvalid_0's auc: 0.843301\tvalid_0's binary_logloss: 0.134708\n",
            "[258]\tvalid_0's auc: 0.843271\tvalid_0's binary_logloss: 0.134704\n",
            "[259]\tvalid_0's auc: 0.843232\tvalid_0's binary_logloss: 0.134689\n",
            "[260]\tvalid_0's auc: 0.843245\tvalid_0's binary_logloss: 0.134668\n",
            "[261]\tvalid_0's auc: 0.843288\tvalid_0's binary_logloss: 0.134647\n",
            "[262]\tvalid_0's auc: 0.843292\tvalid_0's binary_logloss: 0.134631\n",
            "[263]\tvalid_0's auc: 0.843307\tvalid_0's binary_logloss: 0.134618\n",
            "[264]\tvalid_0's auc: 0.843275\tvalid_0's binary_logloss: 0.134606\n",
            "[265]\tvalid_0's auc: 0.843205\tvalid_0's binary_logloss: 0.134609\n",
            "[266]\tvalid_0's auc: 0.843227\tvalid_0's binary_logloss: 0.134596\n",
            "[267]\tvalid_0's auc: 0.843292\tvalid_0's binary_logloss: 0.134588\n",
            "[268]\tvalid_0's auc: 0.843238\tvalid_0's binary_logloss: 0.134582\n",
            "[269]\tvalid_0's auc: 0.84329\tvalid_0's binary_logloss: 0.134552\n",
            "[270]\tvalid_0's auc: 0.843332\tvalid_0's binary_logloss: 0.13454\n",
            "[271]\tvalid_0's auc: 0.843357\tvalid_0's binary_logloss: 0.134518\n",
            "[272]\tvalid_0's auc: 0.843365\tvalid_0's binary_logloss: 0.134513\n",
            "[273]\tvalid_0's auc: 0.843293\tvalid_0's binary_logloss: 0.134513\n",
            "[274]\tvalid_0's auc: 0.843238\tvalid_0's binary_logloss: 0.134513\n",
            "[275]\tvalid_0's auc: 0.843289\tvalid_0's binary_logloss: 0.134495\n",
            "[276]\tvalid_0's auc: 0.843312\tvalid_0's binary_logloss: 0.134474\n",
            "[277]\tvalid_0's auc: 0.843289\tvalid_0's binary_logloss: 0.134471\n",
            "[278]\tvalid_0's auc: 0.843277\tvalid_0's binary_logloss: 0.134466\n",
            "[279]\tvalid_0's auc: 0.843311\tvalid_0's binary_logloss: 0.134456\n",
            "[280]\tvalid_0's auc: 0.843348\tvalid_0's binary_logloss: 0.13444\n",
            "[281]\tvalid_0's auc: 0.843296\tvalid_0's binary_logloss: 0.13444\n",
            "[282]\tvalid_0's auc: 0.843341\tvalid_0's binary_logloss: 0.13442\n",
            "[283]\tvalid_0's auc: 0.84333\tvalid_0's binary_logloss: 0.134407\n",
            "[284]\tvalid_0's auc: 0.843343\tvalid_0's binary_logloss: 0.134399\n",
            "[285]\tvalid_0's auc: 0.843327\tvalid_0's binary_logloss: 0.134395\n",
            "[286]\tvalid_0's auc: 0.843386\tvalid_0's binary_logloss: 0.134383\n",
            "[287]\tvalid_0's auc: 0.843358\tvalid_0's binary_logloss: 0.134378\n",
            "[288]\tvalid_0's auc: 0.843369\tvalid_0's binary_logloss: 0.134364\n",
            "[289]\tvalid_0's auc: 0.843406\tvalid_0's binary_logloss: 0.134351\n",
            "[290]\tvalid_0's auc: 0.843407\tvalid_0's binary_logloss: 0.134338\n",
            "[291]\tvalid_0's auc: 0.843357\tvalid_0's binary_logloss: 0.13434\n",
            "[292]\tvalid_0's auc: 0.843311\tvalid_0's binary_logloss: 0.134337\n",
            "[293]\tvalid_0's auc: 0.843317\tvalid_0's binary_logloss: 0.134326\n",
            "[294]\tvalid_0's auc: 0.843399\tvalid_0's binary_logloss: 0.13431\n",
            "[295]\tvalid_0's auc: 0.843382\tvalid_0's binary_logloss: 0.134309\n",
            "[296]\tvalid_0's auc: 0.843376\tvalid_0's binary_logloss: 0.134304\n",
            "[297]\tvalid_0's auc: 0.843401\tvalid_0's binary_logloss: 0.134293\n",
            "[298]\tvalid_0's auc: 0.843433\tvalid_0's binary_logloss: 0.13429\n",
            "[299]\tvalid_0's auc: 0.843441\tvalid_0's binary_logloss: 0.134282\n",
            "[300]\tvalid_0's auc: 0.843486\tvalid_0's binary_logloss: 0.134268\n",
            "[301]\tvalid_0's auc: 0.843457\tvalid_0's binary_logloss: 0.134263\n",
            "[302]\tvalid_0's auc: 0.843438\tvalid_0's binary_logloss: 0.134263\n",
            "[303]\tvalid_0's auc: 0.843441\tvalid_0's binary_logloss: 0.134253\n",
            "[304]\tvalid_0's auc: 0.843487\tvalid_0's binary_logloss: 0.134243\n",
            "[305]\tvalid_0's auc: 0.843452\tvalid_0's binary_logloss: 0.134233\n",
            "[306]\tvalid_0's auc: 0.843442\tvalid_0's binary_logloss: 0.134227\n",
            "[307]\tvalid_0's auc: 0.843409\tvalid_0's binary_logloss: 0.134223\n",
            "[308]\tvalid_0's auc: 0.843414\tvalid_0's binary_logloss: 0.134213\n",
            "[309]\tvalid_0's auc: 0.843461\tvalid_0's binary_logloss: 0.134199\n",
            "[310]\tvalid_0's auc: 0.843488\tvalid_0's binary_logloss: 0.134186\n",
            "[311]\tvalid_0's auc: 0.843487\tvalid_0's binary_logloss: 0.134181\n",
            "[312]\tvalid_0's auc: 0.843503\tvalid_0's binary_logloss: 0.134169\n",
            "[313]\tvalid_0's auc: 0.843493\tvalid_0's binary_logloss: 0.134166\n",
            "[314]\tvalid_0's auc: 0.843481\tvalid_0's binary_logloss: 0.134162\n",
            "[315]\tvalid_0's auc: 0.843475\tvalid_0's binary_logloss: 0.134154\n",
            "[316]\tvalid_0's auc: 0.843385\tvalid_0's binary_logloss: 0.134146\n",
            "[317]\tvalid_0's auc: 0.843379\tvalid_0's binary_logloss: 0.134142\n",
            "[318]\tvalid_0's auc: 0.843407\tvalid_0's binary_logloss: 0.134133\n",
            "[319]\tvalid_0's auc: 0.843438\tvalid_0's binary_logloss: 0.134122\n",
            "[320]\tvalid_0's auc: 0.843435\tvalid_0's binary_logloss: 0.134115\n",
            "[321]\tvalid_0's auc: 0.843441\tvalid_0's binary_logloss: 0.134105\n",
            "[322]\tvalid_0's auc: 0.843422\tvalid_0's binary_logloss: 0.134103\n",
            "[323]\tvalid_0's auc: 0.843421\tvalid_0's binary_logloss: 0.134101\n",
            "[324]\tvalid_0's auc: 0.843478\tvalid_0's binary_logloss: 0.134089\n",
            "[325]\tvalid_0's auc: 0.843502\tvalid_0's binary_logloss: 0.134078\n",
            "[326]\tvalid_0's auc: 0.843506\tvalid_0's binary_logloss: 0.134065\n",
            "[327]\tvalid_0's auc: 0.843536\tvalid_0's binary_logloss: 0.134053\n",
            "[328]\tvalid_0's auc: 0.84354\tvalid_0's binary_logloss: 0.134044\n",
            "[329]\tvalid_0's auc: 0.843571\tvalid_0's binary_logloss: 0.134035\n",
            "[330]\tvalid_0's auc: 0.843579\tvalid_0's binary_logloss: 0.134032\n",
            "[331]\tvalid_0's auc: 0.84358\tvalid_0's binary_logloss: 0.134023\n",
            "[332]\tvalid_0's auc: 0.843488\tvalid_0's binary_logloss: 0.134024\n",
            "[333]\tvalid_0's auc: 0.843519\tvalid_0's binary_logloss: 0.134015\n",
            "[334]\tvalid_0's auc: 0.843449\tvalid_0's binary_logloss: 0.134012\n",
            "[335]\tvalid_0's auc: 0.843481\tvalid_0's binary_logloss: 0.133997\n",
            "[336]\tvalid_0's auc: 0.843519\tvalid_0's binary_logloss: 0.133986\n",
            "[337]\tvalid_0's auc: 0.843501\tvalid_0's binary_logloss: 0.133982\n",
            "[338]\tvalid_0's auc: 0.843505\tvalid_0's binary_logloss: 0.133974\n",
            "[339]\tvalid_0's auc: 0.843514\tvalid_0's binary_logloss: 0.133969\n",
            "[340]\tvalid_0's auc: 0.843573\tvalid_0's binary_logloss: 0.133959\n",
            "[341]\tvalid_0's auc: 0.843576\tvalid_0's binary_logloss: 0.13395\n",
            "[342]\tvalid_0's auc: 0.843611\tvalid_0's binary_logloss: 0.133936\n",
            "[343]\tvalid_0's auc: 0.843641\tvalid_0's binary_logloss: 0.133932\n",
            "[344]\tvalid_0's auc: 0.843674\tvalid_0's binary_logloss: 0.133926\n",
            "[345]\tvalid_0's auc: 0.843713\tvalid_0's binary_logloss: 0.133913\n",
            "[346]\tvalid_0's auc: 0.843727\tvalid_0's binary_logloss: 0.133909\n",
            "[347]\tvalid_0's auc: 0.843728\tvalid_0's binary_logloss: 0.133902\n",
            "[348]\tvalid_0's auc: 0.843724\tvalid_0's binary_logloss: 0.133898\n",
            "[349]\tvalid_0's auc: 0.843705\tvalid_0's binary_logloss: 0.133898\n",
            "[350]\tvalid_0's auc: 0.843666\tvalid_0's binary_logloss: 0.133894\n",
            "[351]\tvalid_0's auc: 0.843712\tvalid_0's binary_logloss: 0.133879\n",
            "[352]\tvalid_0's auc: 0.84373\tvalid_0's binary_logloss: 0.133874\n",
            "[353]\tvalid_0's auc: 0.84376\tvalid_0's binary_logloss: 0.133864\n",
            "[354]\tvalid_0's auc: 0.843781\tvalid_0's binary_logloss: 0.133861\n",
            "[355]\tvalid_0's auc: 0.843794\tvalid_0's binary_logloss: 0.133855\n",
            "[356]\tvalid_0's auc: 0.843765\tvalid_0's binary_logloss: 0.133854\n",
            "[357]\tvalid_0's auc: 0.843831\tvalid_0's binary_logloss: 0.133838\n",
            "[358]\tvalid_0's auc: 0.843801\tvalid_0's binary_logloss: 0.133839\n",
            "[359]\tvalid_0's auc: 0.843765\tvalid_0's binary_logloss: 0.133837\n",
            "[360]\tvalid_0's auc: 0.843814\tvalid_0's binary_logloss: 0.133824\n",
            "[361]\tvalid_0's auc: 0.843942\tvalid_0's binary_logloss: 0.133813\n",
            "[362]\tvalid_0's auc: 0.843986\tvalid_0's binary_logloss: 0.133801\n",
            "[363]\tvalid_0's auc: 0.843994\tvalid_0's binary_logloss: 0.133793\n",
            "[364]\tvalid_0's auc: 0.843996\tvalid_0's binary_logloss: 0.133794\n",
            "[365]\tvalid_0's auc: 0.844013\tvalid_0's binary_logloss: 0.133789\n",
            "[366]\tvalid_0's auc: 0.844019\tvalid_0's binary_logloss: 0.13378\n",
            "[367]\tvalid_0's auc: 0.84408\tvalid_0's binary_logloss: 0.13377\n",
            "[368]\tvalid_0's auc: 0.844068\tvalid_0's binary_logloss: 0.13377\n",
            "[369]\tvalid_0's auc: 0.844001\tvalid_0's binary_logloss: 0.133776\n",
            "[370]\tvalid_0's auc: 0.84405\tvalid_0's binary_logloss: 0.133771\n",
            "[371]\tvalid_0's auc: 0.844074\tvalid_0's binary_logloss: 0.133761\n",
            "[372]\tvalid_0's auc: 0.844074\tvalid_0's binary_logloss: 0.133753\n",
            "[373]\tvalid_0's auc: 0.844085\tvalid_0's binary_logloss: 0.133755\n",
            "[374]\tvalid_0's auc: 0.844127\tvalid_0's binary_logloss: 0.133742\n",
            "[375]\tvalid_0's auc: 0.844121\tvalid_0's binary_logloss: 0.133738\n",
            "[376]\tvalid_0's auc: 0.844149\tvalid_0's binary_logloss: 0.133728\n",
            "[377]\tvalid_0's auc: 0.844129\tvalid_0's binary_logloss: 0.133727\n",
            "[378]\tvalid_0's auc: 0.844135\tvalid_0's binary_logloss: 0.133723\n",
            "[379]\tvalid_0's auc: 0.844129\tvalid_0's binary_logloss: 0.133717\n",
            "[380]\tvalid_0's auc: 0.844125\tvalid_0's binary_logloss: 0.133716\n",
            "[381]\tvalid_0's auc: 0.844144\tvalid_0's binary_logloss: 0.133706\n",
            "[382]\tvalid_0's auc: 0.844165\tvalid_0's binary_logloss: 0.133703\n",
            "[383]\tvalid_0's auc: 0.844166\tvalid_0's binary_logloss: 0.1337\n",
            "[384]\tvalid_0's auc: 0.844166\tvalid_0's binary_logloss: 0.13369\n",
            "[385]\tvalid_0's auc: 0.844188\tvalid_0's binary_logloss: 0.133682\n",
            "[386]\tvalid_0's auc: 0.844183\tvalid_0's binary_logloss: 0.133678\n",
            "[387]\tvalid_0's auc: 0.844201\tvalid_0's binary_logloss: 0.133669\n",
            "[388]\tvalid_0's auc: 0.844197\tvalid_0's binary_logloss: 0.133663\n",
            "[389]\tvalid_0's auc: 0.844176\tvalid_0's binary_logloss: 0.133664\n",
            "[390]\tvalid_0's auc: 0.844132\tvalid_0's binary_logloss: 0.133659\n",
            "[391]\tvalid_0's auc: 0.84414\tvalid_0's binary_logloss: 0.133651\n",
            "[392]\tvalid_0's auc: 0.84414\tvalid_0's binary_logloss: 0.133653\n",
            "[393]\tvalid_0's auc: 0.8441\tvalid_0's binary_logloss: 0.133658\n",
            "[394]\tvalid_0's auc: 0.844111\tvalid_0's binary_logloss: 0.133651\n",
            "[395]\tvalid_0's auc: 0.844153\tvalid_0's binary_logloss: 0.13364\n",
            "[396]\tvalid_0's auc: 0.844193\tvalid_0's binary_logloss: 0.133634\n",
            "[397]\tvalid_0's auc: 0.844186\tvalid_0's binary_logloss: 0.133628\n",
            "[398]\tvalid_0's auc: 0.844223\tvalid_0's binary_logloss: 0.133616\n",
            "[399]\tvalid_0's auc: 0.844206\tvalid_0's binary_logloss: 0.133619\n",
            "[400]\tvalid_0's auc: 0.844164\tvalid_0's binary_logloss: 0.133618\n",
            "[401]\tvalid_0's auc: 0.844165\tvalid_0's binary_logloss: 0.133619\n",
            "[402]\tvalid_0's auc: 0.844114\tvalid_0's binary_logloss: 0.133621\n",
            "[403]\tvalid_0's auc: 0.844111\tvalid_0's binary_logloss: 0.133617\n",
            "[404]\tvalid_0's auc: 0.844105\tvalid_0's binary_logloss: 0.133615\n",
            "[405]\tvalid_0's auc: 0.844138\tvalid_0's binary_logloss: 0.133607\n",
            "[406]\tvalid_0's auc: 0.844168\tvalid_0's binary_logloss: 0.133602\n",
            "[407]\tvalid_0's auc: 0.844194\tvalid_0's binary_logloss: 0.133594\n",
            "[408]\tvalid_0's auc: 0.844172\tvalid_0's binary_logloss: 0.133598\n",
            "[409]\tvalid_0's auc: 0.84415\tvalid_0's binary_logloss: 0.133596\n",
            "[410]\tvalid_0's auc: 0.844131\tvalid_0's binary_logloss: 0.133595\n",
            "[411]\tvalid_0's auc: 0.844183\tvalid_0's binary_logloss: 0.13359\n",
            "[412]\tvalid_0's auc: 0.844214\tvalid_0's binary_logloss: 0.133584\n",
            "[413]\tvalid_0's auc: 0.844167\tvalid_0's binary_logloss: 0.133584\n",
            "[414]\tvalid_0's auc: 0.844197\tvalid_0's binary_logloss: 0.133575\n",
            "[415]\tvalid_0's auc: 0.844251\tvalid_0's binary_logloss: 0.133569\n",
            "[416]\tvalid_0's auc: 0.844246\tvalid_0's binary_logloss: 0.133564\n",
            "[417]\tvalid_0's auc: 0.844276\tvalid_0's binary_logloss: 0.133559\n",
            "[418]\tvalid_0's auc: 0.844304\tvalid_0's binary_logloss: 0.133551\n",
            "[419]\tvalid_0's auc: 0.844261\tvalid_0's binary_logloss: 0.13355\n",
            "[420]\tvalid_0's auc: 0.844318\tvalid_0's binary_logloss: 0.133543\n",
            "[421]\tvalid_0's auc: 0.844295\tvalid_0's binary_logloss: 0.133544\n",
            "[422]\tvalid_0's auc: 0.844295\tvalid_0's binary_logloss: 0.133549\n",
            "[423]\tvalid_0's auc: 0.844328\tvalid_0's binary_logloss: 0.133541\n",
            "[424]\tvalid_0's auc: 0.844357\tvalid_0's binary_logloss: 0.133538\n",
            "[425]\tvalid_0's auc: 0.84437\tvalid_0's binary_logloss: 0.133532\n",
            "[426]\tvalid_0's auc: 0.84434\tvalid_0's binary_logloss: 0.133533\n",
            "[427]\tvalid_0's auc: 0.844358\tvalid_0's binary_logloss: 0.133525\n",
            "[428]\tvalid_0's auc: 0.844351\tvalid_0's binary_logloss: 0.133519\n",
            "[429]\tvalid_0's auc: 0.844334\tvalid_0's binary_logloss: 0.133516\n",
            "[430]\tvalid_0's auc: 0.844356\tvalid_0's binary_logloss: 0.133512\n",
            "[431]\tvalid_0's auc: 0.844395\tvalid_0's binary_logloss: 0.133509\n",
            "[432]\tvalid_0's auc: 0.844384\tvalid_0's binary_logloss: 0.133504\n",
            "[433]\tvalid_0's auc: 0.844353\tvalid_0's binary_logloss: 0.133506\n",
            "[434]\tvalid_0's auc: 0.844323\tvalid_0's binary_logloss: 0.133508\n",
            "[435]\tvalid_0's auc: 0.844332\tvalid_0's binary_logloss: 0.133509\n",
            "[436]\tvalid_0's auc: 0.844333\tvalid_0's binary_logloss: 0.133508\n",
            "[437]\tvalid_0's auc: 0.844314\tvalid_0's binary_logloss: 0.133506\n",
            "[438]\tvalid_0's auc: 0.844321\tvalid_0's binary_logloss: 0.133503\n",
            "[439]\tvalid_0's auc: 0.844364\tvalid_0's binary_logloss: 0.133496\n",
            "[440]\tvalid_0's auc: 0.844351\tvalid_0's binary_logloss: 0.133493\n",
            "[441]\tvalid_0's auc: 0.844411\tvalid_0's binary_logloss: 0.133485\n",
            "[442]\tvalid_0's auc: 0.844412\tvalid_0's binary_logloss: 0.133483\n",
            "[443]\tvalid_0's auc: 0.84442\tvalid_0's binary_logloss: 0.133477\n",
            "[444]\tvalid_0's auc: 0.844436\tvalid_0's binary_logloss: 0.133473\n",
            "[445]\tvalid_0's auc: 0.844393\tvalid_0's binary_logloss: 0.133474\n",
            "[446]\tvalid_0's auc: 0.844396\tvalid_0's binary_logloss: 0.133468\n",
            "[447]\tvalid_0's auc: 0.844477\tvalid_0's binary_logloss: 0.133459\n",
            "[448]\tvalid_0's auc: 0.844474\tvalid_0's binary_logloss: 0.133455\n",
            "[449]\tvalid_0's auc: 0.844517\tvalid_0's binary_logloss: 0.133449\n",
            "[450]\tvalid_0's auc: 0.84451\tvalid_0's binary_logloss: 0.133445\n",
            "[451]\tvalid_0's auc: 0.844502\tvalid_0's binary_logloss: 0.133439\n",
            "[452]\tvalid_0's auc: 0.844527\tvalid_0's binary_logloss: 0.133436\n",
            "[453]\tvalid_0's auc: 0.844488\tvalid_0's binary_logloss: 0.13344\n",
            "[454]\tvalid_0's auc: 0.84452\tvalid_0's binary_logloss: 0.133433\n",
            "[455]\tvalid_0's auc: 0.844515\tvalid_0's binary_logloss: 0.133435\n",
            "[456]\tvalid_0's auc: 0.844524\tvalid_0's binary_logloss: 0.133435\n",
            "[457]\tvalid_0's auc: 0.844583\tvalid_0's binary_logloss: 0.133426\n",
            "[458]\tvalid_0's auc: 0.844613\tvalid_0's binary_logloss: 0.133421\n",
            "[459]\tvalid_0's auc: 0.844656\tvalid_0's binary_logloss: 0.133408\n",
            "[460]\tvalid_0's auc: 0.84466\tvalid_0's binary_logloss: 0.133405\n",
            "[461]\tvalid_0's auc: 0.844641\tvalid_0's binary_logloss: 0.13341\n",
            "[462]\tvalid_0's auc: 0.844617\tvalid_0's binary_logloss: 0.133412\n",
            "[463]\tvalid_0's auc: 0.844613\tvalid_0's binary_logloss: 0.133411\n",
            "[464]\tvalid_0's auc: 0.844614\tvalid_0's binary_logloss: 0.133411\n",
            "[465]\tvalid_0's auc: 0.844602\tvalid_0's binary_logloss: 0.133413\n",
            "[466]\tvalid_0's auc: 0.844618\tvalid_0's binary_logloss: 0.13341\n",
            "[467]\tvalid_0's auc: 0.844642\tvalid_0's binary_logloss: 0.133404\n",
            "[468]\tvalid_0's auc: 0.844657\tvalid_0's binary_logloss: 0.133399\n",
            "[469]\tvalid_0's auc: 0.84463\tvalid_0's binary_logloss: 0.133395\n",
            "[470]\tvalid_0's auc: 0.844626\tvalid_0's binary_logloss: 0.133393\n",
            "[471]\tvalid_0's auc: 0.844623\tvalid_0's binary_logloss: 0.13339\n",
            "[472]\tvalid_0's auc: 0.844635\tvalid_0's binary_logloss: 0.133385\n",
            "[473]\tvalid_0's auc: 0.844676\tvalid_0's binary_logloss: 0.133379\n",
            "[474]\tvalid_0's auc: 0.844653\tvalid_0's binary_logloss: 0.133381\n",
            "[475]\tvalid_0's auc: 0.844679\tvalid_0's binary_logloss: 0.133381\n",
            "[476]\tvalid_0's auc: 0.844684\tvalid_0's binary_logloss: 0.133372\n",
            "[477]\tvalid_0's auc: 0.84468\tvalid_0's binary_logloss: 0.133371\n",
            "[478]\tvalid_0's auc: 0.844741\tvalid_0's binary_logloss: 0.133364\n",
            "[479]\tvalid_0's auc: 0.844662\tvalid_0's binary_logloss: 0.133369\n",
            "[480]\tvalid_0's auc: 0.844666\tvalid_0's binary_logloss: 0.133368\n",
            "[481]\tvalid_0's auc: 0.844653\tvalid_0's binary_logloss: 0.133367\n",
            "[482]\tvalid_0's auc: 0.844659\tvalid_0's binary_logloss: 0.13336\n",
            "[483]\tvalid_0's auc: 0.844742\tvalid_0's binary_logloss: 0.13335\n",
            "[484]\tvalid_0's auc: 0.84474\tvalid_0's binary_logloss: 0.13335\n",
            "[485]\tvalid_0's auc: 0.844714\tvalid_0's binary_logloss: 0.133351\n",
            "[486]\tvalid_0's auc: 0.844755\tvalid_0's binary_logloss: 0.133345\n",
            "[487]\tvalid_0's auc: 0.844744\tvalid_0's binary_logloss: 0.133345\n",
            "[488]\tvalid_0's auc: 0.844758\tvalid_0's binary_logloss: 0.133341\n",
            "[489]\tvalid_0's auc: 0.844777\tvalid_0's binary_logloss: 0.133338\n",
            "[490]\tvalid_0's auc: 0.844793\tvalid_0's binary_logloss: 0.133334\n",
            "[491]\tvalid_0's auc: 0.844776\tvalid_0's binary_logloss: 0.133336\n",
            "[492]\tvalid_0's auc: 0.844765\tvalid_0's binary_logloss: 0.133337\n",
            "[493]\tvalid_0's auc: 0.844754\tvalid_0's binary_logloss: 0.133336\n",
            "[494]\tvalid_0's auc: 0.844747\tvalid_0's binary_logloss: 0.133335\n",
            "[495]\tvalid_0's auc: 0.844778\tvalid_0's binary_logloss: 0.13333\n",
            "[496]\tvalid_0's auc: 0.844778\tvalid_0's binary_logloss: 0.133331\n",
            "[497]\tvalid_0's auc: 0.844773\tvalid_0's binary_logloss: 0.133329\n",
            "[498]\tvalid_0's auc: 0.844775\tvalid_0's binary_logloss: 0.133329\n",
            "[499]\tvalid_0's auc: 0.844783\tvalid_0's binary_logloss: 0.133325\n",
            "[500]\tvalid_0's auc: 0.844788\tvalid_0's binary_logloss: 0.133325\n",
            "[501]\tvalid_0's auc: 0.84475\tvalid_0's binary_logloss: 0.133323\n",
            "[502]\tvalid_0's auc: 0.844736\tvalid_0's binary_logloss: 0.133321\n",
            "[503]\tvalid_0's auc: 0.844755\tvalid_0's binary_logloss: 0.133315\n",
            "[504]\tvalid_0's auc: 0.844773\tvalid_0's binary_logloss: 0.133311\n",
            "[505]\tvalid_0's auc: 0.844772\tvalid_0's binary_logloss: 0.13331\n",
            "[506]\tvalid_0's auc: 0.844778\tvalid_0's binary_logloss: 0.133299\n",
            "[507]\tvalid_0's auc: 0.844751\tvalid_0's binary_logloss: 0.133296\n",
            "[508]\tvalid_0's auc: 0.844753\tvalid_0's binary_logloss: 0.133294\n",
            "[509]\tvalid_0's auc: 0.844804\tvalid_0's binary_logloss: 0.133289\n",
            "[510]\tvalid_0's auc: 0.844836\tvalid_0's binary_logloss: 0.133284\n",
            "[511]\tvalid_0's auc: 0.8448\tvalid_0's binary_logloss: 0.133286\n",
            "[512]\tvalid_0's auc: 0.844822\tvalid_0's binary_logloss: 0.133281\n",
            "[513]\tvalid_0's auc: 0.844816\tvalid_0's binary_logloss: 0.133283\n",
            "[514]\tvalid_0's auc: 0.844852\tvalid_0's binary_logloss: 0.13328\n",
            "[515]\tvalid_0's auc: 0.844837\tvalid_0's binary_logloss: 0.133281\n",
            "[516]\tvalid_0's auc: 0.844858\tvalid_0's binary_logloss: 0.133277\n",
            "[517]\tvalid_0's auc: 0.844847\tvalid_0's binary_logloss: 0.133275\n",
            "[518]\tvalid_0's auc: 0.844838\tvalid_0's binary_logloss: 0.133272\n",
            "[519]\tvalid_0's auc: 0.844817\tvalid_0's binary_logloss: 0.133269\n",
            "[520]\tvalid_0's auc: 0.844823\tvalid_0's binary_logloss: 0.133265\n",
            "[521]\tvalid_0's auc: 0.844803\tvalid_0's binary_logloss: 0.133269\n",
            "[522]\tvalid_0's auc: 0.844798\tvalid_0's binary_logloss: 0.133265\n",
            "[523]\tvalid_0's auc: 0.844783\tvalid_0's binary_logloss: 0.133262\n",
            "[524]\tvalid_0's auc: 0.844803\tvalid_0's binary_logloss: 0.133262\n",
            "[525]\tvalid_0's auc: 0.844826\tvalid_0's binary_logloss: 0.133254\n",
            "[526]\tvalid_0's auc: 0.844795\tvalid_0's binary_logloss: 0.133256\n",
            "[527]\tvalid_0's auc: 0.844812\tvalid_0's binary_logloss: 0.133251\n",
            "[528]\tvalid_0's auc: 0.844819\tvalid_0's binary_logloss: 0.13325\n",
            "[529]\tvalid_0's auc: 0.844785\tvalid_0's binary_logloss: 0.133252\n",
            "[530]\tvalid_0's auc: 0.844818\tvalid_0's binary_logloss: 0.133247\n",
            "[531]\tvalid_0's auc: 0.844798\tvalid_0's binary_logloss: 0.133248\n",
            "[532]\tvalid_0's auc: 0.844785\tvalid_0's binary_logloss: 0.133251\n",
            "[533]\tvalid_0's auc: 0.844811\tvalid_0's binary_logloss: 0.133246\n",
            "[534]\tvalid_0's auc: 0.844793\tvalid_0's binary_logloss: 0.133248\n",
            "[535]\tvalid_0's auc: 0.844802\tvalid_0's binary_logloss: 0.133245\n",
            "[536]\tvalid_0's auc: 0.844826\tvalid_0's binary_logloss: 0.13324\n",
            "[537]\tvalid_0's auc: 0.844789\tvalid_0's binary_logloss: 0.133237\n",
            "[538]\tvalid_0's auc: 0.844816\tvalid_0's binary_logloss: 0.133232\n",
            "[539]\tvalid_0's auc: 0.844809\tvalid_0's binary_logloss: 0.133232\n",
            "[540]\tvalid_0's auc: 0.844803\tvalid_0's binary_logloss: 0.133235\n",
            "[541]\tvalid_0's auc: 0.844832\tvalid_0's binary_logloss: 0.13323\n",
            "[542]\tvalid_0's auc: 0.844823\tvalid_0's binary_logloss: 0.133228\n",
            "[543]\tvalid_0's auc: 0.844803\tvalid_0's binary_logloss: 0.133228\n",
            "[544]\tvalid_0's auc: 0.844801\tvalid_0's binary_logloss: 0.133222\n",
            "[545]\tvalid_0's auc: 0.844814\tvalid_0's binary_logloss: 0.13322\n",
            "[546]\tvalid_0's auc: 0.844789\tvalid_0's binary_logloss: 0.133222\n",
            "[547]\tvalid_0's auc: 0.844813\tvalid_0's binary_logloss: 0.133226\n",
            "[548]\tvalid_0's auc: 0.844782\tvalid_0's binary_logloss: 0.133231\n",
            "[549]\tvalid_0's auc: 0.844795\tvalid_0's binary_logloss: 0.133227\n",
            "[550]\tvalid_0's auc: 0.844805\tvalid_0's binary_logloss: 0.133223\n",
            "[551]\tvalid_0's auc: 0.844802\tvalid_0's binary_logloss: 0.133223\n",
            "[552]\tvalid_0's auc: 0.844793\tvalid_0's binary_logloss: 0.133224\n",
            "[553]\tvalid_0's auc: 0.844798\tvalid_0's binary_logloss: 0.133224\n",
            "[554]\tvalid_0's auc: 0.84478\tvalid_0's binary_logloss: 0.133229\n",
            "[555]\tvalid_0's auc: 0.844788\tvalid_0's binary_logloss: 0.133226\n",
            "[556]\tvalid_0's auc: 0.84476\tvalid_0's binary_logloss: 0.133234\n",
            "[557]\tvalid_0's auc: 0.844771\tvalid_0's binary_logloss: 0.133227\n",
            "[558]\tvalid_0's auc: 0.844815\tvalid_0's binary_logloss: 0.13322\n",
            "[559]\tvalid_0's auc: 0.844842\tvalid_0's binary_logloss: 0.133212\n",
            "[560]\tvalid_0's auc: 0.844854\tvalid_0's binary_logloss: 0.133209\n",
            "[561]\tvalid_0's auc: 0.84483\tvalid_0's binary_logloss: 0.13321\n",
            "[562]\tvalid_0's auc: 0.844825\tvalid_0's binary_logloss: 0.133209\n",
            "[563]\tvalid_0's auc: 0.844852\tvalid_0's binary_logloss: 0.133201\n",
            "[564]\tvalid_0's auc: 0.844842\tvalid_0's binary_logloss: 0.133203\n",
            "[565]\tvalid_0's auc: 0.844814\tvalid_0's binary_logloss: 0.13321\n",
            "[566]\tvalid_0's auc: 0.844824\tvalid_0's binary_logloss: 0.133212\n",
            "[567]\tvalid_0's auc: 0.844821\tvalid_0's binary_logloss: 0.13321\n",
            "[568]\tvalid_0's auc: 0.844818\tvalid_0's binary_logloss: 0.133209\n",
            "[569]\tvalid_0's auc: 0.844793\tvalid_0's binary_logloss: 0.133213\n",
            "[570]\tvalid_0's auc: 0.844832\tvalid_0's binary_logloss: 0.133207\n",
            "[571]\tvalid_0's auc: 0.844845\tvalid_0's binary_logloss: 0.133205\n",
            "[572]\tvalid_0's auc: 0.844848\tvalid_0's binary_logloss: 0.1332\n",
            "[573]\tvalid_0's auc: 0.844839\tvalid_0's binary_logloss: 0.133199\n",
            "[574]\tvalid_0's auc: 0.844889\tvalid_0's binary_logloss: 0.133194\n",
            "[575]\tvalid_0's auc: 0.844845\tvalid_0's binary_logloss: 0.133198\n",
            "[576]\tvalid_0's auc: 0.844832\tvalid_0's binary_logloss: 0.1332\n",
            "[577]\tvalid_0's auc: 0.844838\tvalid_0's binary_logloss: 0.133195\n",
            "[578]\tvalid_0's auc: 0.844837\tvalid_0's binary_logloss: 0.133195\n",
            "[579]\tvalid_0's auc: 0.844829\tvalid_0's binary_logloss: 0.133196\n",
            "[580]\tvalid_0's auc: 0.844852\tvalid_0's binary_logloss: 0.133193\n",
            "[581]\tvalid_0's auc: 0.844825\tvalid_0's binary_logloss: 0.133195\n",
            "[582]\tvalid_0's auc: 0.844857\tvalid_0's binary_logloss: 0.133191\n",
            "[583]\tvalid_0's auc: 0.844874\tvalid_0's binary_logloss: 0.133187\n",
            "[584]\tvalid_0's auc: 0.844859\tvalid_0's binary_logloss: 0.133189\n",
            "[585]\tvalid_0's auc: 0.844814\tvalid_0's binary_logloss: 0.133193\n",
            "[586]\tvalid_0's auc: 0.844821\tvalid_0's binary_logloss: 0.133191\n",
            "[587]\tvalid_0's auc: 0.844814\tvalid_0's binary_logloss: 0.133189\n",
            "[588]\tvalid_0's auc: 0.844809\tvalid_0's binary_logloss: 0.133192\n",
            "[589]\tvalid_0's auc: 0.844815\tvalid_0's binary_logloss: 0.133194\n",
            "[590]\tvalid_0's auc: 0.844823\tvalid_0's binary_logloss: 0.133193\n",
            "[591]\tvalid_0's auc: 0.844817\tvalid_0's binary_logloss: 0.133191\n",
            "[592]\tvalid_0's auc: 0.84484\tvalid_0's binary_logloss: 0.133184\n",
            "[593]\tvalid_0's auc: 0.844902\tvalid_0's binary_logloss: 0.133179\n",
            "[594]\tvalid_0's auc: 0.84491\tvalid_0's binary_logloss: 0.133179\n",
            "[595]\tvalid_0's auc: 0.844882\tvalid_0's binary_logloss: 0.133175\n",
            "[596]\tvalid_0's auc: 0.844881\tvalid_0's binary_logloss: 0.133171\n",
            "[597]\tvalid_0's auc: 0.844872\tvalid_0's binary_logloss: 0.133172\n",
            "[598]\tvalid_0's auc: 0.844842\tvalid_0's binary_logloss: 0.133174\n",
            "[599]\tvalid_0's auc: 0.84485\tvalid_0's binary_logloss: 0.133172\n",
            "[600]\tvalid_0's auc: 0.844863\tvalid_0's binary_logloss: 0.133173\n",
            "[601]\tvalid_0's auc: 0.844845\tvalid_0's binary_logloss: 0.133174\n",
            "[602]\tvalid_0's auc: 0.84486\tvalid_0's binary_logloss: 0.13317\n",
            "[603]\tvalid_0's auc: 0.844844\tvalid_0's binary_logloss: 0.133173\n",
            "[604]\tvalid_0's auc: 0.84489\tvalid_0's binary_logloss: 0.133165\n",
            "[605]\tvalid_0's auc: 0.844926\tvalid_0's binary_logloss: 0.133166\n",
            "[606]\tvalid_0's auc: 0.84492\tvalid_0's binary_logloss: 0.133162\n",
            "[607]\tvalid_0's auc: 0.844947\tvalid_0's binary_logloss: 0.133163\n",
            "[608]\tvalid_0's auc: 0.844952\tvalid_0's binary_logloss: 0.133161\n",
            "[609]\tvalid_0's auc: 0.844908\tvalid_0's binary_logloss: 0.133164\n",
            "[610]\tvalid_0's auc: 0.844939\tvalid_0's binary_logloss: 0.13316\n",
            "[611]\tvalid_0's auc: 0.844921\tvalid_0's binary_logloss: 0.133159\n",
            "[612]\tvalid_0's auc: 0.844924\tvalid_0's binary_logloss: 0.133154\n",
            "[613]\tvalid_0's auc: 0.844928\tvalid_0's binary_logloss: 0.133151\n",
            "[614]\tvalid_0's auc: 0.844974\tvalid_0's binary_logloss: 0.133146\n",
            "[615]\tvalid_0's auc: 0.844932\tvalid_0's binary_logloss: 0.133154\n",
            "[616]\tvalid_0's auc: 0.844912\tvalid_0's binary_logloss: 0.133159\n",
            "[617]\tvalid_0's auc: 0.844932\tvalid_0's binary_logloss: 0.133156\n",
            "[618]\tvalid_0's auc: 0.844931\tvalid_0's binary_logloss: 0.133153\n",
            "[619]\tvalid_0's auc: 0.844907\tvalid_0's binary_logloss: 0.133156\n",
            "[620]\tvalid_0's auc: 0.844902\tvalid_0's binary_logloss: 0.133151\n",
            "[621]\tvalid_0's auc: 0.844912\tvalid_0's binary_logloss: 0.133152\n",
            "[622]\tvalid_0's auc: 0.844989\tvalid_0's binary_logloss: 0.133146\n",
            "[623]\tvalid_0's auc: 0.844955\tvalid_0's binary_logloss: 0.133148\n",
            "[624]\tvalid_0's auc: 0.844929\tvalid_0's binary_logloss: 0.13315\n",
            "[625]\tvalid_0's auc: 0.844925\tvalid_0's binary_logloss: 0.133148\n",
            "[626]\tvalid_0's auc: 0.84493\tvalid_0's binary_logloss: 0.13315\n",
            "[627]\tvalid_0's auc: 0.844892\tvalid_0's binary_logloss: 0.133153\n",
            "[628]\tvalid_0's auc: 0.844879\tvalid_0's binary_logloss: 0.133154\n",
            "[629]\tvalid_0's auc: 0.844872\tvalid_0's binary_logloss: 0.133154\n",
            "[630]\tvalid_0's auc: 0.844874\tvalid_0's binary_logloss: 0.133154\n",
            "[631]\tvalid_0's auc: 0.844831\tvalid_0's binary_logloss: 0.133156\n",
            "[632]\tvalid_0's auc: 0.844828\tvalid_0's binary_logloss: 0.133158\n",
            "[633]\tvalid_0's auc: 0.844837\tvalid_0's binary_logloss: 0.133157\n",
            "[634]\tvalid_0's auc: 0.844832\tvalid_0's binary_logloss: 0.133154\n",
            "[635]\tvalid_0's auc: 0.844846\tvalid_0's binary_logloss: 0.133154\n",
            "[636]\tvalid_0's auc: 0.844882\tvalid_0's binary_logloss: 0.133155\n",
            "[637]\tvalid_0's auc: 0.844884\tvalid_0's binary_logloss: 0.133155\n",
            "[638]\tvalid_0's auc: 0.844902\tvalid_0's binary_logloss: 0.133155\n",
            "[639]\tvalid_0's auc: 0.844927\tvalid_0's binary_logloss: 0.133149\n",
            "[640]\tvalid_0's auc: 0.84493\tvalid_0's binary_logloss: 0.133146\n",
            "[641]\tvalid_0's auc: 0.844926\tvalid_0's binary_logloss: 0.133144\n",
            "[642]\tvalid_0's auc: 0.844929\tvalid_0's binary_logloss: 0.133138\n",
            "[643]\tvalid_0's auc: 0.844905\tvalid_0's binary_logloss: 0.133142\n",
            "[644]\tvalid_0's auc: 0.844902\tvalid_0's binary_logloss: 0.133144\n",
            "[645]\tvalid_0's auc: 0.844882\tvalid_0's binary_logloss: 0.133145\n",
            "[646]\tvalid_0's auc: 0.844897\tvalid_0's binary_logloss: 0.133144\n",
            "[647]\tvalid_0's auc: 0.844957\tvalid_0's binary_logloss: 0.133133\n",
            "[648]\tvalid_0's auc: 0.844973\tvalid_0's binary_logloss: 0.133131\n",
            "[649]\tvalid_0's auc: 0.844934\tvalid_0's binary_logloss: 0.133133\n",
            "[650]\tvalid_0's auc: 0.844935\tvalid_0's binary_logloss: 0.133127\n",
            "[651]\tvalid_0's auc: 0.844926\tvalid_0's binary_logloss: 0.133128\n",
            "[652]\tvalid_0's auc: 0.844905\tvalid_0's binary_logloss: 0.133128\n",
            "[653]\tvalid_0's auc: 0.84493\tvalid_0's binary_logloss: 0.133127\n",
            "[654]\tvalid_0's auc: 0.844919\tvalid_0's binary_logloss: 0.133125\n",
            "[655]\tvalid_0's auc: 0.844928\tvalid_0's binary_logloss: 0.133122\n",
            "[656]\tvalid_0's auc: 0.844914\tvalid_0's binary_logloss: 0.133122\n",
            "[657]\tvalid_0's auc: 0.844924\tvalid_0's binary_logloss: 0.133116\n",
            "[658]\tvalid_0's auc: 0.844933\tvalid_0's binary_logloss: 0.133116\n",
            "[659]\tvalid_0's auc: 0.844943\tvalid_0's binary_logloss: 0.133114\n",
            "[660]\tvalid_0's auc: 0.844963\tvalid_0's binary_logloss: 0.133109\n",
            "[661]\tvalid_0's auc: 0.844976\tvalid_0's binary_logloss: 0.133109\n",
            "[662]\tvalid_0's auc: 0.844996\tvalid_0's binary_logloss: 0.133107\n",
            "[663]\tvalid_0's auc: 0.844984\tvalid_0's binary_logloss: 0.133109\n",
            "[664]\tvalid_0's auc: 0.844991\tvalid_0's binary_logloss: 0.133111\n",
            "[665]\tvalid_0's auc: 0.844973\tvalid_0's binary_logloss: 0.133108\n",
            "[666]\tvalid_0's auc: 0.844937\tvalid_0's binary_logloss: 0.133108\n",
            "[667]\tvalid_0's auc: 0.844937\tvalid_0's binary_logloss: 0.133106\n",
            "[668]\tvalid_0's auc: 0.844941\tvalid_0's binary_logloss: 0.1331\n",
            "[669]\tvalid_0's auc: 0.844945\tvalid_0's binary_logloss: 0.133096\n",
            "[670]\tvalid_0's auc: 0.844947\tvalid_0's binary_logloss: 0.133094\n",
            "[671]\tvalid_0's auc: 0.844943\tvalid_0's binary_logloss: 0.133091\n",
            "[672]\tvalid_0's auc: 0.844939\tvalid_0's binary_logloss: 0.133093\n",
            "[673]\tvalid_0's auc: 0.84496\tvalid_0's binary_logloss: 0.133089\n",
            "[674]\tvalid_0's auc: 0.844948\tvalid_0's binary_logloss: 0.13309\n",
            "[675]\tvalid_0's auc: 0.84492\tvalid_0's binary_logloss: 0.133093\n",
            "[676]\tvalid_0's auc: 0.844933\tvalid_0's binary_logloss: 0.133094\n",
            "[677]\tvalid_0's auc: 0.844906\tvalid_0's binary_logloss: 0.133098\n",
            "[678]\tvalid_0's auc: 0.84489\tvalid_0's binary_logloss: 0.133094\n",
            "[679]\tvalid_0's auc: 0.844854\tvalid_0's binary_logloss: 0.133096\n",
            "[680]\tvalid_0's auc: 0.84486\tvalid_0's binary_logloss: 0.133096\n",
            "[681]\tvalid_0's auc: 0.844849\tvalid_0's binary_logloss: 0.133097\n",
            "CPU times: user 1min 3s, sys: 363 ms, total: 1min 3s\n",
            "Wall time: 33.3 s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LGBMClassifier(boosting_type='gbdt', class_weight=None, colsample_bytree=0.5,\n",
              "               importance_type='split', learning_rate=0.01, max_depth=9,\n",
              "               min_child_samples=20, min_child_weight=0.001, min_split_gain=0.0,\n",
              "               n_estimators=681, n_jobs=-1, nthread=-1, num_leaves=31,\n",
              "               objective='binary', random_state=42, reg_alpha=1, reg_lambda=0.0,\n",
              "               scale_pos_weight=1, silent=True, subsample=0.8,\n",
              "               subsample_for_bin=200000, subsample_freq=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 100
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "OFf2WKQC-V3R",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "430c8284-548d-43d5-e15d-45a09ef96d7e"
      },
      "source": [
        "y_pred = model.predict_proba(X_val_200)[:,1]\n",
        "label=\"LightGBM Classifer(top 250)\"\n",
        "labels.append(label)\n",
        "y_preds.append(y_pred)\n",
        "i+=1\n",
        "result = plot_auc(y_val,y_pred,label)\n",
        "result_df = result_df.append(result)\n",
        "del result"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\t\t LightGBM Classifer(top 250) on Normal dataset \t\t \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp8AAAEcCAYAAACBJN/jAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd1gUV9vH8e/uwtIVUESwxBZRY8OGvSGKFWusiYmJJsZoiokae01imtFoLNFobDGPUTQUezeJvUcxsaMiIEVhl7Jl3j943Sc8oq4FFuH+XJfXBTNnZn6zi3DvmTlnVIqiKAghhBBCCJEH1LYOIIQQQgghCg8pPoUQQgghRJ6R4lMIIYQQQuQZKT6FEEIIIUSekeJTCCGEEELkGSk+hRBCCCFEnpHiU4jnzPXr1/Hz88NoNFrVfsyYMcyaNSuXUwkhhBDWkeJTPHOtW7emZs2a+Pv706RJE8aMGYNOp8vW5tixY7z66qv4+/tTt25d3n77bS5cuJCtTWpqKjNmzKBly5b4+/vTpk0bZsyYQWJiYq7mP3jwIH5+fkyePDnb8r59+7J+/fpcPbatvfLKK6xdu7bAHEcIIUT+I8WnyBULFizg+PHjbNiwgbNnz7Jo0SLLuuPHj/PGG28QGBjIvn372LFjB35+fvTt25fo6GgAMjMzGThwIBcuXGDx4sUcPXqUX375BXd3d06fPp3r+Z2dndm4cSPXr19/6n1Z20MphBBCFAZSfIpc5eXlRdOmTTl37pxl2ZdffklISAgDBw7E1dUVd3d3PvjgA2rVqsV3330HwMaNG4mJiWHu3LlUqlQJtVpNsWLFGDZsGC1atMjxWMeOHaNHjx7UrVuXHj16cOzYMcu6V155hW+//ZY+ffrg7+/PoEGDHtqD6ubmRvfu3Zk3b16O681mM99//z2tWrWiUaNGjBo1ipSUFOC/l8XXrl1Ly5YtGThwIOvXr6dPnz58+umn1KtXj8DAQI4dO8b69etp0aIFjRo1IjQ01LL/3bt307VrV+rUqUOLFi0sr4s1zp49S7du3fD39+f9998nIyPDsu7OnTu89dZbNGzYkPr16/PWW29x69YtAGbNmsWRI0eYOnUq/v7+TJ06FYDp06fTokUL6tSpQ/fu3Tly5Ihlf6dOnaJ79+7UqVOHxo0b89lnn1nWnThxgj59+lCvXj26dOnCwYMHH3ocIYQQhYQixDPWqlUr5ffff1cURVFiYmKUTp06KdOmTVMURVH0er1SpUoV5c8//7xvu19//VVp0qSJoiiK8v777yujRo2y+phJSUlKvXr1lNDQUMVgMChhYWFKvXr1lMTEREVRFGXAgAFKYGCgcunSJSUtLU0ZMGCA8uWXX+a4rwMHDijNmjVT4uLiFH9/f+XixYuKoihKnz59lHXr1imKoihr165V2rRpo1y7dk1JTU1Vhg0bpnz00UeKoihKdHS0UrlyZeXjjz9WdDqdkpaWpqxbt06pWrWq8uuvvypGo1H55ptvlBYtWiiTJ09WMjIylH379im1a9dWUlNTLRmioqIUk8mknDt3TmnUqJGybdu2bPs3GAz3Zc/IyFBatmypLF26VMnMzFQ2bdqkVKtWTfnmm28URVGUxMREZfPmzYper1dSUlKU4cOHK0OHDrVsP2DAAOU///lPtn1u2LBBSUxMVAwGg7JkyRKlcePGSnp6uqIoivLyyy8roaGhiqIoSmpqqnL8+HFFURTl1q1bSoMGDZTdu3crJpNJ2b9/v9KgQQMlISHhgccRQghROEjPp8gVw4YNw9/fnxYtWuDp6cmIESOArJ43s9mMl5fXfdt4eXmRlJQEQHJyco5tHmT37t288MILdO3aFTs7Ozp16kSFChXYtWuXpU337t0pX748jo6OBAcHZ+uNzYmXlxd9+vRhzpw5960LCwvjtddeo0yZMri4uPDhhx8SGRmZ7RL78OHDcXZ2xtHREYDSpUvTo0cPNBoNHTp0ICYmhmHDhqHVamnatClarZZr164BEBAQgJ+fH2q1mipVqtCxY0cOHTr0yNfh5MmTGAwGBg4ciL29PcHBwdSoUcOy3sPDg3bt2uHk5ISrqytDhw7l8OHDD91nSEgIHh4e2NnZMWjQIDIzM7l8+TIAdnZ2XLt2jcTERFxcXKhduzaQ1XPdvHlzWrRogVqtpkmTJlSvXp09e/Y88hyEEEIUbHa2DiAKpnnz5tG4cWMOHTrEyJEjSUpKokiRIhQpUgS1Wk18fDwVK1bMtk18fDweHh4AuLu7Ex8fb/Xx4uLi8PX1zbbM19eX2NhYy/f/LmadnJzQ6/WP3O/gwYMJCgoiKirqvuOVKlXK8n2pUqUwGo0kJCRYlpUsWTLbNsWKFbN8fa8gLV68uGWZg4ODZWDWyZMn+eqrr/jnn38wGAxkZmYSHBz8yLxxcXF4e3ujUqksy/79uqSlpfHZZ5+xb98+7ty5A4BOp8NkMqHRaHLc55IlS/j111+Ji4tDpVKRmppq+ZAwY8YM5syZQ/v27SldujTvvvsurVq14ubNm2zevDlb8W80GgkICHjkOQghhCjYpOdT5KoGDRrQvXt3Zs6cCWQN5KlduzabN2++r+2mTZto2LAhAI0bN2b//v1WFYgAJUqU4ObNm9mWxcTE4O3t/VT5PTw8GDhwIN9+++19x7tx44bl+5s3b2JnZ5etwPx3Afi4Ro4cSWBgIHv27OHo0aP06dMHRVEeuZ2XlxexsbHZ2v77dfnxxx+5fPky//nPfzh27BirVq0CeOC+jxw5wuLFi/n22285fPgwR44cwc3NzdK+XLlyfPPNN/z5558MHjyYESNGoNfr8fHxISQkhCNHjlj+nThxgiFDhjzxayKEEKJgkOJT5LqBAwfyxx9/WHoPR44cyYYNG1i+fDmpqancuXOHWbNmceLECd59910g61JvyZIlGT58OBcvXsRsNpOUlMSCBQtyvHTbokULrly5QlhYGEajkcjISC5cuEDLli2fOv/rr7/O8ePHuXTpkmVZp06d+Omnn4iOjkan0zFr1izat2+Pnd2zuZig0+koWrQoDg4OnDp1ivDwcKu2q127NnZ2dixfvhyDwcDWrVuzzQ6g0+lwcHCgSJEiJCcnM3fu3GzbFy9e3DLjwL32Go0GT09PjEYjc+fOJTU11bJ+48aNJCYmolarKVKkCABqtZouXbqwa9cu9u3bh8lkIiMjg4MHD1oGN/3vcYQQQhQeUnyKXOfp6UlISIhl5Hi9evVYvHgx27Zto1mzZrRq1Ypz586xevVqypUrB4BWq2XZsmVUqFCBQYMGUbduXXr16kVSUhI1a9a87xgeHh4sWLCApUuXEhAQwOLFi1mwYAGenp5Pnd/V1ZU333yT5ORky7IePXrQpUsXBgwYQGBgIFqtlgkTJjz1se6ZNGkSc+bMwd/fn3nz5tG+fXurttNqtXz33XeEhobSoEEDIiMjCQoKsqwfOHAgGRkZNGzYkN69e9OsWbNs27/66qts2bKF+vXrM336dJo2bUqzZs1o164drVu3xsHBAR8fH0v7ffv20bFjR/z9/ZkxYwazZs3C0dERHx8fvv/+exYuXEijRo1o0aIFS5YswWw253gcIYQQhYdKseZanhBCCCGEEM+A9HwKIYQQQog8I8WnEEIIIYTIM1J8CiGEEEKIPCPFpxBCCCGEyDMFZpJ5s9mMTqfD3t7+qeZXFEKIwkRRFAwGAy4uLqjV0h8hhMh9Bab41Ol0/P3337aOIYQQz6XKlSvj5ub2xNsbDAauX79Oenr6M0wlhHgeaTQa3N3dKV68eI4fagtM8Wlvbw9k/QLVarWPvf2ZM2eoXr36s46Vr8k5Fw5yzoXDk55zZmYmf//9t+V36JO6fv06bm5ulCtXTq4+CVGI3buaEhsby/Xr1ylbtux9bQpM8Xnvl51Wq8XBweGJ9vGk2z3P5JwLBznnwuFpzvlpC8b09HQpPIUQqFQqtFotpUqV4vz58zm2kRt8hBBCPBNSeAoh7nnYPeRSfAohhBBCiDwjxacQQogCp3Xr1jkOQh08eDDXrl175PavvPIKu3btynHd+vXruXz5crZlV65cYcSIEbRu3ZquXbvSpUsXvv76a0wmk2V/gYGBhISE0KFDB4YPH05KSgoABw8exM/Pj5kzZ96Xwc/PD51Ol2OOy5cvM2zYMAIDA+nevTt9+vRh+/btj8z/pHbs2JEt46xZswgODqZfv35PtL/ly5ezZMkSAM6dO0dkZOQzyflvU6ZMITg4mC5dutCnTx9Onz5tWffv9yQkJIR169ZZ1l2+fJnevXvTrl07evfuzZUrV4Cse6S7d+9uee8eZM+ePfj5+bFt27Zsy3N6X0aMGMH69est3+/bt48+ffrQtm1bunfvzltvvfXAy9c5SUtL4/333ycoKIjg4OAH/hyYzWamT59Ohw4d6Ny5M2+88QaxsbHZ2mRkZNCxY0e6d+9uWbZ+/Xrq1atned2GDRtmdbZ78qT4nDlzJq1bt8bPz++BI9JNJhNTpkyhTZs2BAUFsXbt2ryIJoQQohD54YcfchwA8ThCQ0MtxQhAXFwc/fv3p2XLluzcuZMNGzawevVq1Go1mZmZlnbjx49n48aNREREoNFo+Pnnny3rypcvz44dOyzFanR0NHq9/oEZ4uLiGDBgAEFBQezYsYP169fz3XffkZqa+lTn9jCBgYGMHj3a8v3SpUtZvXo1q1evfqz9mEwm0tLSWL58Of379weyis/Nmzc/07wAzZs3JywsjN9++4233nqLDz74INv6e+/Jxo0b6dGjh2X5pEmT6NevH1u2bKFfv35MnDgRyBpX0qVLF5YuXfrQ465bt46GDRtmK2itsX//fsaNG8e4cePYunUr69ev5/333yc+Pt7qfSxZsgRXV1e2bdvGggULGD9+fI4fYHbu3MmpU6f47bffCAsLo1KlSsyfPz9bm1mzZlGrVq37tm3cuLHldZs3b95jnSPk0YCjwMBAXn31VcsPWU7CwsK4du0aW7duJTk5ma5du9KoUSNKly6dFxGFEEI8I9uPJbL1SGKu7LttPU/a1PF84u1bt27NggULqFy5MhcuXOCTTz4hLS2NKlWqcO3aNYYOHUqrVq0AOHToEIsWLSIuLo727dvz0UcfsW7dOs6cOcP06dP59ttvGT16NAcPHiQgICBb75Crq+t9hc49RqOR9PR0ihYtalnm7OzMiy++yP79+2nRogWhoaF07dqVM2fO5LiPVatWERAQQNeuXS3LvLy8sn1/T1hYGMuXL8dgMAAwevRoGjVqhNlsZurUqRw4cACtVouzszNr1qwhISGBkSNHkpCQAECjRo0YO3Ys69evZ/fu3cyZM4d+/fqRkZHBwIEDadq0KaNHjyY0NJTVq1djMplwdXVl8uTJVKhQgfXr1/Pbb7/h4uLC1atX+fLLLzl//jz169fH0dGRpKQk5syZQ2pqKiEhIdSvX5/x48ezd+9evvnmG0wmE56enkydOpUXXniBgwcPMmPGDKpUqcJff/2Fk5MTn3/+OZUqVbrv3O+9lwC1a9fm1q1bmM3mh96PmJCQwNmzZy0FZqdOnZg2bRqJiYl4enrSsWNHevTowYgRI3LcPikpiQMHDhAZGUnHjh2Jj4/Hy8vrgcf7t3nz5vHOO+9Qo0YNy7KqVatate09mzZt4vPPPwegXLlyVK9enb1799K+ffv72mZmZpKRkYFarUan02WruY4cOcKVK1d4/fXXiYqKeqwMj5InxWe9evUe2SYyMpJevXqhVqvx9PSkTZs2bN68mTfffDMPEgohxNNRFAVFsXUK8ThGjRrFwIEDCQkJ4fTp07z88svZ1sfExLBq1Sp0Oh1t2rShZ8+e9OjRgw0bNjBo0CBLYbN06VKaNGnyyOPdK1hjYmIoX7483bp1y7a+W7durFmzhubNmxMREcGaNWuYPn16jvs6e/asVccEaNq0KZ06dUKlUnHp0iVee+019u7dS1RUFAcPHiQyMhK1Ws2dO3eArGK1bNmyLFu2DMCy/N9Wr16Nn58fa9aswcXFhSNHjrBp0yZWrVqFVqtlz549jB07ljVr1gBw8uRJNm7caOl1XrFiBTVr1gTAw8ODESNGWApbyCoAR40axcqVK6lUqRJr167lo48+slwVPX/+POPHj+eLL74gNDSUUaNGZbt0nZNVq1bRsmXLbIXnF198wTfffIOfnx8ff/wx3t7exMTE4O3tjUajAbLmrCxRogQxMTF4enri5eWFvb09Fy9epGLFivcd57fffqNly5YUL16coKAgQkNDGTJkiFXv1dmzZy29rDkZMWIEV69ezXHdL7/8gqOjIzdv3qRUqVKW5T4+Pty6deu+9q1bt+bQoUM0bdoUR0dHKlSoYDm2Xq/n008/Zf78+dl6+e85dOgQISEhuLq6MnjwYFq2bGnV+d2Tb6ZaiomJwdfX1/L9g14sIYR4WoqiEH/HgNn8ZNViwl0jB87dwWDM2l6fYWLvqTtkGMzPMuYjqTDTyv4w3bS7Sa7WC+rWzdPjP0ibOk/XO5kXUlNT+fvvv+ncuTMANWrUwM/PL1ub4OBg1Go1bm5uVKxYkWvXrlGuXLlH7nvRokVERESQnJzMrFmzqFOnDpB1ibdVq1aYTCYmTZrEl19+ybhx4yzbBQQEMGXKFLZv307lypXx8PB4JucaHR3NyJEjiY2Nxc7Ojtu3bxMfH0+ZMmUwGo2MGzeOgIAASzFdq1Ytli1bxsyZM2nQoAFNmzZ95DF27txJVFQUvXr1ArL+j929e9eyvk6dOtlud4iNjc3WK/m/Tp48SZUqVSy9mT169GDKlCmW2wpeeOEFGjRoAEBISAgTJkwgNTUVV1fXHPcXERFBWFgYq1atsiz74osv8PHxwWQysXDhQt5///1st0I8jJeXF7GxsTkWn+vXr2fMmDFA1geKcePGPbL4tHamiHvF+bPw119/cfHiRfbu3YuLiwszZszg888/Z+LEiXzxxRf069cPb2/v+4rPli1b0qFDBxwdHTl79iyDBw9m+fLlOb4WD5Jvis9n5UGXKKxx9OjRZ5jk+SDnXDgUhnNOSIU7+nu/wFVc2nQsx3axd+HwJTVxd59+WiBH+6ziU6UCv5IKxd3yruuziCGBZkkb8Mm4QoxjBZyKexWK9/lZe9gf/X/Pm6rRaCz3Y/6vatWqZRvIMmTIEIYMGUL37t0tl7r/TaPREBQUxBdffHFflvbt2zN+/Hg+++yzh+b+32M+zIcffsiYMWNo06YNZrOZWrVqkZGRgZeXFxERERw8eJA//viDr776itDQUPz9/QkNDeWPP/5g48aNLFq06JFFmaIo9OjRg/feey/H9S4uLtm+d3BwICMjw6r8T2vbtm3MmjWLZcuWUbx4cctyHx8fIOv9ePXVV5k7dy5msxkfHx9iY2MxmUyW9z0uLs7SHrIG4uQ0r+6ZM2e4cOFCtg8VcXFxHD16lLp16+Lp6UlycnK2bZKSkvD0zPqwVq1aNU6dOvXAS+3W9Hz6+vpy48YNyz5jYmIICAi4r31oaCgNGza0PNmsS5cujB07Fsj6m7F3716+//57MjIyuHPnDp07dyYsLMyy33t569Spw6lTp57P4tPHx4ebN29auuH/tyfUWtWrV3+iiZbv/WAUJnLOhUNBOGdduokfN8Vw7lrOo34VBa7EWv9YRzcnDT2be1K2hOMTZ/ItpuWlcjn3suQmxWwi49AG0vauRKWxx6nDCKrWasuxY8ee6H3OyMh4qg/tzytXV1defPFFwsPD6dy5M3/99ZfVj2h2cXHJNtq5X79+dOvWjQ0bNljuuTSZTDkWnvccPHgwx17U3r174+TkRPPmzR+a4d4xw8LCLL23CQkJ7Nu37777PlNSUiz38q1bt84yCCoxMRGNRkOzZs1o3Lgxu3fvJjo6mvT0dEqWLEnHjh2pV68eQUFBmM0P79Vv3bo1o0ePpnfv3pQsWRKTycS5c+ce+NStypUrZ5sxwNXVNdtrWrt2bcaOHWu5tB0aGkq1atUsPZvXrl3jyJEj1KtXj7CwMCpXrpxjr+euXbv47LPPWLp0abb7GY1GI8nJyZZiNCIigsqVK6NWqylWrBhVq1YlPDyckJAQwsPDqVq1qqXoMplMREdHU7ly5fuOt27dOt58881s9/suXLiQdevWUbduXZo0acKGDRvo0KEDDg4OREVFcfHiRUvtM3ToUCZMmED16tV56aWXAIiKiuL27ds0bdrUqp7P4OBgfvnlF2rUqMGVK1c4ffo0X3/99X3tSpcuze+//85rr72Gvb09e/bs4cUXXwSybr245+DBg8ycOdNyW0NsbCze3t4A3LhxgxMnTjB06NBH5vq3fFN8BgcHs3btWtq2bUtycjLbt2/P1j0uhCh4zGaFTYcTSEoxPrBN/B2DZfBKfT837DQ591SV8nKgVgVXXvB25O+/z1O5sl+O7dRq8CvtjL3d8zfTnDHuMvqI2Zhi/sG+ckOc272D2q2YrWPlW6+//rrlvj3I/gcVsmZiGTt2LIsWLaJy5cpWP9++d+/efP755yxZsoTRo0fTuHFjVq5cyTfffMOcOXNwd3dHq9XSpk0bSwEB/73n02g04uPjw5QpU+7bt7e3N4MHD35kBm9vb1asWMFXX33Ft99+i7OzM87Ozjlu+8knn/DOO+9QtGhRmjVrhru7O5DVyTNhwgSMRiMmk4nmzZtTu3ZtQkNDWbZsGWq1GrPZzJQpUx46QAegfv36vP/++wwdOtRSeAcHBz+w+Gzbti1Tpkxh+PDhQNagph9//JEuXbrQoEEDy/2cH330EUajEU9PT7788kvL9pUrV2bt2rVMnjwZR0fH+3qR/33u9vb22QYHLVu2DAcHB4YMGWL5gFCiRAm++eYbS5vJkyczZswYvv/+e4oUKZJtiqljx45Rq1at+35WMjIyiIiIuK+XuFOnTnTp0oXx48fTs2dPbt68Sc+ePVGr1Tg6OjJr1izLLRbNmzdn6tSpTJ06leTkZOzs7ChdujQjR4586Ov/b2+88QZjxowhKCgItVrN1KlTLYX57NmzKVGiBH379qV///78888/dOnSBTs7O3x8fJg2bdoj979q1Sp27Nhh+b/14YcfUq1aNavzAagUJfdvkZ8+fTpbt27l9u3beHh44O7uTkREBIMHD2bEiBHUqFEDk8nE1KlT+f3334Gsudh69+5t9THufXqXnk/ryTkXDrl5zrp0E3HJmdzRGTl5MZXbdx7c05OTfaetv09yQBtv+geWtKptQXufFaOB9N/XkP7nWlSObji3exv7Kk2zXTJ+0nN+2t+d95w7d+6xR+Xamk6nw9nZGZVKxYULF3jllVfYvHlztlHoIve88cYbjBw58rELl//tictrI0eOpGfPnjRq1Mgmx3+ePOj3Qp70fI4fP57x48fft/yHH36wfK3RaHL8FCiEyFvnruo4fUVH4l0DyalG7uiMmHP4jHpXb+LKrf9e6lapwKuo/WMdq6iLhkbVPHmrk688mvEBjDei0EXMxnz7GtrqrXFqMxi1cxFbxyoQjh8/zhdffMG9Pphp06ZJ4ZmHJkyY8MD7F/OrzMxM6tWrJ4XnU8o3l92FELkrNc3ErhNJmHIY4W0yK/xzPY1Tl1JJSv3vJXDfYlqKutihyeFSd1EXOwa08aaMlyMajYoS7va8WMo5V8+hMFEy00nbs5yMw7+hKlIc195TsK/46GnrhPWaNm1q1UhukTvKlStn1ewB/ysgIMBmvZ5arZa+ffva5NgFiRSfQhQAJpNCSpqJTKMZXZqJHceTLNMIxcapOXj9OicupRId9+DRpR5udtSqmHXPZBkvB2pWcMXNWX5F2ILh8gn0m+ZgTo7FoU5HnFq9hspBCnshRMEgf1mEyOeMJoW/r+stPZbH/knhwo007uqN3NWbSNWbSE2/fwoYB3s1ajWYTSrso5Px9tTyVidfAuvkPHegq6NGLn3bmDk9lbQdS8g8uRW1py+uA2ZiXzbnARtCCPG8kuJTiDx0V2fk6D8pHP07hatWTg2UlGog4e79o8HrvuiGbzEH3Jw1FHG2w81Zg6O9GlTg7KChWY2sEa1ZA1FqP9PzEM9e5t9/ot/8PYouGYdGPXFq2g+V/ZMPABJCiPxKik8hcoHZrJCWaea70Oscu5AC/3+bZUraf3soX/B2xNtD+8h9FS9qT7/WRfAt/t+2rk4aKvnKZdiCwJyahH7bQgzn9qEpUR7nXhOx83nR1rGEECLXSPEpxDNiMins/yuZQ+dSOH4xxTJ3ZY3yLpQr+d/JzMuXdKKenxteRR9deIqCS1EUMs/sIm37IpTMNBxbvIpjwx6oNPJr+Vm5c+cOzZo14+WXX7bMuPLdd9+h1+sZPXq0pd3KlSs5c+YMn3/+OQCXL1/mq6++IioqiqJFi6LVannzzTdp06aN1ceeN28eoaGhQNYjFocNG5ZjOz8/P8vk5pD1yMf/fcynEAWN/JYT4gmZzQqrd8by1xUdigLR8ekk/n/B2axGUSqXdqakh5Ym1YvKvZQiG/OdOHSb52G8eARNqSq4dHwPTfGyj95QPJbw8HBq1apFREQEo0aNQqt99Ae+uLg4BgwYwMcff8y8efMAiI+Pt8xBbY3Dhw+zefNmwsPDAejVqxcNGjSgfv36ObZfs2bNfY+fFKIgk+JTiCdwPlpP6P549pxKxk6jonJpJ6qXd6VWBVdqVXSlVHG5V0/cT1HMZByLJG3XMlAUnILewqFuR1RqzSO3FY9v3bp1fPzxxyxcuJAdO3bQvn37R26zatUqAgICsj2i0svL675HVj5MZGQkXbt2xdEx64pH165diYyMfGDxKURhI8WnEI/p5u0MPlp4AaNJoe6Lbowf8AKOWikexMOZEm6gj5yNMfov7Mr749x+OBp3b1vHyhUZp3eQeXJbruxbWysIhxqBj2wXFRVFcnIyDRs2JD4+nnXr1llVfJ49e5YmTZo8cP2iRYuIiIjIcd2ECROoV68eMTExNGjQwLLcx8eHw4cPP3Cfr7zyiuXxlsOHD7eqh1aI55kUn0I8hMmksDDiBsn/mnj90s007DQq5r9XmVLFHeSSungoxWwi4+B60vauQmWvxbnT+2hrtJGfm5Yg3mIAACAASURBVFz266+/EhISgkqlom3btkyfPp3Y2NgHvu7Wvh9DhgxhyJAhzyzn7t278fHxITU11XKp/4MPPnhm+xciP5LiU4gH0KcbmflLNIei7uLsoKbY/z860kGrZnAHX0p7OT5iD6KwM8ZeQh/xLaZbF7H3a4xzu6GoXT1tHSvXOdQItKp3MrdkZmYSHh6OVqtl48aNABgMBtavX4+Hhwc3btzI1j4pKQlPz6z3pVq1apw+ffqB+7am59PHx4ebN29alsfExODj45PjNveWu7q60qtXL5YuXWr9iQrxnJLiU4gcHIq6y5f/uYY+3cTAtiXp3bKE9FQJqynGTNL3ryH9wK+onNxw6T4WbZUHX8oVz9aOHTsoX748P//8s2XZ8ePHGT16NPPnz2f+/PncunWLkiVLkpycTGRkJBMnTgSgX79+dOvWjbCwMDp37gxAQkIC+/bto2vXrlb1fAYHBzN9+nT69+8PwIYNG5gwYcJ97e7cuYODgwOOjo4YjUa2bNlC1apVn9XLIES+JcWnEP/jWlw6k366TFEXDeMHlKNRtaK2jiSeI8brZ9FFzMaccB1tjUCc2gxG7eRm61iFyrp16yyF4z3+/v6YzWYSEhIYO3Ys77zzDiaTCUVRGDBgAI0aNQLA29ubFStW8NVXX/Htt9/i7OyMs7MzgwcPtvr4AQEBtG3blo4dOwJZA47u3QO6Y8cOdu7cyYwZM7h06RITJ05EpVJhNBrx9/fnvffee0avghD5lxSfQvy/Db/Hc/j8XY79kwrAZ29UpLyPk41TieeFkplG2u7lZBwJQ12kOK59pmJfoa6tYxVKixcvznH59u3bLV936NDhgdtXrFiR+fPnP1WG4cOHM3z48PuWBwYGEhiYdUuCv78/YWFhT3UcIZ5HUnyKQutaXDo/bYkhOj4DswI3bmdQ2suBKmWcqVHBRQpPYTXDpWPoN32H+U48DnU74tRyICoHeQKVEELkRIpPUegoisLuk8l8t+E6ZrNC3cpFsNOoqO/nxsC2JWXaJGE1c1oKaTsWk3lqO2rP0ri9MhO7Mi/ZOpYQQuRrUnyKQuXabVgx/wLno/VU9HXiw55lqCA9nOIJZEb9jn7LfBT9HRwbv4xj076o7GR+RiGEeBQpPkWhoCgKs36NZtsxO4oVyWRkrzK0ru2BWi0j2MXjMacmot+yAMP539F4V8S5zxTsvCvaOla+oCiKzAohhADAbDY/cJ0Un6JAUhQFo0n5/6/h09VXOBiVQoMKZj4ZWEUurYvHpigKmae3k7Z9MYohA6eWA3EI6I5KI79GARwdHUlISKBYsWJSgApRiCmKgsFgIDY2FhcXlxzbyG9NUaBcj09n1q/RXI5NJy0j+6euro2LUdcnVgpP8dhMybHoN83FePkYdqVfwrnjCDTFSts6Vr5SunRprl+/Tnx8vK2jCCFszM7OjqJFi1K8ePGc1+dxHiGeKZNZ4exVHQajwsmLKfxnT9YfvqA6HvgUd+Be/4u3uz0ta3tw7Fis7cKK546imMk4Ek7a7p9ApcKp7VAc6nZApVLbOlq+Y29vT/ny5W0dQwjxHJDiUzxXMg1m9p+5Q3qmmcQUA3tPJRMdn2FZX7WsM6+1LUnNijKpt3g6ptvR6CLnYLp+FrsKdXFu/y6aoiVsHUsIIZ57UnyKfCclzchfl3Wcv64Hsu7Z/HVvHHYaFRkGxdJOpYJibvYMCylF+ZJO2GlUVCrlhEYGEYmnoJiMpB9YR/r+1ai0Tjh3Hom2eiu5j1EIIZ4RKT5FvmE0KYz+4QJnr+oty9RqQAEFKOftRPXyLpT2cqBe5SIUcdGgtZPLn+LZMd66gD5iNqbYS9hXaYpz27dRu3rYOpYQQhQoUnyKfGNx5E3OXtXTspY7tSu5UqO8K77FHGwdSxQCiiGD9P0/k35gHSrnorj0GIfWr7GtYwkhRIEkxaewqTOXU/luw3X0GWZu3zHg6WbHR73KotHIJU6RN4zRf6GLmI058QbaWkE4tX4DtZPcMyyEELlFik+R5xRFIS7ZwNjFF7mZmIlGDa39PVAUaFHTXQpPkSeUDD1pu38i42g46qLeuPadjn15f1vHEkKIAk+KT5GnEu4amLD0EpdvpQNQxsuBya+Wx7e4XF4Xecdw8Qi6TXNR7t7GoX4ITi1eRaV1tHUsIYQoFKT4FHkiPjmTpVti2HUiGYAujYrzUjkX/Mo44+0hz8MWecOsv0va9h/IPLMTdbEyuL76JXalq9o6lhBCFCpSfIpcdeDsHeZuvEHCXQMANcq7EFC1CD2ayXyJIu8oioIhaj/6LQtQ0lNwbNIHxyZ9UNnZ2zqaEEIUOlJ8ilxhMimM+/ESJy+lUs7bkZa13PGv5ErdykVsHU0UMubURPSbv8fw959oSlbCud907ErIk3iEEMJW8qz4vHz5MmPGjCE5ORl3d3dmzpxJuXLlsrVJSEjgk08+ISYmBqPRSEBAAOPHj8fOTmrk583Pu2I5eSkVR62aL96qiJuTvIcibymKQuapbaRtX4xiMuDUehAODbqiUmtsHU0IIQq1PJuhe9KkSfTr148tW7bQr18/Jk6ceF+bBQsWULFiRcLCwvjtt9/466+/2Lp1a15FFM9AfHIm4X/eZs2uWLw97Fn5STUpPEWes9MnkvrzOPQRs9GUKE+RN+fi2LCHFJ5CCJEP5ElVkJCQwNmzZ1m6dCkAnTp1Ytq0aSQmJuLp6Wlpp1Kp0Ol0mM1mMjMzMRgMeHt750VE8Qwc+yeFcT9eAkAFDO9aGhdH+WMv8o5iNpFxJIzSfyzDqLHDOXgYWv9gVCp5EpYQQuQXeVJ8xsTE4O3tjUaTVYhoNBpKlChBTExMtuLznXfeYfjw4TRt2pS0tDT69+9P3bp1H+tYZ86ceeKcR48efeJtn1dPe86KAldvw5Erak5czfoD36ehkcolgZR/yI8vqbzPBZN9ahxeZzbgeOc66cVfJL5aF0xKUTh23NbR8kxheJ+FEM+/fHU9dPPmzfj5+fHTTz+h0+kYPHgwmzdvJjg42Op9VK9eHQeHx58z8ujRo49d6D7vnvaczWaFNbviWLHnFgDlSjoytHMpalZwfVYRnzl5nwsexWQg/c9fST+wBpXWGeeQj7mU5krdevVsHS1PPen7nJGR8VQf2oUQ4nHlSfHp4+NDbGwsJpMJjUaDyWQiLi4OHx+fbO1WrlzJp59+ilqtxs3NjdatW3Pw4MHHKj5F7klNMxFx8DbRcRkcPn8Xg0khLcNMSQ8tb3f2JaBqUVtHFIWMMeYf9BGzMcVdxr5ac5yD3kbtUpR82eUuhBACyKPis1ixYlStWpXw8HBCQkIIDw+natWq2S65A5QuXZq9e/dSs2ZNMjMz+fPPPwkKCsqLiOIh4u9ksut4Eku33LIs09qpaFfPkzIlHOkYUAy1Wh6JKfKOYkgnbd9qMg6GonLxwKXnBLSVG9o6lhBCCCvk2WX3yZMnM2bMGL7//nuKFCnCzJkzARg8eDAjRoygRo0ajB07lkmTJtG5c2dMJhMBAQG8/PLLeRVR5OA/e+JYujkGgNLFHejSuDgdA4qhUmUNEBMirxmunkYfOQdz0k20tdvh3PoNVI4uto4lhBDCSnlWfFasWJG1a9fet/yHH36wfF22bFnLiHhhW4qisOtEMks3x+DqpOGDHmV4wduRUvIMdmEjSoYe/c6lZB6PRO1eEtd+n2JfrpatYwkhhHhM+WrAkcgfTGaFBWE3CD+QAMDYfi/gX8nNxqlEYWa4cAjdpnkoqYk4NOiGU4sBqOwdbR1LCCHEE5DiU9xnTuh1th5JxL+SK+MHlMPZQebqFLZh1t8hbdsiMv/ajbr4C7h2/wS7UlVsHUsIIcRTsLr4/P3334mIiCAxMZEFCxZw+vRpUlNTadSoUW7mE3nsUNQdth9LpGpZZz7oUUYKT2ETiqJgOLsX/dYFKBl6HJv2w7HJy6g09raOJoQQ4ilZ9diPFStWMHnyZMqVK8fhw4cBcHR0ZPbs2bkaTuQdRVE4eO4uk5dfwcVRw/j+5fBy19o6liiEzCm30f06Dd3GL1C7l6TIoNk4Ne8vhacQQhQQVvV8/vTTTyxbtozSpUtbBghVqFCBy5cv52o4kXdWbI/l552xuLva8cXgingWkT/0Im8pikLmiS3ody4BkwmnwDdwqB8iz2MXQogCxqriU6fTWSaEvze9jtFoxN5eCpTnWaYx63nsRpPC9qOJONir+PrtSvgWkxHtIm+ZkmLQR87BePUUdmVr4NxhBBpPX1vHEkIIkQusKj7r16/PokWLGDp0qGXZ8uXLCQgIyLVgIncpisL8HRriUy5ZlnVv6iWFp8hTitlExuHfSNuzAjQanNsPR1u7LSqVVXcECSGEeA5ZVXyOHz+et99+m7Vr16LT6WjXrh0uLi4sXLgwt/OJXBJ1TU98iooODTxpW68YGjVU8HGydSxRiJjirqCLnI3p5t/Yv9gA5+BhqN2K2zqWEEKIXGZV8VmiRAnWrVvH6dOnuXHjBj4+PtSsWRO1WnonnkfRcelMX3UFjVphUHtfXBzlnjqRdxSTgfQ/1pL++y+oHJxxCRmFfbXm8sQsIYQoJKwqPocOHcr8+fOpWbMmNWvWtCx/9913mTt3bq6FE8/WlsMJbPzjNpdvpQPQra5ZCk+Rp4w3z6OLmI05/iral1riFDQEtXNRW8cSQgiRh6wqPg8ePJjj8kOHDj3TMCJ3ZGSamLXuOntOJQPweruS1PMrQtLNszZOJgoLxZBO2t6VZBzaiMrVA5dek9C+2MDWsYQQQtjAQ4vPe/N4GgyG++b0jI6OxtdXRqPmZ0aTwtLNN9l6NInUNBNt63rwZgdf3Jyz3vajN20cUBQKhisn0UfOwZx8C61/e5xbvY7K0cXWsYQQQtjIQ4vPW7duAVkjo+99fY+Pjw/Dhw/PvWTiqSiKwq9741i//zbFitjx8cvlaVCliK1jiUJESdeh37mEzBNbUHv44Nr/M+xfqPnoDYUQQhRoDy0+P/vsMwD8/f15+eWX8ySQeHpnrqQydvElDCaFCj6OfP32izhqZXCYyDuZ/xxEv2keii4Jh4DuWU8osne0dSwhhBD5gFX3fN4rPFNTU0lKSsq2rkyZMs8+lXhiKXojq7bHYjAp9G5ZgleDSqJWyyhikTfMujvoty3AcHYvaq8XcO05HjvfyraOJYQQIh+xqvi8ePEiI0eOJCoqCpVKhaIolmlRzp07l6sBhfUiDtxm7sYbALSt68Fr7XxsnEgUFoqiYDi7B/3WhSgZehybD8CxUU95HrsQQoj7WFV8Tp48mYCAAJYvX05gYCA7d+7k66+/xt/fP7fzCSv9c0PPiu1Z9+W+06UUzWq42ziRKCzMd+PRb56H4cJhNL5+uHR8D43XC7aOJYQQIp+yqviMiorixx9/xN7eHkVRcHNzY9SoUXTq1ImQkJDczige4cbtDEbM/QeA97qXJrh+MRsnEoWBopjJPL4Z/c4fQTHj1GYwDvU6o1LL3LFCCCEezKri08HBAaPRiL29PR4eHty8eZMiRYqQnJyc2/nEIyTeNTBxWdbz2b98qyLVy7naOJEoDEyJN9BHfofx2mnsytXCuf1wNB5ym4cQQohHs6r4rFu3Lps2baJ79+60a9eOwYMHo9VqadiwYW7nEw+QaTDz9a/R7P3/ieOnvVZeCk+R6xSziYxDG0jbuxKVxh7nDiPQ1morj8YUQghhNauKz39PMP/hhx9SqVIl9Ho93bp1y7Vg4uHG/niJv67ocHXS0Ku5F/X8ZA5PkbuMcZfRR8zGFPMP9pUb4tzuHdRucouHEEKIx2NV8flvarWarl27kpmZydq1a+nfv39u5BIPcD5ax4zVV4lPNtCqljsf9y4rvU4iVylGA+m/ryH9z7WoHN1w6TYG+ypN5edOCCHEE3lk8fnnn39y7tw5ypYtS5s2bTAajaxevZoffvgBd3d3KT7z2MrtscQnG3ijvQ8tarpLASBylfFGFLqI2ZhvX0NbvTVObQajdpZediGEEE/uocXnokWLmD9/PpUqVeLChQv07duXQ4cOodVqmTZtGi1btsyjmOKeG7czqF7OhZ7NS9g6iijAlMx00vYsJ+Pwb6iKFMe19xTsK9azdSwhhBAFwEOLz19++YUVK1ZQvXp1Tpw4Qd++fRk9ejSvvfZaHsUT/7bzeBIxiZnyjHaRqwyXj6OP/A7znVgc6nTEqdVrqBycbR1LCCFEAfHQ4jMpKYnq1asDULt2bbRaLQMHDsyTYCK7XccT+fI/0QC0q+9p4zSiIDKnp5K2YzGZJ7eh9vTFdcBM7MtWt3UsIYQQBcwj7/lUFMXyz8HBAQCz2WxZr1arcy+dALLeg+3Hk3B11LB8TFWcHGQSb/FsZZ7/E/2W71F0yTg26oVj076o7B1sHUsIIUQB9NDiU6/XU61aNcv3iqJYvr/3fHd5tnvuW7cvjmP/pNK0elEpPMUzZU5NQr91AYao/WhKlMe510TsfF60dSwhhBAF2EOLzx07duRVDvEA56N1/LwzDm8Pe0b3kedli2dDURQyz+wibfsilMw0HFu8imPDHqg0jz37mhBCCPFYHvqXplSpUnmVQzzA+B8voSgw5dXy2GlkWiXx9Mx34tBtmovx0lE0pavi0mEEmuJlbR1LCCFEIZFn3RyXL19mzJgxJCcn4+7uzsyZMylXrtx97SIjI5k/f77lsv7SpUspXrx4XsXMV+KTM0hNN9OqljsvlHSydRzxnFMUMxlHI0nbvQwUBaegt3Co1wmVSu7bFkIIkXfyrPicNGkS/fr1IyQkhI0bNzJx4kSWL1+erc3p06eZO3cuP/30E15eXqSkpKDVavMqYr4Rm5TJV/+5xoWbaWjUMCDI29aRxHPOlHAdfcQcjNf/wq68P87th6Nxl58rIYQQeS9Pis+EhATOnj3L0qVLAejUqRPTpk0jMTERT8//Thu0bNkyBg0ahJeXFwBubm55ES9fiU/O5I2vzmEyQ2t/D5rVKIpvMUdbxxLPKcVsouilvdzdvgeVvRbnTu+jrdFGnowlhBDCZh6r+IyJiSE2NpbatWs/1kFiYmLw9vZGo8kaqa3RaChRogQxMTHZis+LFy9SunRp+vfvj16vJygoiKFDhxaaP5QZBjPjfryEyQxDOvrSramXrSOJ55gx9iL6iNkUu3URe7/GOLcbitpV5ogVQghhW1YVnzdv3uTDDz8kKioKlUrF8ePH2bx5M/v27WPGjBnPLIzJZOL8+fMsXbqUzMxM3nzzTXx9fenatavV+zhz5swTH//o0aNPvO2zsPaQmuh4NT3qmyjrdI2jR6/l+jFtfc62UNDPWWUy4H5pD+6X92OydyahVm90JV+C85eBy7aOl2cK+vuck8J4zkKI549VxefEiRNp2bIlq1evJiAgAIAmTZowc+ZMqw7i4+NDbGwsJpMJjUaDyWQiLi4OHx+fbO18fX0JDg5Gq9Wi1WoJDAzk1KlTj1V8Vq9e3TIZ/uM4evQodevWfeztnpU/ziRz8tpVXJ00vNGtZp709tr6nG2hoJ+zMfosuohFmBOvo63ZBqfAN7l29u8Cfc45Kejvc06e9JwzMjKe6kO7EEI8LquGuZ4+fZohQ4agVqstRZGbmxspKSlWHaRYsWJUrVqV8PBwAMLDw6latWq2S+6QdS/o/v37URQFg8HAgQMHqFKlyuOcz3PrYkwaAN+PqFxobjMQz46SmYZ+6wJSVowCUyaufabh0ukD1E6F775pIYQQ+ZtVxWexYsW4evVqtmUXLly4r+fyYSZPnszKlStp164dK1euZMqUKQAMHjyY06dPA9CxY0eKFStGhw4d6Nq1K5UqVaJnz55WH+N5duDcXUq42+PlXvhG94unY7h0jLs/vEPGkXAc6nWiyODvsa9Qx9axhBBCiBxZddl90KBBvP322wwZMgSj0Uh4eDgLFy5k8ODBVh+oYsWKrF279r7lP/zwg+VrtVrNJ598wieffGL1fguCc9d0XIpJp21dGQwirGdOSyFtx2IyT21H7Vkat1e+wK5MtUdvKIQQQtiQVcVnz549cXd355dffsHHx4cNGzbw3nvv0aZNm9zOV+ApisKWw4kAvNq2pI3TiOdFZtTv6Ld8j6K/i2Pjl3Fs2heVnfSaCyGEyP+sKj5NJhNt2rSRYvMZM5kV5v92gy1HEqn7oivFitjbOpLI58ypiei3zMdw/g803hVx7jMVO++Kto4lhBBCWM2q4rNJkyYEBwfTuXPnQjeCNDftOZlMxMEESrjbM65/OVvHEfmYoihknt5O2vYfUAyZOLV6DYeA7qjUGltHE0IIIR6LVcXnjz/+SHh4OCNHjkStVtOxY0c6deqEn59fbucr0LYcybrc/vmbFXFykCJC5MyUHIt+03cYLx/HrvRLOHccgaZYaVvHEkIIIZ6IVcVntWrVqFatGqNGjeLQoUOEh4czcOBAvLy8CAsLy+2MBdJdnZFTl1Lp0MATn2KPPy+pKPgUs4mMoxGk7f4JVCqc2g3FoU4HVCqrJqkQQggh8qXHfrZ7hQoVqFixIr6+vly5ciUXIhUOCSkGAGpVdLVxEpEfmW5fQxc5B9P1c9hVqItz+3fRFC1h61hCCCHEU7Oq+Lx79y5btmwhPDyckydP0qRJE958800CAwNzO1+BFZ+UCYCHqwwyEv+lmIykH/iV9P0/o9I64dx5JNrqreTBA0IIIQoMq4rPZs2a4e/vT6dOnfjuu+8oUqRIbucq8PaduYOdRkWlUk62jiLyCeOtC+jDv8UUdxn7Kk1xbvs2alcPW8cSQgghnimris9t27ZRooRc8ntWjl9IYfuxJDo3LCYDjQSKIYO0/avJOLAelXNRXHqMQ+vX2NaxhBBCiFzxwOLz8OHD1K9fH4CLFy9y8eLFHNs1atQod5IVYFuPJKJWwWvB1j+eVBRMhmtn0EfOwZx4A22ttji1HiTPYxdCCFGgPbD4nDJlCuHh4QCMGzcuxzYqlYodO3bkTrICKj3TzIFzd2lX3xNn6fUstJQMPWm7lpFxLAJ1UW9c+07Hvry/rWMJIYQQue6Bxee9whNg586deRKmMAjdH096ppmWteRevsLKcPEIuk1zUe7exqF+CE4tXkWldbR1LCGEECJPWDVh4NChQ3Nc/u677z7TMIXB4fN3AXipnIuNk4i8ZtbfRffb16T+MgmVvSNur36Jc9AQKTyFEEIUKlYNODp48GCOyw8dOvRMwxQGN25n0LKWOxq1TJ1TWCiKguHcfvRb56Okp+LYpA+OTfqgspNptoQQQhQ+Dy0+Z8+eDYDBYLB8fU90dDS+vr65l6wA0qWbuKs34VVUio7CwpySgH7L9xj+PoCmZCWc+07HzruCrWMJIYQQNvPQ4vPWrVtAVs/Nva/v8fHxYfjw4bmXrADadzoZAP9KMpq5oFMUhcyTW0nbsQTFZMCp1es4BHRDpZZBZkIIIQq3hxafn332GQD+/v68/PLLeRKoILtyKx21CqqXl/s9CzJTUgz6yO8wXj2JXZnqOHccgcazlK1jCSGEEPnCA4vP69evU7p0aSBrLs/o6Ogc25UpUyZ3khVAl2+lUcJdi72dVeO8xHNGMZvIOBJG2p7loFLjHDwMrX8wKpW830IIIcQ9Dyw+O3fuzPHjxwEICgpCpVKhKEq2NiqVinPnzuVuwgLCZFK4GptOmRIysrkgMsVfRRcxG9PN89hVrIdL+3dRF/GydSwhhBAi33lg8Xmv8ASIiorKkzAF2fr98dzRmehfo6ito4hnSDEZSP/zV9L3r0Hl4Ixzl4/QvtQSlUpmMxBCCCFyYtVUS/8rOjoalUpluSwvHm3z4QSqveBMw2pSfBYUxpt/o4+YjSn+CvbVmuMc9BZqF3dbxxJCCCHyNatuRvvwww85duwYAOvWraNjx4506tSJtWvX5mq4guLizTRuJmTSoqYHXkW1to4jnpJiSEe/YwkpP43EnJaCS88JuHYdLYWnEEIIYQWris8///yT6tWrA7Bs2TKWLl3K2rVr+eGHH3I1XEGxdk8cdhoVrWpLcfK8M1w9xd3Fw8k4uB5trSCKDpmPtnJDW8cSQgghnhtWXXY3GAxotVpiY2NJTk6mbt26ANy+fTtXwxUE56P17DmVTIMqRXBzfqK7HEQ+oKTr0O9aSubxTajdS+La71Psy9WydSwhhBDiuWNVNVS1alUWLlzIjRs3aNmyJQCxsbG4urrmZrYCYeaaqwC8EuRt4yTiSRkuHEK3aS5KahIOAd1waj4Alb3MWiCEEEI8CauKzxkzZjB79mzs7Oz4+OOPgazR8J07d87VcM+7xLsGYhIzqeDjSEUfJ1vHEY/JrLtD2vZFZP61G7XXC7j2GIedr5+tYwkhhBDPNauKz7Jly/L1119nWxYcHExwcHCuhCooNh1KAOD9HmVk6p3niKIoGM7uRb91AUqGHsdm/XFs3AuVxt7W0YQQQojnntU3Ia5bt46NGzcSGxuLt7c3ISEh9OjRIzezPfdOX0lFrYJKvtLr+bww372Nfss8DP8cQuNbGZcO76EpUc7WsYQQQogCw6ric/78+WzYsIFBgwbh6+vLzZs3Wbx4MXFxcQwdOjS3Mz6XMgxmLtxIo2Utd+n1fA4oipnME1vQ7/wRTCacAt/EoX4XVGqNraMJIYQQBYpVxefatWtZsWIFpUqVsixr2rQpAwYMkOLzAf66okOXbqZZDZleKb8zJd5EHzkH47XT2L1QE+cOI9B4+Ng6lhBCCFEgWVV8pqWl4enpmW2Zu7s76enpuRKqIDh7VYdaBTUqyIwA+ZViNpFxeCNpe1aCRoNzhxFoa7WVnmohhBAiF1k1yXyzZs346KOPuHTpEunp6Vy8eJExY8bQtGlTqw90+fJlevfuTbt27ejduzdXrlx5YNtLly5Rq1YtZs6cafX+85u/rugoV9IRF0e5bJsfmeKuN3IVwQAAIABJREFUkPLTR6TtWIJ9+VoUHTIfh9rtpPAUQgghcplVxefEiRNxcXGhS5cu+Pv707VrV5ycnJgwYYLVB5o0aRL9+vVjy5Yt9OvXj4kTJ+bYzmQyMWnSJNq0aWP1vvMbk0khKlrPSy+42DqK+F9mI2l7V3H3x/cw34nFpetoXHpORO1W3NbJhBBCiELhkZfdU1JSuHbtGhMnTuTzzz8nKSkJDw8P1Gqr6lYAEhISOHv2LEv/r707j46yPtQ4/p2Z7AkhBBNIREQoQthkCSCVWEQ0gMOiENMKHJVeoLhgvfd6QCjI1ipePQoIl0ILVrGiubWgIQJF2WWNUcSwyI4mJJAQAplJZjLz3j+QVJbgJCQzWZ7POZzDDO+87/ObOUyevNtv2TIArFYrs2bNIj8//5rD+YsXL6ZPnz7YbDZsNlsFh1MzHD1tp9jhpl0Llc+apDTrIM22L6L4Yi4B7fsQ/MBYzCENfR1LRESkXrlh+dy4cSO///3vKS4uJjQ0lAULFnD33RWfxzo7O5smTZpgsVw6BG2xWIiOjiY7O/uK8nngwAG2bt3KO++8w8KFCyu8HYB9+/ZV6nUA6enplX7tT23/zgRYcBUepYpWWW2qasw1manUQaPDn9PwxHbMQQ3I7joCe1Qb2H/Y19G8pj58zlfTmEVEaqYbls+5c+fy3//93wwbNowPP/yQN998kxUrVlRLEKfTydSpU3n55ZfLSmpldOjQgcDAwAq/Lj09vWzO+pu17uBxoiNs3J9Qs+f+rsox11TO419jS1uIu+A0gV0HcqBRZ7r2vMfXsbyqPnzOV9OYPVdSUnJTv7SLiFTUDcvnqVOnGDlyJAAjRoxg0aJFldpITEwMOTk5uFwuLBYLLpeL3NxcYmL+fTubM2fOcPLkScaOHQtAYWEhhmFw8eJFZs2aVant+oJhGHx7ooiOd+gqd19yF1/E/vlSHF+txdwolrARr+B/e0cM7RkSERHxqRuWT7fb/e8F/fxwuVyV2kjjxo2Ji4sjNTWVIUOGkJqaSlxc3BWH3GNjY9m5c2fZ4/nz52Oz2Zg4cWKltukruQVO8gpLdbGRDzkO7cC2ZiFG0TkC7x5GcMIITP4V3xsuIiIiVe+G5bO4uJgRI0aUPS4qKrriMcB7773n0YamT5/OpEmTWLhwIeHh4WW3URozZgwTJkygY8eOFc1eI+0+UAigi418wF1UgG3dn3Hu34wlqgUhSVPxi2nt61giIiLyEzcsn3/84x+veDx8+PBKb6hVq1akpKRc8/ySJUuuu/yzzz5b6W350vovzxEaZOb2JkG+jlJvGIaB49uN2P+1GMNhI+jeUQT1GobJ4u/raCIiInKVG5bPhx9+2Fs56oSNX5/j4Pc2RvZrgsWsm5V7g7vwDEWfvkXpkT1Ybm1L6MDnsEQ193UsERERKYdH02vKz7tgK+X1lFOYTTAsIcrXceo8w3Dj+PJTbBuWgeEmuN9YAuOtmMyaUUpERKQmU/msIp9/dY5Sl8F/Jd1GUIAKUHVy5f+AbfU8Sk/tw69FZ0IGPosloqmvY4mIiIgHVD6ryOa9BcREBtCva+TPLyyVYrhdlOz8J/Yt72Gy+BPy0HMEdHpA87GLiIjUIiqfVWDf8YtknrDRv7uKZ3UpzTmKbfVcXKcP439nL0ISx2Nu0NjXsURERKSCPCqfDoeDBQsWkJqaSkFBAenp6WzdupXjx4+X3YS+Pvtkex4Aw++N9nGSuscodVK8bQXF21MwBTUg9OEX8W97j/Z2ioiI1FJmTxb605/+xKFDh3jttdfKfui3bt2a999/v1rD1QaHf7CxeW8BUQ39ufUW3ci8KpV+v5/Cvz5L8bYVBLTvQ/jY/yUgrreKp4iISC3m0Z7P9evXs27dOkJCQjCbL/XVJk2akJOTU63haoM9hy4A8F9Jt/k4Sd1hOOzYN71Lye6PMYffQljyDPxbxfs6loiIiFQBj8qnv7//NVNr5ufnExERUS2hapP8QidhQRbuatXA11HqBOexDGxp83GfzyGwm5XgPo9jCgzxdSwRERGpIh4ddu/fvz8TJ07k1KlTAOTm5jJz5kweeuihag1X05274GTDVwU0i9Lh9pvltl+gaPWbXHz/D2DxI2zkHEISx6t4ioiI1DEelc/nn3+eZs2aMXjwYAoLC0lMTCQ6Opqnn366uvPVaB9szOFisYvk+3Sh0c1wHPyCwsXjcez9jKBeSYT/x1v4N+/g61giIiJSDTw67B4QEMDkyZOZPHky+fn5NGrUSBd9AKfOlABwd1xDHyepndwXz2Fbtwjnga1YmrQkJHk6fk1/4etYIiIiUo08Kp+XD7dfVlRUVPb3226rvxfaZOU5aNE0yNcxah3DMHDs+xz7vxZjOEsI6vM4QT0fwWTRbWdFRETqOo9+2j/wwKVZZAzDKHvu8p7P/fv3V0+yGs4wDAqLSomOCPZ1lFrFdT4X26dvUXo0HUuzOEIHPofllvr7C4yIiEh941H5PHDgwBWPz5w5w1tvvUV8fP29/c26PfnYStz07dzI11FqBcNwU5Kehn3j22AYBD8wjsB4KyaTR6cdi4iISB1RqeOcUVFRTJkyhcTERAYNGlTVmWo8wzD4c2oWAH1UPn+WK+97bKvnUfr9t/jd0ZWQAc9giWji61giIiLiA5U+ye7o0aPY7faqzFJr/HC2BLvDTXKfaAL9teeuPIarlJKdH2Hf8ndM/oGEWJ8noOP9ulhNRESkHvOofD722GNXFAa73c7hw4fr5a2WDMNg49cFAPTrqr2e5Sk9fQTb6rm4co7g3/YeQh78HeawSF/HEhERER/zqHwmJSVd8Tg4OJi2bdvSokWL6shUo63Znc97n+UQ2cBPc7lfh1HqoHjr+xRv/z9MIeGEPjKZgLb3+DqWiIiI1BA/Wz5dLhc7duxg1qxZBAQEeCNTjbb+y3z8LCZmPH6HDh9fpfRUJkWr5+LO/56ATg8QfP9vMQdr2lERERH5t58tnxaLhW3btqloATnnHGSesPFEYlN+caumfbzMKLFh3/g3StJXY24YRdivZ+HfsquvY4mIiEgN5NHVMo8//jjz58/H6XRWd54abXvmeQDu7RTh4yQ1h/NoOoVLnqIkfTWB8YMIH7NQxVNERETKdcM9n6mpqVitVpYvX87Zs2dZtmwZkZGRV+wF3bhxY3VnrDFW78yjQbCFmEid6+m2X8C+fgmObz7D3LgZDUa9it9t7XwdS0RERGq4G5bPadOmYbVa+Z//+R9v5amxNu09x/dnSrg7LtzXUXzOcWAbtrULMewXCPplMkG9f43JT+cDi4iIyM+7Yfm8PJ1mjx49vBKmJvvnlrOEBVt47pH6OxWk+2I+trX/i/PgF1iatiLk1zPxa9LK17FERESkFrlh+XS73ezYseOKOd2v1qtXryoPVdMYhsGhH2z07dyIiLBK35e/1jIMA8c367GvX4LhdBB83xME9nwEk9ni62giIiJSy9ywSTkcDqZMmVJu+TSZTHz22WfVEqwmyS1wYhjQIKT+lS1XQQ62T+dTeiwDv9vaEzLwOSyNb/V1LBEREamlblg+g4OD60W5/Dn5Fy5d5d+iaZCPk3iP4XZRkr4a+8a/gclEcOJ4ArsOxGTSdKIiIiJSefXvGHIlXLS7AGgeVT/Kp+vsSYrS5uH6fj9+reIJ7f805obRvo4lIiIidYBHFxzVd5fLZ1hw3T7sbrhKKd7xfxRvfR9TQDAhg/+LgPb3aYIBERERqTI3LJ8ZGRlVtqFjx44xadIkCgoKiIiIYM6cOdfMDb9gwQLS0tIwm834+/vz/PPPk5CQUGUZKuvMeQdQt8/5LD19GFvqm7hyj+Efl0DIg7/DHKqb6YuIiEjV8tph95deeonHHnuMIUOGsGrVKqZNm8Y777xzxTKdOnVi9OjRBAcHc+DAAUaOHMnWrVsJCvLt4e79J2yEh1hoGFr3zlIwnCXYt/6dkh0fYQqNIHTYHwhoU/fvYCAiIiK+4ZWrR/Ly8sjMzMRqtQJgtVrJzMwkPz//iuUSEhIIDg4GoE2bNhiGQUFBgTcilsswDA59b6Pd7aF17vBz0LnjFP71WUq2/x8BnfoRPvZ/VTxFRESkWnllV152djZNmjTBYrl02NpisRAdHU12djaRkZHXfc3KlStp3rw5TZs29UbEcmXnO8i/UEq3Oxv4NEdVMkps2De8TeyXqyGiCWG/+SP+d3T2dSwRERGpB2rkceRdu3Yxd+5cli5dWuHX7tu3r9LbTU9Pv/a54ybAgtl2gvT0E5Ved00RfOYQUZmfYCku5PztvTj3i/sx8l2Qf+3Y66rrfc51ncZcP9THMYtI7eOV8hkTE0NOTg4ulwuLxYLL5SI3N5eYmJhrls3IyOCFF15g4cKFtGzZssLb6tChA4GBgRV+XXp6Ot26dbvm+Xd2HCI81MGA+zrV6sPublsh9vVLcOz7HPMtzQlNnsax00XXHXNdVt7nXJdpzPVDZcdcUlJyU7+0i4hUlFfO+WzcuDFxcXGkpqYCkJqaSlxc3DWH3Pfu3cvzzz/PvHnzaN++vTei3ZBhGBzJstO0UUCtLZ6GYeDI3ELh4t/hyNxEUO/fED56Hn63tvV1NBEREamHvHbYffr06UyaNImFCxcSHh7OnDlzABgzZgwTJkygY8eOzJgxg+LiYqZNm1b2uldffZU2bdp4K+YVTuaW4HJD386NfLL9m+W+kIdt7UKch3ZgiWlNyGN/xC/6Dl/HEhERkXrMa+WzVatWpKSkXPP8kiVLyv7+j3/8w1txPHLwlA2AO2KCfZykYgzDwPH1Ouyf/RXD5SS472gCewzFZK679ykVERGR2qFGXnBUU+w9epHwEAsdWoT6OorHXOeysaXNp/TE1/g170jIwAlYImN9HUtEREQEUPm8oYPf27gtKhCzueaf72m4XZTs+Rj7pnfBZCak/zMEdEnEZPLKab0iIiIiHlH5LIfbbZBzzkHHO2r++Z6uMycoWj0XV9ZB/H/RnZD+z2AOv8XXsURERESuofJZjux8B85Sg9ub+HZqzxsxXE6Kv0iheNsHmAJDCB3yAv7tflVrr8wXERGRuk/lsxw550oAiIms+D1DvaE06xC21XNxnTmOf7tfEfLAOMyhDX0dS0REROSGVD7LcSSrGIA7YmrWnk/DWYx983uU7FqJKawRoUnTCGjd09exRERERDyi8lmOzBNFREf4E9UwwNdRyjhP7MWWNh/3uSwCOvcnpO9oTEG150p8EREREZXP63C7Db4+epHudzbwdRQAjOIibBuW4cj4FHOjGMIe+xP+Le7ydSwRERGRClP5vI6cAgf2Ejfta8D9PR3f7cK25i2Mi+cI7PkIwfeOwORfs04FEBEREfGUyud1rN2dD8AdTX03s5G76Dy2f/0ZZ+YmzFG3EzZsCn6xvplmVERERKSqqHxex6e78gDfTKtpGAbOzE3Y1v0Zo8RGUMIIgn6ZhMni7/UsIiIiIlVN5fMqLrfBBZsL692NCQv27lzo7sKz2NYswHl4F5bYOwkd+ByW6BZezSAiIiJSnVQ+r3IqtxgDaBblvft7GoYbx1drsX2+FFwugvv9B4HxgzGZvVt+RURERKqbyudVdh0oBOCe9t65YbsrPwtb2jxKT36D3+13ETLwWSyNYryybRERERFvU/m8yvdnS2gU5sct1Xx/T8PtomT3KuybloPFQsjACQTc9aCmxhQREZE6TeXzKll5DhqEVO/hblfucYpWz8WVfQj/1j0J6f8U5ga3VOs2RURERGoClc+r5J5zcEvD6rmy3Ch1UvzFBxR/8SGmoDBCh07EPy5BeztFRESk3lD5/InT+SWcOe+kV7uqP9+z9IcDFK2eh/vsCQI63EdwvzGYQ7xzXqmIiIhITaHy+ROfZZwDoFe78Cpbp+Eoxr75XUp2rcLUoDFhj76E/y96VNn6RURERGoTlc+fOJplB6BTy7AqWZ/z+NfY0ubhLjhNYNeBBN/3JKbAkCpZt4iIiEhtpPL5E4d+sNOlVRhm882dg+kuvoj986U4vlqLOTKWsJFz8G/eoYpSioiIiNReKp8/crog77yTezve3HmYjkPbsa1ZiFFUQODdwwhOGIHJ33s3rBcRERGpyVQ+f3TeBgYQHVG5+3u6iwqwrfszzv2bsUTfQUjSNPxiWldtSBEREZFaTuXzR9kFlw61RzeqWPk0DAPHtxuw/2sxhsNO0L2jCOo1HJNFb62IiIjI1dSQfpR7aVZN4pqHevwad+EZij59i9Ije7Dc2pbQgc9hiWpeTQlFREREaj+Vzx8VlZgwmyDcg9mNDMON48tPsW1YBoab4H5jCYy3YjJX78xIIiIiIrWdyuePbCUQFmz52SvdXXk/YEubR+mpffi16EzIwGexRDT1UkoRERGR2k3l80cXSqBBcPl7Lg23i5Kd/8S+5T1MFn9CHnqOgE4PaGpMERERkQpQ+fyRvcTELZHXfztKc45iW/0mrtNH8L+zFyGJ4zE3aOzlhCIiIiK1n8rnj2xOiAjzv+I5o9RJ8bYVFG9PwRTUgNCHX8S/7T3a2ykiIiJSSSqfXLpdkt0BjcL+/XaUfr+fotVzceedIqDj/QTf/x+YQ6puzncRERGR+shr5fPYsWNMmjSJgoICIiIimDNnDi1atLhiGZfLxezZs9myZQsmk4mxY8eSlJRU7dlsJW5cbhNNGgVgOOzYN71Lye6PMYffQljyDPxbxVd7BhEREZH6wGvl86WXXuKxxx5jyJAhrFq1imnTpvHOO+9cscwnn3zCyZMnWbduHQUFBQwdOpRevXrRrFmzas1WcLEUgOjCbylc8i7u8zkEdrMS3OdxTIEh1bptERERkfrE7I2N5OXlkZmZidVqBcBqtZKZmUl+fv4Vy6WlpZGUlITZbCYyMpJ+/fqxZs2aas+Xn1/ImMB/0Omb18DiR9jIOYQkjlfxFBEREaliXtnzmZ2dTZMmTbBYLt3KyGKxEB0dTXZ2NpGRkVcsFxsbW/Y4JiaG06dPV2hb+/btq3C+s5n7udc/g9PNErC37YNxpgTOpFd4PbVRenr9GOdPacz1g8YsIlIz1bkLjjp06EBgYGCFXmN07cKObS3p1fuX1ZSqZkpPT6dbt26+juFVGnP9oDF7rqSkpFK/tIuIVJZXDrvHxMSQk5ODy+UCLl1YlJubS0xMzDXLZWVllT3Ozs6madPqnz3IZDITEFyxwioiIiIiFeeV8tm4cWPi4uJITU0FIDU1lbi4uCsOuQP079+flJQU3G43+fn5rF+/nsTERG9EFBEREREv8Er5BJg+fTrLly8nMTGR5cuXM2PGDADGjBnDN998A8CQIUNo1qwZDz74II8++ihPP/00t912m7ciioiIiEg189o5n61atSIlJeWa55csWVL2d4vFUlZKRURERKTu8dqeTxERERERlU8RERER8RqVTxERERHxmjpzn0/DMABwOByVXkdJSUlVxak1NOb6QWOuHyoz5svfmZe/Q0VEqpvJqCPfOBcuXODQoUO+jiEiUivdeeedNGjQwNcxRKQeqDPl0+12U1RUhL+/PyaTyddxRERqBcMwcDqdhIaGYjbrTCwRqX51pnyKiIiISM2nX3NFRERExGtUPkVERETEa1Q+RURERMRrVD5FRERExGtUPkVERETEa1Q+RURERMRrVD5FRERExGvqVfk8duwYycnJJCYmkpyczPHjx69ZxuVyMWPGDPr168cDDzxASkqK94NWIU/GvGDBAh566CEGDRrEI488wpYtW7wftAp5MubLjh49yl133cWcOXO8F7AaeDrmtLQ0Bg0ahNVqZdCgQZw9e9a7QauQJ2POy8tj7NixDBo0iAEDBjB9+nRKS0u9H7YKzJkzh759+9KmTZtyZ3Ora99fIlJHGfXIqFGjjJUrVxqGYRgrV640Ro0adc0y//znP43Ro0cbLpfLyMvLMxISEoxTp055O2qV8WTMmzdvNmw2m2EYhrF//36jW7duht1u92rOquTJmA3DMEpLS42RI0ca//mf/2m88sor3oxY5TwZ8969e40BAwYYubm5hmEYRmFhoVFcXOzVnFXJkzHPnj277LN1OBzG8OHDjdWrV3s1Z1XZvXu3kZWVZdx3333GwYMHr7tMXfv+EpG6qd7s+czLyyMzMxOr1QqA1WolMzOT/Pz8K5ZLS0sjKSkJs9lMZGQk/fr1Y82aNb6IfNM8HXNCQgLBwcEAtGnTBsMwKCgo8HrequDpmAEWL15Mnz59aNGihZdTVi1Px/z2228zevRooqKiAGjQoAGBgYFez1sVPB2zyWSiqKgIt9uNw+HA6XTSpEkTX0S+afHx8cTExNxwmbr0/SUidVe9KZ/Z2dk0adIEi8UCgMViITo6muzs7GuWi42NLXscExPD6dOnvZq1qng65p9auXIlzZs3p2nTpt6KWaU8HfOBAwfYunUrTzzxhA9SVi1Px3zkyBFOnTrFiBEjePjhh1m4cCFGLZ1d19MxP/XUUxw7dozevXuX/enWrZsvIntFXfr+EpG6q96UT/l5u3btYu7cubz++uu+jlKtnE4nU6dOZcaMGWXlpT5wuVwcPHiQZcuW8e6777J582ZWrVrl61jVas2aNbRp04atW7eyefNm9uzZoz2BIiI+Vm/KZ0xMDDk5ObhcLuDSD+Lc3NxrDmPFxMSQlZVV9jg7O7vW7gX0dMwAGRkZvPDCCyxYsICWLVt6O2qV8WTMZ86c4eTJk4wdO5a+ffvyt7/9jQ8//JCpU6f6KvZN8fRzjo2NpX///gQEBBAWFsb999/P3r17fRH5pnk65uXLlzN48GDMZjMNGjSgb9++7Ny50xeRvaIufX+JSN1Vb8pn48aNiYuLIzU1FYDU1FTi4uKIjIy8Yrn+/fuTkpKC2+0mPz+f9evXk5iY6IvIN83TMe/du5fnn3+eefPm0b59e19ErTKejDk2NpadO3fy+eef8/nnn/P444/z6KOPMmvWLF/Fvimefs5Wq5WtW7diGAZOp5MdO3bQtm1bX0S+aZ6OuVmzZmzevBkAh8PB9u3bad26tdfzektd+v4SkbrLZNTWk74q4ciRI0yaNInCwkLCw8OZM2cOLVu2ZMyYMUyYMIGOHTvicrmYOXMm27ZtA2DMmDEkJyf7OHnleTLmYcOG8cMPP1xxIcarr75KmzZtfJi88jwZ80/Nnz8fm83GxIkTfZT45nkyZrfbzZw5c9i8eTNms5nevXszceJEzOba+TuoJ2M+efIkL730EmfPnsXlctGzZ0+mTJmCn5+fr+NX2OzZs1m3bh1nz56lUaNGREREsHr16jr9/SUidVO9Kp8iIiIi4lu1c5eHiIiIiNRKKp8iIiIi4jUqnyIiIiLiNSqfIiIiIuI1Kp8iIiIi4jUqn1KrjRo1ipSUFF/HuKGPP/6Y0aNHl/vve/bs0b0YRUSk3lD5lBqjb9++dOrUiS5dupT9ycnJ8XqOUaNG0bFjR7p06ULPnj155plnyM3NrfT6Bg8ezNKlS8set2nThhMnTpQ9jo+PZ+3atTeV+Xrmz59P+/bt6dKlC/Hx8fz6178mIyPD49dfnVNERKQqqHxKjbJo0SIyMjLK/vz0xvfeNG3aNDIyMli7di2FhYW8/PLLPslxswYMGEBGRgY7duygZ8+ePPfcc76OJCIi9ZzKp9Ro58+fZ9y4cdx99910796dcePGcfr06esue+LECUaOHEm3bt3o2bMnv//978v+7ciRIzz55JP06NGDxMRE0tLSPNp+REQEiYmJfPfddwB8+eWXDBs2jG7dujFs2DC+/PLLsmU/+ugj7r//frp06ULfvn35+OOPy57/zW9+A8CIESMAGDJkCF26dCEtLY2dO3dy7733ArB48WImTJhwRYbZs2cze/ZsAC5cuMDkyZPp3bs3CQkJvPHGG2Xzm9+In58fgwYNIicnh/z8fODStKrJycnEx8fTu3dvZs6cicPhKDcnwIYNGxgyZEjZntQDBw549D6KiIhcpvIpNZrb7eaRRx5hw4YNbNiwgcDAQGbOnHndZefOncs999zD7t272bx5MyNHjgTAZrMxevRorFYrX3zxBW+88QYzZszg8OHDP7v9/Px81q5dS1xcHAUFBYwbN45Ro0axc+dOnnzyScaNG8e5c+ew2WzMnj2bJUuWkJGRwYoVK4iLi7tmfe+99x4Aq1atIiMjg4EDB17x7w899BCbNm3i4sWLALhcLtasWYPVagVg0qRJ+Pn5sW7dOlauXMm2bds8OufV4XCwcuVKIiIiCA8PB8BsNvPiiy+yY8cOVqxYwfbt2/n73/9ebs7MzEwmT57MzJkz2blzJ8nJyTz11FNlhVVERMQTKp9Sozz99NPEx8cTHx/PU089RaNGjUhMTCQ4OJiwsDDGjx/P7t27r/taPz8/srKyyM3NJTAwkPj4eAA2btzIrbfeyrBhw/Dz86Ndu3YkJiayZs2acnPMnj2b+Ph4hgwZQlRUFC+++CIbN27k9ttvZ+jQofj5+WG1WmnZsiUbNmwALpW57777juLiYqKjo2ndunWFx3/rrbfSrl071q9fD8COHTsICgqic+fOnD17lk2bNjF58mRCQkJo3LgxTzzxBKtXry53fWvWrCE+Pp677rqLlJQU5s2bVzaveYcOHejcuTN+fn40a9aM5OTkct9bgA8++IDk5GTuuusuLBYLDz/8MP7+/nz11VcVHqeIiNRffr4OIPJTCxYs4Je//GXZY7vdzssvv8yWLVs4f/48AEVFRbhcLiwWyxWvfeGFF5g7dy7Dhw+nYcOGPPnkkwwfPpwffviBvXv3lpVRuLRHcfDgweXm+MMf/kBSUtIVz+Xm5hIbG3vFc7GxseTk5BASEsIbb7zB0qVLmTJlCl27dmXixIm0atWqwu+B1WolNTWVoUOHkpqaWrbXMysri9LSUnr37l22rNvtJiYmptx19e/fn9er/OU8AAADHklEQVRee438/HwmTJjAt99+S8+ePQE4duwYr7zyCvv27cNut+NyuWjfvn2568rKymLlypUsX7687Dmn03lTF2OJiEj9o/IpNdrSpUs5duwYH374IVFRUezfv5+hQ4diGMY1y0ZFRZWdG7lnzx6efPJJunfvTkxMDN27d2fZsmU3lSU6OpqsrKwrnsvOziYhIQGAhIQEEhISKC4u5s0332Tq1Kllh7ErYsCAAcyZM4fTp0/zr3/9iw8++ACApk2bEhAQwI4dO8r2XnoqMjKSmTNnMmzYMKxWK9HR0UyfPp127drx+uuvExYWxttvv33Dq+5jYmL43e9+x/jx4ys8JhERkct02F1qtKKiIgIDAwkPD6egoIC33nqr3GU//fTTsouRGjZsiMlkwmw206dPH44fP87KlStxOp04nU727t3LkSNHKpTlV7/6FcePH+eTTz6htLSUtLQ0Dh8+TJ8+fTh79izr16/HZrMREBBASEgIZvP1/3vdcsstnDp1qtztREZG0qNHD1588UWaNWtWtvc0Ojqae+65h1deeYWLFy/idrs5efIku3bt8ih/y5YtSUhI4C9/+Qtw6b0NDQ0lNDSUI0eO8P77798wZ1JSEitWrODrr7/GMAxsNhsbN24sOz9VRETEEyqfUqM9/vjjlJSUcPfdd5OcnFy2l/F6vvnmG5KSkujSpQvjx49nypQp3HbbbYSFhfHXv/6VtLQ0EhIS6N27N6+99lqFL5Rp1KgRixYtYtmyZfTs2ZO//OUvLFq0iMjISNxuN2+//TYJCQn06NGD3bt3M3369Ouu55lnnmHSpEnEx8eXe9X95YujLh9yv+zVV1/F6XQycOBAunfvzoQJEzhz5ozHY/jtb3/Lhx9+SF5eHhMnTiQ1NZWuXbsyderUay5+ujpnx44dmTVrFjNnzqR79+48+OCDfPTRRx5vW0REBMBkXO/4pYiIiIhINdCeTxERERHxGpVPEREREfEalU8RERER8RqVTxERERHxGpVPEREREfEalU8RERER8RqVTxERERHxGpVPEREREfEalU8RERER8Zr/BxVTMPy0DtvQAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "The log loss and auc was found to be 0.133 and 0.845 for LightGBM Classifer(top 250) on Normal dataset.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Ks4GX3k7hv_",
        "colab_type": "text"
      },
      "source": [
        "# Combined"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bfDFKhQ9DajM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 518
        },
        "outputId": "f8f6b1fc-909a-41fc-ce9b-424079cf893e"
      },
      "source": [
        "plot_auc(y_val,y_preds,labels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAH1CAYAAAAuzC8kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3gVVfrA8e/MrblppBFAOtKrUkJbwICAVCmiuKDYENSNuro0WVcpgoIPgmIB2cUCiBKU7iLCz0ZRECnSpBMCKaTem9w2M78/LlyJIRDYCArv53kw986cOefMnZhk3jnnPYphGAZCCCGEEEIIIYQQV5F6rTsghBBCCCGEEEKIG48EJIQQQgghhBBCCHHVSUBCCCGEEEIIIYQQV50EJIQQQgghhBBCCHHVSUBCCCGEEEIIIYQQV50EJIQQQgghhBBCCHHVSUBCCCH+pF5//XWeffbZEvf37NmTLVu2lHm7W7ZsoUOHDmVerxBCCCGEuLGYr3UHhBDierNixQr+85//cOTIEUJDQ6lXrx4jRoygRYsWV7Ufq1atuqrtnVO3bl1CQkJQFIWwsDB69OjBqFGjMJlMwTIbNmxg9uzZHDx4EJvNRvv27fnHP/5BhQoVgmXS09N57bXX+Prrr3G5XMTHx9OjRw8efvhhHA7HtTg1IYQQQghRhmSEhBBClKH//Oc/vPTSS4wYMYLvvvuODRs2cO+99/Lll19e665dVcuWLWP79u18+OGHrF69muTk5OC+zz//nGeeeYb777+fzZs3s3LlSqxWK/feey+5ubkA5OTkcM899+DxePjoo4/Yvn07//nPf8jLy+P48ePX6rSEEEIIIUQZkoCEEEKUkfz8fGbNmsXzzz9P165dcTgcWCwWEhMTGT16NABer5fJkyfTvn172rdvz+TJk/F6vcCvUyHmzp1LmzZtaN++PevWreOrr76iW7dutGrVirfffrtIm16vl6eeeopbbrmFfv36sW/fvuC+xMRENm7cCASmdzz55JOMGjWKW265hZ49e7Jr165g2bS0NP72t7/RunVrEhMTef/994P73G43Y8aMoWXLlvTo0aPIcZdSrVo1br31Vvbu3QuAYRi8/PLLjBw5kt69e2O324mLi2Py5Mk4HA7mz58PBAI7oaGhTJs2jcqVKwNQsWJFxo8fT7169UrdvhBCCCGE+OOSgIQQQpSR7du34/F4uP3220ss89Zbb7Fjxw6WLVvG8uXL2bVrF2+++WZwf2ZmJh6Ph6+//pqkpCTGjx/P8uXLSU5OZsGCBbz55pucOHEiWP7LL7+ke/fufP/99/Tq1YvHHnsMn893wbbXr19Pz5492bp1K4mJiUycOBEAXdcZOXIkdevW5euvv+a9997jvffe45tvvgHgjTfe4Pjx43zxxRfMmzePzz77rNSfyaFDh9i2bRvVqlUD4PDhw6SmptK9e/ci5VRVpWvXrsEAyqZNm7j99ttRVfk1JYQQQghxvZK/9IQQoozk5OQQFRWF2Vxyep4VK1bw+OOPExMTQ3R0NI8//jjLly8P7jebzYwcORKLxUKPHj3Izs7mvvvuIywsjNq1a3PzzTezf//+YPmGDRvSvXt3LBYLDzzwAF6vlx07dlyw7ebNm9OxY0dMJhN9+/YNjqbYtWsXWVlZPPHEE1itVqpUqcKgQYNYvXo1AGvWrGHEiBGUK1eOihUrMnTo0Et+Fv369aNZs2b06NGDVq1ace+99wKQnZ0NQPny5YsdExcXF9yfk5NDXFzcJdsRQgghhBB/XpLUUgghyki5cuXIzs7G7/eXGJRIT0+nUqVKwfeVKlUiPT29SB3nkj/a7XYAYmJigvttNhsulyv4/vwkkKqqEh8fX6S+88XGxgZf2+12PB4Pfr+fkydPkp6eXiTppqZpwffp6elUrFixSJ8v5dNPP6Vq1aqsWbOGV199lYKCAqxWK1FRUcE6q1SpUuSYjIyM4P5y5cqRkZFxyXaEEEIIIcSfl4yQEEKIMnLLLbdgtVpZt25diWXKly9Pampq8P2pU6cuOFqgtE6fPh18res6aWlpl11fxYoVqVy5Mlu3bg3+2759O3PnzgUCIxdOnTpVpM+loSgKPXr0oFmzZsyePRuAmjVrUqFCBT7//PMiZXVdZ+3atbRu3RqANm3a8MUXX6Dr+mWdixBCCCGE+POQgIQQQpSR8PBwkpKSmDBhAuvWraOwsBCfz8dXX33FK6+8AkDPnj156623yMrKIisri9mzZ9O7d+8rbvPnn39m7dq1+P1+3nvvPaxWK02bNr2sOpo0aUJoaChz5szB7XajaRoHDhxg586dANxxxx3MmTOH3NxcTp8+zQcffHBZ9Q8fPpxPPvmEjIwMFEVh9OjRvPXWW6xYsQKPx0NGRgbPPfccTqeTYcOGAfDAAw/gcrkYPXo0J0+eBAKJN6dMmVIkcacQQgghhPjzkoCEEEKUoQcffJAxY8bw5ptv0qZNGzp16sSCBQvo0qULAI899hiNGjWiT58+9OnTh4YNG/LYY49dcXudO3dm9erVtGzZkmXLlvH6669jsVguqw6TycTbb7/Nvn376Ny5M61bt2b8+PE4nU4AnnjiCSpVqkTnzp158MEH6du372XVX7duXVq0aMG8efMA6NGjB6+88grz588nISGBnj174vF4WLRoUZEpG4sWLcJsNjNo0CBuueUW7r//fsLDw4MJMoUQQgghxJ+bYhiGca07IYQQQgghhBBCiBuLjJAQQgghhBBCCCHEVScBCSGEEEIIIYQQQlx1EpAQQgghhBBCCCHEVWe+1h0oK7qu43K5sFgsKIpyrbsjhBBCCCGuU4Zh4PP5CA0NRVXl+Z4QQlyp6yYg4XK5OHDgwLXuhhBCCCGEuEHUqVOH8PDwa9a+ruukpKTgcrmuWR+EEOJiQkNDqVy5conB2+smIHFumbs6depgtVqvevu7d++mUaNGV71dcfXINb4xyHW+Mch1vv7JNb4xXKvr7PV6OXDgwGUvs1zWMjMzURSFunXrykgNIcQfjq7rnDx5kszMTMqXL3/BMtdNQOLcNA2r1YrNZrsmfbhW7YqrR67xjUGu841BrvP1T67xjeFaXudrPU04JyeH6tWrSzBCCPGHpKoq8fHxHDt2rMSAhPz0EkIIIYQQ4k9I07RrPkpDCCEuxmKx4Pf7S9wvAQkhhBBCCCH+pK71KA0hhLiYS/2MkoCEEEIIIYQQQgghrjoJSAghhBBCCCH+Z4mJiWW26t2XX37Jyy+/fNEyixYtYv78+ZdV75gxY+jQoQN9+/alW7duTJw4EV3X/4ee/r769u2L2+0u83qfeuopWrdujc/nC25LSUkhISGhSDmXy0XdunWD730+HzNnzqRbt2707t2bO++8k6lTpxap51J++ukn+vTpQ7du3XjwwQc5c+bMBcsdOXKEoUOH0rdvX+644w5ef/314L5ly5bRu3dvGjRowIcffljkuBdffJHu3bvTp08f7rnnHnbt2lXqvomr77pJaimEEEIIIYS4PnTu3JnOnTtftMzgwYOvqO7hw4czZMgQnE4n/fr1o3nz5vTo0eOK6votv9+P2Vx2t1jLli0rs7rOycnJYePGjdSoUYP169fTrVu3Uh87duxYPB4PycnJhIWF4ff7SU5Oxuv1liqfia7r/OMf/2DKlCm0aNGCN998k+nTpzNlypRiZadNm0a3bt0YMmQILpeLXr160bFjR5o0aUL9+vWZMWMGc+bMKXZchw4dGDduHBaLhQ0bNvD000+zbt26Up+juLokICGEEEIIIcR1YN2PWazdmvW71N21RTRdbo2+omM/++wz5s2bB0DVqlWZMGECMTExeL1eJk6cyPfff090dDT169cnMzOTWbNmsXTpUv7v//6PWbNmcfjwYcaOHUthYSG6rtOvXz8eeughXn/9dQoKChg9ejQA77zzDitXrkRRFBwOBwsXLrzoCiRhYWE0bNiQ1NRUAJxOJ1OmTGH//v14PB4SEhIYO3YsJpOJgwcPBvtQr149jh8/zsiRI7ntttsYOnQo9erVY8eOHURGRjJ37lzmzJnD2rVr0TSN+Ph4Jk6cSFxcHOvWrWPmzJmoqoqmafzzn/8kISGBN954g5UrV2Kz2VAUhffff5+IiAjq1q3Ljz/+SGhoKDt37mTy5MkUFBTgcDh47rnnaNKkCSkpKQwYMIB77rmHr776isLCQiZPnkyLFi0ueN4rVqygY8eOtG/fnuTk5FIHJI4ePcq6dev46quvCAsLA8BsNnP33XeX+nth9+7d2Gy2YN/uueceOnfufMGAhKIo5OfnA+B2u1EUhejowPdgnTp1AC54fW+77bbg62bNmnH69Gl0XZfVaP6gJCAhhBBCCCGE+F0cOHCA6dOns3TpUsqXL89rr73GxIkTee2111i8eDGpqamsWrUKTdMYOnQoFSpUKFbHwoULSUxM5NFHHwUgNze3WJlPP/2U9evXs2jRIsLCwsjOzr7kDeiZM2fYv38/f/vb3wCYMmUKLVu2ZPLkyei6zrPPPktycjKDBg1i1KhR3H///fTt25ddu3YxaNCgInWdOHGChQsXYjabWbZsGSdOnODjjz9GVVUWLlzI1KlTefXVV5k1axYTJkzglltuQdM0CgsLycnJYf78+Xz77bfY7XacTid2u71I/V6vl6SkJKZMmUKbNm3YuHEjSUlJrF27FgiMemjWrBlPP/00y5cvZ/r06Xz00UcXPO/k5GRGjx5Ns2bNmDx5MmlpacTHx1/0swLYs2cP1apVIzIy8oL7Dx48yDPPPHPBfW3btmX06NGcOnWKSpUqBbdHR0ej6zo5OTmUK1euyDHjxo1jxIgRLFy4kLy8PEaNGkXlypUv2c/zLViwgE6dOkkw4g9MAhJCCCGEEEJcB7rceuWjGH4vW7ZsoWPHjpQvXx4IPBHv27dvcF/fvn0xm82YzWZ69uzJtm3bitXRsmVLpk2bRmFhIQkJCbRu3bpYmQ0bNjB48ODgk/uoqKgS+zRnzhwWL17MkSNHGDx4MLVq1QJg/fr17Ny5k//85z9A4Kl8fHw8TqeTAwcO0Lt3bwAaN25cJK8CQO/evYNTNdavX8/u3bvp168fEFie9Vy/WrduzZQpU+jatSsdOnSgTp06aJpG1apVGTVqFO3bt6dTp07B8uccOXIEi8VCmzZtgMANvsVi4ciRI4SGhuJwOIIjA5o1a1Zi/o09e/aQl5dH69atURSFrl278tlnn/Hoo4/+zyu23HzzzWU6xWTx4sX07duXhx9+mPT0dIYOHUqjRo1o2rRpqY5ftWoVK1asYMGCBWXWJ1H2JCAhhBBCCCGE+MPq1q0bzZo147vvvmPu3LkkJyczffr0K67vXA6Jw4cPM3jwYNq3b0/Hjh0xDIM333yTKlWqFCnvdDqBiy9f6HA4gq8Nw2DkyJEMHDiwWLlx48axf/9+Nm/ezJNPPskDDzzAoEGD+Pjjj/nxxx/ZvHkz/fv3591336VevXqlPier1Rp8raoqfr//guWSk5PJy8sL5ufwer2Ehoby6KOPEhUVhdPpRNM0TCYTANnZ2cTExADQoEEDjh07Rm5u7gVHSZRmhETFihWDU2QAsrKyUFW12OgIgA8++CCY+6F8+fK0bt2aH374oVQBiS+++IIZM2Ywf/58YmNjL1leXDsydkUIIYQQQgjxu0hISOCrr74iIyMDgI8//pi2bdsC0KpVK1asWIHf78fj8bBmzZoL1nHs2DHi4uLo378/jz/++AVXTbjttttYtGhRMHiQnZ19yb7VrFmTpKQkZsyYgWEYJCYmMmfOHDRNAwI3yydOnCAsLIzatWuzcuVKAH7++eeLriaSmJjIwoULg1NLvF4v+/btA+Dw4cPUrVuX+++/nz59+rBr1y6cTidZWVm0atWKpKQk6tSpwy+//FKkzho1auDz+di8eTMAmzZtwu/3U6NGjUue5zler5eVK1eSnJzM+vXrWb9+Pd9++y0AW7duxeFw0LhxY5YsWRI8ZtGiRbRr1w6A6tWrk5iYyPPPPx/8nDVN45NPPsHlcgVHSFzo37k8H40aNcLtdrN161YAPvroI7p3737B/lauXJlvvvkGCASFtm3bRu3atS95nhs2bGDKlCnMmzfvsqd4iKtPRkgIIYQQQgghysQDDzwQfLoOgQSKzz77LA8++CAAVapUYcKECUBg+sa+ffvo2bMnUVFR1KxZ84J1rlmzhhUrVmCxWFAUhXHjxhUrc+edd5KWlsbdd9+N2WzG4XCwYMGCS+YOuPvuu3n//ff58ssvGTduHNOmTaNv374oioLFYmHcuHFUqVKFl19+mXHjxjFnzhzq1KlDnTp1CA8Pv2Cdd955Jzk5OQwZMgQIjJgYPHgw9erV49VXX+XYsWOYTCYiIiKYPHkyTqeTv/3tb7jdbgzDoEGDBnTt2rVInVarlVmzZhVJajlz5swiIyMuZd26dVStWpVq1aoV2d67d2+Sk5Np0aIF06ZNY9KkSSxcuBDDMKhduzbjx48Plp06dSqzZ89mwIABWCwWdF2nY8eOpe6Hqqq88sor/Otf/8Lj8XDTTTcxbdq04P6+ffsyZ84c4uPjmTJlCpMmTeLf//43fr+fHj160LFjRwBWrlzJK6+8Ql5eHl9++SVz5szh3//+NzfffDNjx47FYrGQlJQUrHf+/PkXncYjrh3FMAzjWneiLHg8Hnbv3k2jRo2w2WxXvf1t27bRvHnzq96uuHrkGt8Y5DrfGOQ6X//kGt8YrtV1vtZ/d56zd+9e6tevf83aLwtOp5OwsDC8Xi8jR46ke/fu3HXXXde6W8W4XC4cDgeKonDw4EGGDh3K559/XmKCRyHEry72s0pGSAghhBBCCCGuiQceeACv14vH46Ft27bBRJB/NNu3b+eVV17h3LPciRMnSjBCiDIgAQkhhBBCCCHENfHJJ59c6y6USvv27Wnfvv217oYQ152rktTy5ZdfJjExkbp165aYAEbTNF588UW6dOnC7bff/qf54SSEEEIIIYQQQojLd1UCEp07d2bBggXcdNNNJZZZsWIFx48fZ+3atSxevJjXX3+dlJSUq9E9IYQQQgghhBBCXGVXZcpGixYtLllm9erV3HXXXaiqSnR0NF26dOHzzz/n4Ycfvgo9FEIIIYT4YzB0g/3fp5B5Mq+EAgbaqQzUHCcWTeGknkmu4bponbqhEZj5bqDrfgzDW9bdLhsGlDbbekSkm1C7X5KXCiHEn9gfJofEqVOnqFSpUvB9xYoVOX369GXXs3v37rLs1mXZtm3bNWtbXB1yjW8Mcp1vDHKdr3+lvsZ+A9sxH+Yz2u/bod8wDCjI9mNoBgYGut+J5ndhaH4MAxTAg599ajo+An1TFAWzYkVVAssqnjHycOHGjAnlchq/rMJXUSn6ZTbpNKnlonZlNzm5Vvl/WQgh/sT+MAGJsiLLforfi1zjG4Nc5xuDXOfrX2mvsT/XzalFP+M95cUSG4JiLjqb9bTnDBtzdqEZGoYBhmFg6Eapn+IHCpZUWg9EJYoWBlUJvveg4Vb8RGALFtGNQjTdh1/xo+MlTD+FmfwS7+UVFELMDsLN4YRHxVGl3q006dQHq91xye5rmoHHr5OV7yMzx0dmro8Cj4bZrKJpBn7NQNPPftUM/HrgvaaB/9z735bTzyurgf+84w+fKqRmpRAGt48gMmc/4Wd+JiR9N+bsYwC4YyLJqx6FZtKIimtFYWT9a7rspygqMTERq9WK1WqlsLCQm2++mUceeYRbb731iutctGgRHo+HYcOGlVhm5syZ1K5dmx49elxxO+fs37+fUaNGAZCbm4vT6QxOOx80aBB//etf/+c2fmvnzp3MmDGDEydOEBkZSWhoKH/7299o2bIliYmJvP3229SpU6fM2vvtZzp27Fh27txJ7dq16d27N1u3bmX06NFl0tZXX33F8OHDeeONN7j99tuD24cOHcqDDz7IbbfdFtyWlJREp06d6N+/PwDffPMNs2fPJisri7CwMOLi4vj73/9O3bp1S9V2YWEhY8eO5eeff8ZkMjF69Ogi7Z3z/vvvk5ycHHx/4sQJ7rrrLsaOHXvRfZqmMXXqVDZt2oTP5yMxMZFRo0ahKH/UiO8f0x8mIFGxYkVSU1Np0qQJUHzEhBBCCCGuX4ZuYHivfISCz+dD0wPHe3QPBQVOzpzJoDDfhduZx4EN/8V7PJ9q1maYlF8fXJgUC7qhsS3vM9LTDga3u0Mq4bFVwFAtKLoPs9dZtMFS/L2pXCRqYaADoCoWVDXw2mI2YzVZUZVAUMQwKXjKO6hftx5DOg8tzcdwUbpuUODRKfBopOZoFHpcFHi04LZCj05mrpe0bB/ZTj+5Tj95BX7yC7WicZOLUBQwqQpmk4JZVTCZFMymX7eZTErgdXBf4HWIWaey/xh31TxEI9NBlM8PgK6ByYy5ckOUxm3Jtp0hP3cv1tA4qtS/D0e5mqTK6Ig/nFmzZgVvnteuXcvw4cOZN28eTZs2vaL6Bg8efMkyTz755BXVfSF169Zl2bJlACxdupT/+7//Y9asWcXK+f1+zOb//VZq//79PProo7zyyiv85S9/AeD48ePs3bv3f667JOd/ppmZmfz3v/9l69atqGrgZ0/nzp0vu05N0zCZTMW2Jycn07p1a5KTk4sEJC7l22+/5bnnnmP27Nk0btwYgL1795KRkVHqgMS8efMICwvjiy++4OjRo/z1r39l7dq1hIaGFil33333cd999wGB3yUdOnSgV69el9y3ZMkSDh8+zKeffoqiKIwYMYLVq1fTs2fPUp+n+AMFJLp3784nn3xC165dycnJYd26dSxYsOBad0sIIYS44RmGgS+zAEMr9biAUtHdfvy5Hvw5blK++wW3231F9eQZBazXdvDbcQtr1vy3aEELfG9svfCAhbAYIKbIJqtXw+z3EeL2YwszYQ7VUOx+zPbzypwpwHEy94L9MgB3fDjuOAdamP03e1VUqwVHQkOs4eGBulQrbWLaU95eHgiMMDiR4QYDjpwqvMSnEODXDfYdL2D/iQLyCvzkFWg4z311ly6wYLeqRDhMRIaaqRTjICrcQmyEhbiowNfYSCuhdlMg0HAu8KD+GmwoDcMw0DKO4j/6E74jP+E/8TN4CwEFU4VamBP6YaneDNNN9XBm7yF1/0doeS5ia/QgtsYdqKqlVO3caDy7vsS744vfpW5r09uxNb68G9WuXbuyc+dO5s2bx6xZs/B6vcyYMYMffvgBr9dL3bp1eeGFFwgNDSU/P5+XXnqJ3bt3oygKLVq04Pnnn+f111+noKCA0aNH8+OPPzJx4kR0Xcfv9zNy5Eh69erFmDFjaNSoEUOGDMHlcjFp0iR27doFQN++fXnkkUeAwFP5Ro0a8dNPP5Gens4dd9zBs88+W6pzSUlJYcCAAfTv35/NmzczaNAgOnfuzKRJk0hNTcXj8dCzZ09GjBgBwOHDh3nppZfIzs7G5/Nx//33M2DAgGL1zp07l4EDBwaDEQBVq1alatWqxcr++9//ZtWqVWiahs1m44UXXqB+/foUFhYyevRoDh48iNlspkaNGsycOZPDhw8zduxYCgsL0XWdfv368dBDDwU/08cff5z77rsPt9tNv3796NevHxEREUWCMJ9++ikLFy5E0zTCwsJ44YUXqFmzJkuXLmX58uWEhoZy7Ngxpk2bRv369Yv0Nzs7m82bNwdv0jMyMoiLiyvV5z179mwee+yxYDACKFb/paxZs4apU6cCUL16dRo1asTXX3/NHXfcUeIxGzZsIC4urki7Je3bt28fbdq0wWIJ/Dxq164dK1askIDEZboqAYlJkyaxdu1aMjMzeeCBByhXrhyrVq3ikUceISkpicaNG9O3b1927NhB165dAXj88cepUqXK1eieEEIIIS5A92gUHsvBtSeT/O2Xn9epVG0YOjv0I/yoH7x04Yuw6AYVfMXvtk0mByabHUwWNLsVv6Kh/Oam2aJaCTGFnH1n4D11CtXlprzHjM3wEec/idnth8wLtoz1pqaUO+9mIrgnLo6IK5xOkJHrZdy8w6RkeK7o+KgwM7GRFsJCTFSIchDuMBHuMBMeYsJhV3HYTIRYVRx2EyG2wHuHTcViVjGpYDYpZT7sWMtJCwQgjv6E/+hOjIIcANToStga3Ya5ejPM1ZqghgQCND5PLif3zic/Ywf28KpUu+VJ7OGVy7RP4vfXtGlT1q9fD8C7775LeHg4S5YsAWDatGnMmTOHp59+mpdeegmHw8GyZctQVZWsrKxidc2dO5eHHnqIXr16YRgG+fn5xcq8+eab6LrOihUrcLlc3H333dSpU4eOHTsCgVHYCxYswOVy0aVLFwYOHEj16tVLdS45OTk0btw4OJ3hgQce4LHHHqNly5Z4vV6GDRtG48aNSUhI4Nlnn2XatGnUqlULp9PJgAEDaNasGbVq1SpS5549e+jevXup2r/zzjt58MEHAdi4cSP/+te/+Pjjj/n2229xuVysXr0aCEw1AVi4cCGJiYk8+uijRbafExYWxpw5cxgwYECRUSHnbN26lTVr1rBgwQKsVitfffUV48aN46OPPgJgx44dLFu27ILBE4Dly5fTqVMnYmNjuf322/n0008ZPnx4qc51z549PP/88yXuT0pK4tixYxfct3jxYux2O6mpqUVWeSxNjsLk5OTglJFL7WvYsCFLly7l3nvvBWDdunXk5ZWQjFiU6KoEJMaPH8/48eOLbZ87d27wtclk4sUXX7wa3RFCCCFECXS3H9f+M/hz3OR+n4rmDKzGYK8eSWTCTbj8LjI8Gbj8RW8EdM2g4JgOF5h14clJxe8pQPe68eWm4XNlA5CrQrrdjscSitmbi92VSkn5FlQ1HtVUvfgOwyDEcBHpKyTCX1BkV5yuE21cQSBFVYkfPBhzVFSpiltiYwkp5Q3NOT6/TuF5UyUKirzWWLg+jex8PyN6VyI24vJGA9SoEILNqhJzmceVNb0gF/+xnYEREMd2oGefAkAJjcJSoxnm6s2wVG+KGlm+yHGGYZCTupG0X5Zg6H7K39yfmKqdUdTiw8FFUbbGnS97FMPvzThvWM769etxOp3897+B0Uter5d69eoBgafPS6V6M3oAACAASURBVJcuDU4biI6OLlZXQkICb731FsePH6ddu3YXnAayadMmxo0bh6IohIWF0bNnTzZt2hQMSHTv3h1VVQkPD6dWrVocP3681AEJm80WfLpeUFDA999/XyRw4nK5OHToEPHx8Rw6dIi///3vwX0+n4/Dhw8XC0hcjt27d/POO++Qm5uLoigcPXoUgHr16nHo0CFefPFFWrVqRadOnQBo2bIl06ZNo7CwkISEBFq3bn1Z7a1fv559+/Zx1113AYFref4N96233lpiMAICwY0xY8YA0K9fP5577rlLBiRKGwi90DSa/1V6ejqbN29mypQppdrXv39/Tpw4weDBgwkLC6NJkyZs3ry5zPt1vfvDTNkQQgghxNXhz/Xg3JOBPzfw9N3QNDwnUyjMyUdxRgb/IMzxZ3DYvQuXlo/npwK07X50Q8OvmCiwODDO+7vR0C1gFL9hNAwPhpF99p0CSiiqNQ5FCcdlDzyts2sa7bJ8KPw6lNdvN2Pp0haTYsZiCyEkvPjNiaEbpOb4yckPQ0Hh/HBE1pksCmKiufDzs4vz2cLweiMhrZQHpAE/pwbf6oaB2xsIMhSezdFQ+JtcDT7/xedOhNpV/jmkGq3qRV7BGVwbhs+N/8TP+I7uwH/kJ7S0w4AB1hAs1Rpja9EnEICIrVriTYe3IINT+xbgytqHo1xtKjYYgs0Rf3VPRJSpXbt2Ubt2bSBwQ/uvf/2LNm3aXFFdw4YNIzExkY0bNzJx4kTatWvH008/fVl1nJ/83mQyoWmlz10TEhIS/N7VdR1FUViyZElwyP45v/zyC1FRUcFRBxfToEEDdu7cSZcuXS5azuv18uSTT/Lhhx/SsGFD0tLS6NChAwBVqlRh5cqVbN68ma+//poZM2awYsUKunXrRrNmzfjuu++YO3cuycnJTJ8+vdTnaxgGAwYMKDFHx29zMZxv9+7dHDx4kOeeey64LT09PZhwODo6mpycnCLHZGdnBwNR5z6XkqZplGaERKVKlTh58mSwzlOnTpGQkFBinz/77DM6dux4wWDYhfapqsrTTz8d/B6cO3fu/xRwulFJQEIIIYT4Eyo8coTCAwdwHz+O7ik6rF8zNAr8rsBSkoaOboDiD0PVIjH5IzBrMSioaIYGKBiGDugYhOA38tmetZxs70l8evGcDl5bOfKjAk800Yv/IV/8NtOOQkVU1F8LqAaK2YXZMNOlS0caNayHyWRC18DwmcjI1cnxWTidq3Ey00NWth+yf63RVahxKstLRo4X3QCbRUH9zQ2upsdjyrrSJ+p+4MwVHgsogTwM56ZBhNhUypezBqdGhNgCUyUM3SA20orDrhJiPW8qhU2lXKgZq0W9dFvXkKFraKkHAgGIoz/hP7kXND+oZsyV62Pv8NdAHohKdS45usEwdLJObCD94DIURaVCvXuJuqk9ivLH/gzExa1bt45FixYxb948ILAKx/z587nllluw2+04nU7S0tKoVasWt912G/PmzWP8+PEoikJWVlaxG8MjR45Qo0YNqlatisPh4LPPPivWZps2bUhOTqZ58+bBaQznVs0oS2FhYTRv3pw5c+bw+OOPA4Eb3nM5HOx2O5999hl33nknQHDkRFhYWJF6Hn74YYYNG0br1q1p27YtEFjJYc+ePXTr1i1Yzuv14vf7qVixIhCYjnHO6dOniYyMpEuXLrRr146//OUv5OTk4Ha7qVKlCv3796datWqMGzfuss4xMTGR0aNHc/fdd1OhQgU0TWPv3r00atTokscmJyfz8MMPFwkYvfPOO8Fr065dOz777DN69OiBzWZj3759HDp0KLjAwciRI/nnP/9Jo0aNaNiwIRDI2ZCZmUn79u1LNUKie/fuLF68mMaNG3P06FF27drFq6++etE+l/QZXWifx+PB6/USHh5OamoqixYt4vXXX79kv0RREpAQQgghfkeGYWB4vegeD7rbXfyr243mcqE5ncGvusuFVlAAhhEc7mwYOj7dj1f34XUXUpiVgQ8dt8POCUckZ0w2NCOweoN+9hhFUbGbIjErNhTFA6SjGan4dS8evQDd0AHt7IoPCorJgTkshKgGTTApzfjtLaRhGJxJTcVms/HXv/6Vm2+++Yo/F2ehn33HCzhwooBNX3g4lVVAZq6X7Hw/+nmDBxw2lZhIS5GAg92qUr+qg8Rm5YiNtNKtZXSxZIqBp3DNrrh/ojjDMNAzj58NQGzHd3w3eALjUkzxtQIjIGo0C6yKYf1tEs+SeZyppO79gMLcI4TFNKJi/Xux2Is/oRR/DklJScFlP2vVqsWcOXOCUyvOLf84cOBAFCWQp+SJJ56gVq1ajB07lpdeeolevXphMplo1apVsSnfH3zwAVu2bMFisWC1Wi84Jfyxxx5j4sSJ9O7dG4A+ffoERxKUtenTpzNlypRgW6GhoUyePJm4uDjefvttXnrpJebNm4eu68TExPDaa68Vq6NevXq8/fbbzJgxg+eff56QkBCioqJISkoqUi4sLIykpCQGDhxIuXLligQr9u/fH7zR1nWd4cOHEx8fz9tvv82KFSuwWCwoinLZAYmWLVvy1FNPMXLkSDRNw+fz0b1790sGJDweD6tWrWLRokVFtvfq1Ys+ffowfvx4Bg4cSGpqKgMHDkRVVex2OzNmzCDq7FS5Dh06MGHCBCZMmEBOTg5ms5nKlSvzzDPPlLr/Dz30EGPGjOH2229HVVUmTJgQDAjNnDmT8uXLB1cc2bZtGwUFBbRv375YPSXty8/PZ+jQocFpRs8++2wweCJKTzGM0i7k9Md2bj3oRo0aFRmKdbXImvbXP7nGNwa5zjeGkq7zb0canM/QDQy/Xmy75nKRvWEDrl270PLyzmZAOPtfv4bh9aIZ5z3ltcagxrRGUczBksbZ9SEMwFCUs4ef+/Uc2JdODieMLDLIw403WJ0HPwCxhGOj6LBhVQmEFAq0fLy6B5/uCQQhFFBNKmazitlmJSQsArsjDGtIycNvf6XQvGXCZQUj/BqcyvJwLM3NzkNODqQUkJLpQTv7cZpNCrERFspHWagUY6NKnI2bKzmoGm8jMtR8RckV5f/lsqHnZQSSUB75Cd/RHRhn83+oURUxV2saCEBUa4LquPypJYbuJ/Pof8k8sgbVbKNCnUFEVGh1Wdf7Wl3na/135zl79+697JUHhBDiarvYzyoZISGEEOK6ZxgGnpRT+HMCUxBMJ7JwWg8XKZO1YQMFZ5eJuxBz+a4o5pJu2OPBEo8p5tf2cinAwCDTyOOEnkEBbpyGmwLcGGwuKXdjydsBMyqRSgixagSKop69cVOpZK1AeXscJsWESTFhVn/99e41qxTGOYLv7aEWajevRFSFwEoGh04WsOSbDFwGeHz6ef8MvGdfu306Xp+Bx6djGPDRPjewu+SOXoRJhUoxNjo1LUfD6qE0qBZKfJQVm0Ut81UdxOXTC/PxH9t5djWMHehZJwFQHJGYqzfFUj2QjNJU7n/L61CYd4zUPe/jcZ4kIr4FFeoOwmyNKItTEEII8SciAQkhhBDXLW9aGvnbtpGz6RiGUQvFFHiSGY6VtO0nflP6ZiyVLv7U32/yk+/IxdA1DEND17VAkkeTGWeuQo7XhVMtJEPNxqUWzb8QgZkQv59I3aBQz0Gzqpit4Vjs4VjtkdhCIrCY7dgdNsIr2nCYHUWOL1euHFWqVAkOZy0NTTfw+3XyCjXO5PrIyveRluvj55/yyXZmk53vZ39KAW6vTvlICzaris2iYrWohIWo2C0KVktgm82qYjMH3ptNlxc4UBSoEGWlSnk7MRFmHDaTBB/+IAyfB3/KnrNLce5AO3UQMMBix1y1EbZb7wgEIOKqlUk+B13zknF4BWeOrcNsi6BK05GExxVfKUEIIcSNQQISQggh/tQMXSd30yZ093kBAMOg8GgK+T8eRg2pghpaD3O4jqO2BUVVSEtLIz7+1ye8bq2QA3n72elNp1DRg3VYc7zYMlQUvxcDDacvDd3wF21fNeGzRuKzRII1MEVC0byYvYVnl7EEk1ZAdGQEcVVu5pYuA6hS9xZU05UvYWgYBroeWM1B1w3O5PtJzfRwMtPDqSwvp7M8pOf4yHb6cRZq+LXiwy5MKoSFmIgJt/BYn5toXDPsAi2J642ha2inDwaW4jy6A3/KHtB8oJowV6qH/S/3YqneNJCI0lS2y4a6sg9was8HeAszKHdTe+JrD8BkDinTNoQQQvy5SEBCCCHEn1rGkiWcWb26yDbFEo25wh1Y4mqACUJqRRPfvx6Hf04nIyWPNMNDnlEOzdBId6exN3c3/uwzePP9v6ldBQtgsQbehlS/YB9CQ0OoFB9BtaqVuLlSTUIdv95kqaqJqPgqWEMCIx5ynD4OnXZT6AksDen2ahR6dQrcGm6vTqFXP7tdx3Pu9XlTJrw+Ha/fwOsPbPP69WAuhnPCQkxEh5upUcFOdLiZqHALMeFmYiOtxESYiY6wUC7UjNmsFksGKa4vhmGgn0kJjoDwH9uJ4XEBYCpfA1vzXliqN8VctRGK9fcJDmj+QtJ/+ZTsk19jCYml2q1PERpd73dpSwghxJ+LBCSEEEL8IXkKfGScyGX/9yfx+35dXtLw+9FTUilv2FEAwxuCUvEuTBERqAZY/ToWDfwKbLIeJ8vIQz+q45++Dl+ehq5nAl44EKhPV1T8JgXDZMHkLyAiazfnFq8MCW1K3YS7admjDmbzhUc0mEwmQkJKvpEzDAO/ZpCV5+VASiFTPzqGx3fpfNImFSxmFatZwWpWsZ6dPhHmMGOzKMFpFHaLSkyEhcqxNirFWKkUayM8xIzpMqdViOuHnp95diWMHfiO/IThDCxhqkbGY6nfHku1ppirN0UNLfe79yU/cxen9i7E78khumoXytfqg2qy/u7tCiGE+HOQgIQQQohrzp+fz+n33iMtzcMvrkrk+0MoZ7ZTyWYhXNHw4sGDF/fZf7rqZy9eXLqHQpMXXdHBGcgHqWGgGwZeDDxeFyoWFAJz3w2TC0XxYqhKMIeBip8wRaN8XARtE/tTt/EtqOqVTafIzvexcU8uqZke0nN9nMn1kZ3vJ9flp9AbGMZgt6o80qMCDrsJu1XFblUJsarYbSoOW2Cbw2rCYlZQFFCVs19lJIMoge524j+262wiyp/Qz6QAoIREnE1E2TSQByKq4lXrk9/rJO3Ax+Se/h5baCUqNxmOI7LGVWtfCCHEn4MEJIQQQvwuDH9gRYbge0PH8AemRGg+DS3fiy/Pgzstm9NrvyPXXxHDFEZNO4SaFPYbJ/mvfhg/WvHKdbBZbZgVDX9BLqCjG9rZBTJBOTvCwWZohLuPY1YCIyv8bje1mnegcut+tGje4orPTdcN3D6dtCwvp7I8pGR6+Pmoi+2/OPGdzddgt6qUCzUTFW6m1k0hxEVYiI+ycEvtcKrFy7x5ceUMvxd/yt5gAEI7dRAMHSw2zFUaYmvaNRCAiK9RJokoL6tvhkFe2lZO71+M5i8krmYvYqt3R1HlT04hhBDFyW8HIYQQF+V2u8nPzw++1wp86J7iQQKtwIXu9gBQeDgf1768i9Z72shii7Y/EHBQCORqCFb268u2rZpz5tgezhzbi8+ZjaJ7UXUfytngQ7XGrTikpODW3MTZy1PeFk+k9fyh6LUBUBSFag1bEl+tLr8cTSmxX+dWpshy+snM8ZGR5yUjx0dmro+sPB/pOYHVKnKcfvTzAi5Wi0LLuuF0bRFD7ZvshDvMmE2KrCYh/meGrqGlHQ4uxek/8TP4vaComG6qi73tIMzVm2G+qR6KuWwTUV4OnzubU/sW4czciT2iOtUaDMUedtM164+4+hITE7FarVitVnw+Hw8++CB33XVXmbfx9ttvU6dOnTKt95yhQ4eSmppKWFgg0W+NGjV47bXXfpe2zsnLy2Px4sU88sgjJZbJyMhg2rRpbNu2jYiICEwmE4MGDWLQoEGMGTOGRo0aMWTIkDLr065du5g/fz6vvvoqAAsXLuT999/Hbrcze/ZsxowZwwcffFAmbeXm5vKXv/yFQYMGMX78+OD2119/nYKCAkaPHh3c9uGHH7J7926mTp0KwJEjR5g+fTr79u0jMjISq9XKww8/TJcuXUrd/uzZs/n0008B6NevH48//vgFyy1ZsoT33nsPVVUxmUyMGzeOFi1aoOs6gwcPprCwEIC4uDhefPFFKleufNF9IkACEkIIIdAK/XhOBgIIfr+f42knMQydXGceX3z/f/j8vt+l3RhTBOXt0eh2jZC4kMDUBDN4XPm4Xfn4zhxj//KZoEDNJm2pdPMdnMvvABAaGc1/Y3ZxvCCPe6s+Qbu4v5Su4fMCEpqms/uoi7VbsziYWkhegUZ+gb9YokgAh10lOsxC1fJ2Wta1UiHayk0xVirF2qkYY8Vhu/KVM4Q4xzAM9OxU/EfOBiCO7cBwOwFQ46phuyWwFKelaiMUm+MStf3+DMMgJ/Vb0g4kYxga8bUHEl018aqPzhB/DLNmzaJOnTocOHCA/v3706FDhyKrGv0ZjB8/nttuu+2KjvX7/ZjNl3eLlZeXx7vvvltiQKKwsJAhQ4bQv39/pk6diqqq5Ofns2rVqivqY2k0btw4GIwA+OCDD3jllVdo0qRJ8P3lKumzWblyJU2bNmXVqlWMGjUKq7V0eWbS09MZMmQI//jHP5g9ezYQCNx89913pe7TDz/8wOeff87KlSsBuOuuu2jVqhUtW7YsUi47O5uXXnqJtWvXEhsby5dffsnzzz/P6tWrUVWVd999l/DwcADee+89pk6dyhtvvHHRfSJAAhJCCHGDMDQD554M8nJyOXz6GLrx6x13wS9Z6IWB6RTb9UMU4AnuU1FoqzbAqgR+ZSgWA2u0ASbAMPCcPInmcmGrVAlzVBQQGOCQXeBDMQVuSHTd4MhRL9EVwmn0l2rYHBZsNhsVKlQg6/hePn1tNFA00aNqtWIpF0Fsu1ZUbNuGkNhYvEXPiJWnPyer4Ay3le9SJBhhGAYFbp0cl58cZ2A0Q3a+nxynnxyXj6MpKh//eJD8Ao0cV2CfSYWq5e3UqhRCdLiZmHALceUsxEZaiYu0EBtpIcRmwizJIsXvQHdm/ZqI8uhPGHkZACgRcVjqtAnmglDDoq9xT4vyFmSQuvdDCrL344iqS6X6Q7A64q51t25YW85sZFNm6W/GLkeb2HYkxLQtdfk6deoQERERXGZ5xYoVvP/++/h8gQD36NGjadOmDRAY9dC3b182btxIRkYGDz74YPBp/9atW3nxxRcBaNmyJcZ5cwF37tzJ5MmTKSgowOFw8Nxzz9GkSRNSUlIYMGAAgwYN4ptvvsHtdjN9+nQ++ugjduzYgd1u58033yQurvTfq5dqq3///mzevJlBgwbRuXNnJk2aRGpqKh6Ph549ezJixAh0XWfChAls3rwZq9WKw+Hgo48+YsKECeTn59O3b19CQkL46KOPirS9cuVKypUrx6OPPhrcFh4ezj333FOsn5s2beK1117D4/GgaRojRoygZ8+eALzxxhusXLkSm82Goii8//77WCwWRo8ezcGDBzGbzdSoUYOZM2eyZcsWXn75ZZYuXcpTTz3FiRMnGDVqFA0bNuTpp59mwIABbNmyBYAdO3Ywffp0XK7ACj5JSUl06tTpgp/N4MGDi/U5OTmZf/zjH7zzzjt8+eWX3HHHHaW6JgsWLCAhIYE777wzuC0uLq7I+0tZvXo1d955J3a7HYA777yT1atXFwtIGIaBYRi4XC5iY2PJz8+nQoUKwf3nAg4ATqcTVVVLtU9IQEIIIa57ul/Hk5rPmTWHOH3yFMv9my+cl+E89cNjaBhVHgCHyUKkNfCL2rV7J74TR+Dwr38QmqOiiOndi+jOndmz8XOO793P4Z1p+NxFl9B02ExUq1wZT2pmMNzx87ojHN21GX+YiYzbKqBbTRgq+MMt6HYTKAqQCwWfw/EL97VdbAf6V/51SLBhGEz/+Djrf8op8fysZoWIUC+hdpXKsTZ6JcRwe/Mo4srZLvq5CFFWDE8BvuO7gith6JnHAFDsYZirNcXS5i7MNZqhRlX6Q077MQydrOPrST+0DEUxUbH+XylXqf0fsq/i2ti2bRtRUVHUqxdY4rV9+/b06tULRVE4fPgww4YN4+uvvw6Wd7vdLF68mJSUFHr37k2/fv2wWCw8/fTTTJ8+nYSEBFavXs2CBQsA8Hq9JCUlMWXKFNq0acPGjRtJSkpi7dq1AOTk5NC8eXOeeeYZ3n33XYYNG8YHH3zApEmTeOGFF/jwww95+umnL9j3SZMmBadp3HffffTu3fuSbTVu3Dg4teCBBx7gscceo2XLlni9XoYNG0bjxo2Jiopiy5Ytwafqubm5ADz//PMMGDCAZcuWXbA/P//8c3BkwqU0aNCAhQsXYjKZyMzMpH///rRv3x7DMJg/fz7ffvstdrsdp9OJ3W5nw4YNuFwuVp9dPvtcn8732muvkZiYGBz9kpLy6yjDvLw8/vWvfzFnzhzKly9Peno6AwcODI44+O1n81v79u0jJyeH1q1bk5GRQXJycqkDEnv27KFdu3Yl7p8zZ06Jo0j++c9/0qJFC06dOkWrVq2C2ytWrMgPP/xQrHx0dDQTJkygX79+REREoOt6sVEijzzyCHv27CEqKop58+aVet+NTgISQghxnTIMg5MnTrJl0Xqy83IwgHRzHrpqMOL+4YSHhZG5bDl527aCagLdDQQmRNjT0lDYQ6ESyi/mGniUQBJGxVwJe6M2mMLDyc/+mdysnaDoKGtXYKxdTm7Gj2dbV1HUovkTPIWwcwPBxJNg4LebcN0cht66Js+2HEuI6fKSPaqYCLcEnjxk5flYuy2L//spm2PpHprVCqNBtVDKhZmJCgsklywXZqFcqJndu3+iZfP6snKFuGoMvw//yX2/JqJMPRBIRGm2BhJRNr7tbCLKmihXuMrL1eJ2niR1zwe4844SFtuEivUGY7FHXetuCSAhpu1ljWL4PSQlJWEYBsePH2fmzJnB4fcnTpzgmWeeIS0tDbPZTGZmJhkZGcFRCj169ACgcuXKREREcPr0aXw+HyEhISQkJATLPP/880Agd4DFYgmOsmjbti0Wi4UjR44QGhqKw+GgU6dOADRs2JAKFSpQv3794PuNGzeWeA6/nbKxf//+i7Zls9mCN9EFBQV8//33ZGVlBY93uVwcOnSIfv364ff7ee6550hISLjiaSEXk5WVxbhx4zh27Bgmk4nc3FyOHDlC48aNqVq1KqNGjaJ9+/Z06tSJsLAw6tWrx6FDh3jxxRdp1apV8DMrre3bt5OSklJkuomiKBw7doyoqKgin82FLFmyhL59+6IoCl27dmXSpEnBUTUlBThLG/gcPnw4w4cPv6zzKYnT6WTBggUsWbKEmjVrsnr1ap544gmWL18e7M/cuXPRdZ133nmHt956ixdeeCF4/MX23egkICGEENcRv9PL5vfXsT1jL/m+Apx6IQoK4ZhB0YjRoIHJhL74Q3IB/+HDlIuLo+bUqUV+weu6xrr3F7Jn00YMbQsm89kbJD9wCHS9EM2XgaqGoqj2s0cZmCxxVKg+gC739Sb2pogifdMMP1P3TCLVfTK4LcTkoE54HYbXunACqd/SNJ1vd+eS69LOm46RSVq2lyOn3BhA+XIWBnaI477bK2AxX3hYpFmVZTTF78swdLT0I2fzQPwUSETp8wQSUVasjb3NwEAiysr1Ucylmy99rRm6n4wja8g8+jkms4ObGj1MRHxzGRUhijj3FH3NmjWMHTuWW2+9ldjYWP7+978zZswYunTpgq7rNG3aFI/n1+mBNtuvI9RMJhOaduGRfKX9fjs/D4GqqkXeX6z+KxESEhLsl67rKIrCkiVLsFiKJ5ldtWoVW7ZsYePGjUyfPj2YTPFiGjZsSHJycqn68sILL5CYmMgbb7yBoih069YNj8eDyWTi448/5scff2Tz5s3079+fd999l3r16rFy5Uo2b97M119/zYwZM1ixYkWpz90wDOrWrRscuXK+lJSUIp/Nb3m9XlauXInVag2ODvH5fCxdupSRI0cSFRXFyZMnixyTnZ1NdHRg6lqDBg3YtWtXiX0rzQiJihUrkpqaGtx+6tQpKlYsvkTyt99+S3h4ODVr1gQCwbGxY8cW6Q8EvtcGDhxI165diwUdLrbvRiYBCSGE+JMxDAPN9+t0CF3XOXnkBFuXfssB13FcBEY6VNMjaGpEU91hJ9yRdcG6wpo2peKwYcX+WNj/w3fs+noOqvr/7N13eFRV+sDx7/TJZNJJgwBJgAAh9BJKCEhRlA4CCupigVVWWRXX0BVUlLILsmJBd0VYEPQXIiCWdWEpgoCg9EQIkBDSezKTTLv3/v4YGMiGpgJBPJ/nmQfm1vfMJJm57z3nPd6ERjdF0cpUOitwKi5cshMnEtXRUdhaBqHS1fwoKeAgh4sPQnHNczlkB9VSFfeFD6FzYDz+ej/06usfIpFdZOPtjdn8cPJ8cT8V+Hlr8TNrCfTRMrZvKG2ivImN9EarEeMzhVtPKs29OBNGxiGUanehWHVQQwxt+qONaoe2UWvURnMdR/rzVZefIef4KuzWHPzCuhAaMxqt/rfXDuHWuffee/nyyy957733mDFjBpWVlZ6ZBZKTk3E4HNc4AkRHR2Oz2di/fz+dOnXiq6++oqLC/XsVFRWF0+lkz549dO3ale+++w6Xy0VUVBQFBQU3tC0/51xms5mOHTuyfPlyz2wNubm5aLVaNBoNGo2Gnj170r17d7Zt20ZWVpannVcq+jhw4EDef/99PvjgAx5//HFUKhUWi4XNmzczZsyYGttWVlbSoEEDVCoVu3btIjPTPRzMYrFQVVVFly5d6NKlCwcPHuTkyZP4+/vj5+dHv3796NGjBz179qSs7MpDHv9X+/btyczM9Lw24K630bp162vuu2XLFqKiovj44489y378N78nsAAAIABJREFU8UeSkpJ46qmn6Nq1K++88w55eXmEhYVRVlbGF1984eklM3bsWIYPH86mTZsYPHgwAMXFxezcuZNhw4ZdVw+JAQMG8OqrrzJu3DgAPvvsM2bNmlVru4iICI4fP05xcTFBQUHs2bMHs9lMQECApzfMhcTEV199RfPmzQGuuk5wEwkJQRCE35Cqqio2vrWOw2UnL7s+VNIT7pSoX1VOpFJM/ccew/d8V9dLKYpCRXE1KApWB1BoPb8cKoqt7FjvHvspj07k++BsHLIT8MJH64u/3h9/XQD1teafVUVfhYqWPrG0D+x4XdsrikJxuZN9P1Xw7dFyDp22IMswomc9RiWG4GPSohG9HIQ6JFvLPEUoXRmHkMvzAVD5BKFr2tk9E0ZkG9Q+9eo40l9OlhwUnNpIydktaA3+NGz3J3zqXftCQxAApkyZwogRI5gwYQLTpk1j0qRJ+Pn50bNnT/z9/a+5v16v529/+1uNopb169f3rFu6dGmNQpOXDhG5kX7uuRYtWsTrr7/uuUj29vbmtddew2azMWvWLFwuF5IkkZiYSLt27VCr1QwePJjBgwfj5+dXq6ilyWRi1apVLFy4kL59++Lt7Y1Wq2Xs2LG1zj1lyhTmzJnD3//+d1q3bu25+LVYLDzzzDPYbDYURSE2Npa7776bPXv2eGbTkGWZiRMnEhoaSkZGxnW9Nn5+frz99tssXLiQefPm4XQ6adiwIe++++41901OTva8Rhe0b98eWZbZt28fXbp0Yfr06UyaNAlJklAUhYceesgzdCY0NJRVq1axaNEilixZgslkwmQyXXX61P8VHx/P3Xff7Sn8OWzYME9NiS1btrB161Zee+014uLieOKJJ3jooYfQ6XTo9XrefPNNVCoVhYWFTJs2zVOwtUGDBixcuBDgqusEN5Vyaana3zC73c7Ro0eJi4ur0eXrVjlw4AAdO17fl2zht0m8x78Pt+v7rCgK546e4YNPP8LpclJfHUTjkGBsme5Kj76yhkaRUfjW16DXGyiuUKGPboYu6PIXQoe3nSHzWM27Ou6PgyokVxpOxzdIAY1RTYgl0jsaf50/zcwtiDRH3pD2SJJC6lkrTsn9EWRzyJRbXe6HxUWZ1cXp3GrOFthRFDDq1XSL9WVEQjBNG/z6aQ5v1/dZuHFuxnusOKpxnT16vgfEQaSCMwCoDN5oG7c5n4Boizoo4o4YxmAt+Ymc1FU4q4sIaJBISLPhaLQ/r87LzVZXv8t1/b3zgtTUVE9dBEEQhNvV1f5WiR4SgiAItzFJkjhy5AgVqfkcOnoYFIVeutY0aajGtfc/AIQ9/DBaf3/M5ytwb1l1kJ/2ZcO+U8CpKx67oG0aDl8L6upqzAePoLVYUEnuqUA19QIYlTSHRkFNbnibDp+28On2AvafqLzseqNejZ+3Bi+9hgGdA4lv4UebaG+8DLd3oT/hzqNILqScn9wJiDMHceWkgSyBRoc2IhZj7z+gi2yLJqzpbV+I8ueQXNXkn0ymLPtb9F7BNO74PN4BMXUdliAIgnAHEgkJQRCE25DiclG0eSvfHEvjqCXPs7yTVaKRZQfOzGqMTZoQMWkSuqAgz/qKYis/fZ9Nkw7htEmMrHVcq8tKvi2Pr8s3UcZZ2mQH4zhyEntVNVE9+2IKCiY0NIqWrRLRe/36ngj/q7jCSdL7F5Mk8x6PRqdVYdCp3fUgvLUYdKL+g1A3FEVGKsx014E4c74QpaMaUKEJb4oxfgTayLZoI2JR6e7MKWIrCw+Rm7YGl72CoMZ3Exw9CLXmt1F0UxAEQfjtEQkJQRCE24jikin+5jSWY2eortBw1JWHHybuphWZlnxMce3IOt8VXK3Xc2pzFpDl2b84uwIU6DygGVp9FXs/X4nT4S5yaXFZSC0/igJoXQqNM62US6cIjWxBx/ufpmXXfje0LVU2iZPZVRRXuMguslNQ5mDPcXcxsudGNiShtR8m0etBqGNSWb5nKk5XxmGUKncxN3VgAwxxfdwJiMZtUHv51HGkN5fLUUHeT59Qkb8fg7kBDdtOwsu3cV2HJQiCINzhREJCEAThNiBV2Tj37i5c5RpABejZ5zwKKgj3a89350Cra4wpt/qSvapqHUejUxMUsZ9/f/QlZUXZWMtLwMc9LaesyBgUCT+dP1q1ltB2HejQfxQRMW1vSBvsDonTudX8cNLCD+mVpGdX43C5a0SoVRDoq8Ns0hAX5c3dnQKvcTRBuDnkqnJcmYfdPSAyDyGX5gKg8g5AF90ebWRbdJHtUPsG13Gkt4aiKFTk7SPvxCfILjvB0UOoF3k3KrX4iigIgiDcfOLTRhAE4RaoPFJA6dYMlPM1Gk7bc/jOegwF3FNbeMoLy6CSAAWrSsZMMH5eQQTGqul4T1PCoy9/Ia8oCj+WHWDrlg9RfXEYV6gZyaSiomtDwtu082wX49OCe8MH3bB2WapdrNtWwNEzVjLybdgc7vYFmLV0buFL5xgfYht7ExqoR68VQzGEW09x2HCdO+ZOQGQcQso/DSig90LXuA2GTkPcCYh6De+IQpQ/h9NWQm7ax1iKjuDlF0X9lg9jMNev67AEQRCE3xGRkBAEQbjBJKsT2eHyPK8+W07hhhPo/I14RQdgcVSx53gqqKGB7IXscqDS6tD4+qAL9wONmqJzFWgKZO4Z1pv2PWvPV21xVXKgZD/Hyo9wrvos5c5yACKOZGMIMBM+cTQqlYoWvi3pHFh72s8bwVLlYvo/T3Myu5p6vjraNTET29hEu6Y+RIUZ0WpEAkK49RRZQso5gf+pbVSmforrXCrILtBo0TZoiTHxIXRR7dCEN7ujClH+HIoiU5r9LQUn16MoMqExowhseNfPmsZXEARBEG4EkZAQBEG4AVwWB85CK1K1i/z/SwWp5ozKGrOO+uPbsuvHPXzzzTfIskz/4mIa6PUEjxyJJq4zn8z/Fss5dyJDg5nI+j61khGKorCtcCvrsz5BRsagNhInR2OubozKIXM2K42mHXsxJOrRm97mlF2FnMyuZvw9YYxKDEGt/n3dXRZuD4qiIBedPT8V5484zx4FexUBqFBCozF0GYoush3ahrGodMa6DrfO2avyyT3+L6rKTuId2ILwlg+h97r89MCC8HOUlZUxZMgQ3nrrLdqcn/Xp3Xff5dixY/z9738H4MiRIyxZsoQzZ87g7++PoigMGjSIxx9/HIA+ffqg1+sxGAzY7XY6derESy+9hE6nu6Gxnjt3jl27djFmzJgrbnPmzBkWLVpEWloafn5+6PV6nnjiCfr168fDDz/MY489xl133XXDYtqyZQv79+8nKSkJgMWLF/P1118TGBjImjVrfvXxt2/fzsSJE3nrrbfo37+/Z/nl2jJ58mR69+7NiBEjANi5cyfLli2jpKQEs9lMcHAwzz//PM2b175hcjnV1dVMmzaNY8eOodFoSEpKuuxrJ8sy8+bNY/fu3Wg0GkJCQpg3bx6hoaGebex2OyNGjMBgMLB+/XrP8u+++46FCxdit9sB+Otf/0qLFi1+3osk1AmRkBAEQfgVHPlWiv+bQdWJ4otJCDV4t1S5S0GcV6wuZvO6lew9c4Ywh4PmVgemVnfjateOXK2Ona9txW5LRW9U06R9OAC+QXaOfvuF+wAK5NpyOGk5Qab1DI0NIcT4tCBIF8QPyR9RartYT6J14uCb0lZJkjmRXc2ZPBvZhXZ2HinD16RhdK+Q311Xd6FuyRWF7iKUZw7izDiEYi0FQB0Qjj62F7rIthwrU2jfLbGOI719KLJE8dktFJ7ehEqtJbzlw/jX7y5+d4Ubxt/fn9mzZzNt2jRSUlI4c+YMq1ev5rPPPgPgp59+YsKECcyfP59evXoBUFxczIoVK2ocZ+nSpcTExCBJEuPGjeObb77hvvvuu6GxZmdns27duismJAoKCnjooYf4y1/+wrJlywAoLCxk165dNzSOS/Xt25e+fft6nn/44Yds27aNwMCfV3NJkiQ0mtq9v5KTk+natSvJyck1EhLX8u233zJjxgyWLVtG69atAUhNTaWwsPC6ExL/+Mc/MJvNfPPNN2RkZDBu3Dj+/e9/4+3tXWO7rVu3cvjwYTZu3IhWq+X111/nnXfe4eWXX/Zss3jxYtq2bUtaWppnWX5+PjNmzOCDDz4gOjoam82Gy+VC+G0QCQlBEIRfqOKHXAo3ngCVCnNcMD5tQ8j71yrsWadxnKnwbJen1/PfwEBcajURNhsJoU3Z7YpEPqnAyTQUxY7LuQuXcx9OOxzedvXzhriPykkOcxJQqTUM/ONLePsHodUZCItqecPbWlnt4s/LTpJb7ABAr1XRoJ6BPu0DxAWNcNPJ1ZW4Mg+fnw3jEHJJNgAqk//5IpRt0Ua2Q+N/8S6afOBAXYV727FVniPn+EpslWfxCW5LWIsH0Rn86zos4SYo27WL8h07bsqx/RIT8e/R46rb9OvXj6+++opFixbx/fffM23aNILOT039/vvvM2rUKE8yAiAoKIgpU6Zc9lh2ux273Y6vry8AVquVV199lSNHjgAwdOhQJkyYAEBmZiazZ8+mpKQErVbLc889R2JiItXV1SQlJZGeno5WqyUqKoo333yTuXPncu7cOYYOHUrjxo1ZunRpjXOvXr2a+Ph4hg0b5lkWHBxc4/kFmzZtYuXKlTidTgCSkpLo1q0bsiwzd+5c9uzZg16vx2QysXbtWoqLi5kyZQrFxcUAdOvWjenTp7N+/Xq2bdvG0qVLGTt2LHa7nT/84Q8kJCSQlJRESkoKa9asQZIkzGYzL7/8MtHR0axfv56NGzfi7e1NZmYmCxcupGXLmt8DSktL2bNnD1988QUDBw6ksLCQ4ODrK9y7bNkyJk2a5ElGALWOfy1ffvklb7zxBgCRkZHExcWxY8cO7r333lrbOhwO7HY7arUaq9VKRESEZ93+/fvJyMjg0UcfrZGQWLNmDUOHDiU6OhoAo1H0hvstEQkJQRCEX0B2ShR9eQpdkImwB2LRB3tTumMH9oyDhD3yCN7nP7gPHj/ON19/jY+3N8MHDMDgb+Sb9zKRJQVtUztK42NYvvovOJ1oo8LxG5xAsb2Ys1UZtc7ZwKshf4h6DKPGq8Zynd6IyTfgprb3rZRz5BU7mDSkAV1a+FLPT4dGDNEQbhLFacd17vj5qTgPIeWm4ylE2SgOQ4d70UW2Rx3cWCTErkKWnRSd+ZKijK/Q6LyJaD0Bn5AO4jUTbqpZs2Zx11130b179xo9G44fP86AAQOuuf/kyZMxGAycPXuWhIQEEhISAHj77beRZZlNmzZhtVoZM2YMMTEx9OrVixdeeIHRo0czatQo0tPTGTduHF9++SUHDhzAarXyxRfu3obl5e56S7Nnz2b+/Pk1uvxf6vjx4/S4RvLlgoSEBAYNGoRKpeL06dOMHz+eHTt2kJaWxt69e/niiy9Qq9Wec2/atIlGjRp5eoZcWH6pNWvW0Lx5c9auXYu3tzf79+/nyy+/ZPXq1ej1erZv38706dNZu3YtAIcOHWLDhg00atTosjFu3LiR3r17U69ePfr3709KSgoTJ068rvYdP36c2bNnX3H95MmTyczMvOy6devWYTQaycnJoUGDBp7l4eHh5OXl1dq+T58+7Nu3j4SEBIxGI9HR0Z5zV1VVMW/ePN555x0yMjJq7Jeenk6DBg145JFHqKioID4+nilTpqDX66+rjULdEgkJQRCEn8l6spjCDSdQHBJB90SjD/ZGstko/PRTDA0b4t+7Ny6nTFFuOZu3/BetRkeLgAS+31ZMYXEBuiqJnPgd+GYWoNtYhhToRXXf5rga+FCoygYT1DM3ZmKTSYQaw+qsnYqiUGpxcfSMhR1HyrmnUwCDu4mx5sKNp8gSUl66ZyYM17njIDlBrUHboAXGnmPRRbZDUz8GlUZ8dbkeVWWnyUldicOah194V0Kb3Y9Wb67rsISbzL9Hj2v2YrjZvvvuO8xmM6dPn8bhcFzxovDVV1/l+++/p7i4mE8//ZTwcPdwxQtDNux2O8888wwrVqxg/PjxfPfdd0yfPh2VSoXZbGbgwIF89913dOzYkdTUVEaOHAlA06ZNadmyJQcPHqRFixacOnWKOXPm0KVLF3r37n3D25uVlcWUKVPIz89Hq9VSVFREYWEhDRs2xOVyMWPGDOLj4z01E9q2bcuKFSuYP38+Xbp08SRcrmbr1q2kpaUxatQo4Px0vRUXe2J26NDhiskIgPXr1zN16lQAhg8fzowZM66ZkLjexOX/9i75NY4dO8apU6fYsWMH3t7evPbaa7zxxhvMnj2bBQsWMHbsWEJDQ2slJCRJ4ocffuDDDz/EYDDwwgsvsHz5cp5++ukbFptw84hPdUEQhOtkSTtL/trtoLg/9JXq7zn+3g7OuiKokMwoSgJayZ8ds7ZgKTuHVXMap1nCr7yUjIK3kOQyVJINB1Dvv6A3mmjV7366DXkUL7Nv3TbuEp9sy+eLfSWUWZzYne66GCajmsfvFdMBCjeGoijIxec8PSBcmYdR7FYANCFRGDoOQhfVDm3DVqj0Xtc4mnApWbJTkL6Bkqz/ojMG0KjdM5jrtarrsITfiZKSEubNm8fy5cv55z//ydKlS3nhhRcAdzf/I0eO0K9fPwBmzpwJQHx8PJIk1TqWwWCgd+/ebNu2jfHjx/+ieBo2bMjnn3/Onj172LFjB4sXL2bTpk3X3C82NtYzNORann/+eaZOnUq/fv2QZZm2bdtit9sJDg5m8+bN7N27l927d7No0SJSUlJo3749KSkp7N69mw0bNrB8+XI+/vjjq55DURRGjhzJn//858uu/99aDJc6evQo6enpzJgxw7OsoKCAAwcO0LFjRwIDAykrK6uxT2lpqad2RWxsLIcPH77iMI3r6SFRv359srOzPcfMzc0lPr72DGApKSl07doVHx8fAIYMGcL06dMBOHDgADt27ODtt9/GbrdTXl7O4MGD2bRpE/Xr1ycuLs6z34ABA9iwYcMVXxPh9iISEoIgCJeSZZznx3VeoEgSlQcOULI1D7UpGpW2Cq13Jc4Qfw781AKNSsbPp4hUzVnssgNkBXwBlQoUGcl4AquPDqN/MI3rt6Jlg474BoVi8g0kqH7jumnn/0jPqSJlZyF5pQ6OZ7oLZI5ICCY0QE+Iv47IMCM+JvGRIfxycmXR+ZkwDuE8cxDF4v49U/uFomuZ4J4Jo3Eb1N6itsEvZSlOJTf1XzhtxQRE9CKk6XA0WjGWWrh15syZw+jRo2nRogUzZsxg6NCh3HPPPbRu3ZoJEyYwfvx4OnbsSGKiu+Csw+FAluXLHkuWZb7//nsiIyMBd62F5ORkOnbs6BmG8eKLL2I2m2nZsiUpKSmMHDmSU6dOkZaWRrt27cjLy8PPz49+/frRo0cPevbsSVlZGWazGYvFcsV2jB07luHDh7Np0yYGD3YXii4uLmbnzp216khUVlZ66hwkJyfjcLhrLZWUlKDRaOjZsyfdu3dn27ZtZGVlYbPZCAsLY+DAgXTq1In+/ftf8TW4oE+fPiQlJTFmzBjCwsKQJInU1FTi4uKu+Z4kJyfzxBNP8Nxzz3mWvffee57XskePHnz22Wfcd999GAwG0tLSOHXqlGemlKeeeopZs2YRFxdHq1bu5GZaWhpFRUUkJCRcVw+JAQMGsG7dOlq3bk1GRgZHjhzhr3/9a63tIiIi2LVrF+PHj0en07F9+3aaNWsGUCORtHfv3hpDbgYNGsTixYuZOHEiOp2Ob7/9Vsyw8Rsivl0KgiBcQv/VV6SfOlVjmcqrIRq/NqhN0ejD9Kjuacah774j85iDaiWdkMaVpJ49hd27PkbrOVSKjNrLiDkuBr9G9WjQvDctfeOIMEVc4ax1R1EUNu4u4p9f5aIAMREmBsYHcV98ENHh4s608MvJNguuzCPnC1EeRC4+B4DKy/d8Icp2aCPbogkIr+NIf/skp5X8k8mU5exGbwqlcccpeAc0q+uwhN+ZL774goyMDBYtWgSAn58fs2fPZvr06SQnJ9OiRQvee+893nzzTV5++WUCAwPR6XQ8+eSThISEeI5zoYaE0+mkWbNm/OlPfwJg0qRJvPLKK54EwZAhQzyJjUWLFjF79mxWrFiBVqtlwYIFBAYGsn37ds+FryzLTJw4kdDQUIKCgoiKimLQoEFER0fXuqgODQ1l1apVLFq0iCVLlmAymTCZTJ4impeaNm0akyZNws/Pj549e+Lv706q5ubmMmvWLFwuF5IkkZiYSLt27UhJSWHFihWo1WpkWWbOnDmo1eqrvradO3fm2Wef5amnnkKSJJxOJwMGDLhmQsJut7N58+ZaPTAGDRrEkCFDmDlzJvfffz85OTncf//9qNVqjEYjixcvJiDAXZsqMTGRuXPnMnfuXMrKytBqtURERFyxGOnlPP7440ydOpX+/fujVquZO3cuZrN7CNmbb75JSEgIDz74IOPGjePkyZMMGTIErVZLeHg4r7zyyjWP36FDB3r27MmwYcPQaDTExsbyxz/+8brjE+qWSlEUpa6DuBHsdjtHjx4lLi4Og8Fwy89/oduTcOcS7/Gdy3riBK7CQhRJImfFRxgbtsCrWUdclSBZQarUIOmh0FTC7rLDlDps7t4P/0Pjb+XeTt3w0fnQuGVHfOvd+voPsizz/U+VlFS6sDtl98MhY3cq2Dz/v/gos7g4W2AnMszIjLGNiAj+fSQhxO/zjae4HLjOpXoSEFJuOigy6AxoG8a5h2BEtkUTEoVKdfUv3zfC7+U9rij4kby0j3E5LQQ16k9w9EDUmt9PIbe6ep/r+nvnBampqT97xgNBEIRb7Wp/q0QPCUEQftecpaWcW7oOlT4MlT4AfYOxKGipOgmoQDZqOWM9y+GcTygLjEJW6zFWl+HVwI8qfweV+nI0KhVB/vWZcPfLBBhv7mwXteJ3yZw8V8WxTCs/ZVVzMruKgjJnjW1UKjDo1OcfKgx6NcbzzwPMOjo08+HRe8LR627+RaJw51BkCSn/tGcqTlfWMXA5QKVG06A5xh5j0Ea2Q9ugOSqNrq7DveO47BXk/bSWioIfMJob0rDd03j5XrmonSAIgiDcjkRCQhCE363qzHLy1v6INigRlU5BH+JNhVRFeJtoDPV9+DH3W/Z8tIpqvZqqejEoKjXOOCtVjfQEBRho4BVFA2NDeob0wqy9ddXrFUXBJSmUWVzM+ziTtLPumg8GnYqGwUb6dQhkQOdAjHp30kGnVYlp/oRfTVEU5NIcXGfOJyAyD6HY3GOw1cGNMbS/F21kO3SN4lAZTHUc7Z1LURTKc/eSd+ITFMlBcJOh1Gt8Nyq1pq5DEwRBEISfTSQkBEH4XZAdDvL+9S+cRUUosh7Z0QhFCkJx2VBrM4lMegK1TseBAweo8pNY/ffpVNhyqfKPRtYYCIj0Jy6xFW0btyPEEIpWfev+fGbkVbM3tYLcEjt5pQ6Kyp0UV7iwOdxFsIb1qEevtv5EhRox6MVFiXDjyJaSi4UoMw6iVBQCoPYNRte828VClObAOo7098FpKyEndTXW4mN4+UVTP/YRDN51NzWwIAiCIPxaIiEhCMIdz1lUxNnFi3Hk5KNv0B00UYCCSpOJKUpFyAOPotbpcNqrSd+1mW+Of0+VORKXf0uMPkbGjhpL06ZN6yT2KpuLKe+mU2WXUavA36wlyFdHZIwX4YF6WkSY6NFazEog3BhyRSGurGM4zx7DlXUMucg9lZvKywdt4zbouo9GG9kOdUC46HVzCymKTOm5HRSkp6AAYc3HEBDR65bU4hAEQRCEm0kkJARBuCMpLhlnRTU5775D9alTqI0NMDR9CMWhwtwmBP+7GuFUt6W8MIe0H7ZRkpvFvi2fU2kIxhHYGrVGzeDBA4nvFH/N6tc304df51Fll5kxtjEdmvlg1KtRq8WF4O+J4rSh2Kw359g2K65zx3FlnU9AlOe7V+i90EbEYmjdB21UOzSh0eLit47YrXnkpP6L6rJ0vANjCW85Fr1XvboOSxAEQRBuCJGQEAThjiE7JPI/OY6zpApniQ0UgM7oIzoDoA0ycaRsM3n7f6JqSwkOW5VnX6fOh4rAtqBSERwbwpP3/xEv462bcUJRFOwOmdxSB+cKbGQV2jlXaGf74TK6x/qSIHpB3FEUyYliLUe2liJbSlGspcjWUhTLhX/L3OuspeCovunxqEz+aBu2wtB5KNpGcWhCIkVNgjqmyBLFZ7+h8PTnqNR66sf+Ab/wrqJniiAIgnBHEQkJQRDuGKXbM6k6WYLizEeuykNxWdGFhhDQqxf60AB27f6AtBNbCAxvTONWndHqAjiXCdZyAxbfAsBF17vjGdJ76C2N+8hpC3/7vyyKKpy4pIszMWs1KhqHGXl6eMQtjUf4dRRFRi7KwpV/yp1guJBwsJQiW8tQrKUo1RWX3Vdl8EblHYDaHIAmrCk6cwBqb39URh+4GdehWgPa+s1RB9YXF7q3keqKs+SmrsJWmYVPSHvCmz+A1uBX12EJgiAIwg0nEhKCIPwmyQ6JigO5KE53YUdnaTWVP+QBoNgPEvbQCNRGI8aICLT+/lRVlJG69z80iu1KdOs/kpNTwJkfSrAZyyj2T0Wn1/LgyHHYbLZb3pbNe4vJK3XQq60fjUOMNAw20jDEQKi/HoNeLS4Ub3OKvQpXdpp72EN2Gq7cE2C/2PsGrQG12R+VdwCawPqoG7VyJx28A1CZ/c//G+hOPGj1ddcQoc7JkpOiM5spyvw3Gp2ZiNYT8Q3tUNdhCcJ169OnD++++y4xMTE1lk+YMIFZs2bRqNHVp6Z9+OGHeeyxx7jrrrtqrVu/fj3t27cnKirKsywjI4O//e1vHD16FF9fX2RZplevXjz77LNoNBoefvhhcnJyMJvNOJ1OmjRpwrx58/Dx8WHv3r088sgjPPbYYyQlJdWIYd++ffzwww94e3vXiuPMmTMsWrSItLTqSGUJAAAgAElEQVQ0/Pz80Ov1PPHEE/Tr1++q8f9SW7ZsYf/+/Z4YFy9ezNdff01gYCBr1qz51cffvn07EydO5K233qJ///6e5Zdry+TJk+nduzcjRowAYOfOnSxbtoySkhLMZjPBwcE8//zzNG/e/LrOXV1dzbRp0zh27BgajYakpKTLvnayLDNv3jx2796NRqMhJCSEefPmERoa6tnGbrczYsQIDAYD69evB9w/M/PmzaNBgwYAREREsGzZsp//Igk3lUhICILwmyJVObFlVVC+N5vqU6WXrFFAKcaRt4PoudPQhYRw7tw5is+cIS83lwP/3YBkjMBSYuLEN5tRVAqOcAtqxYVRMvDHP/6R0NBQDhw4cFPjVxSFjDwbheVOcovtpGVVsf1wGd1ifZn6QORNPbdw4yguJ7Y9/4fz5D6kvHRQZFCp0YREYWjVG02DFmjDmqH2rQd6L5FUEq6pqiydnOOrcFTl4xfejbCY+9Hoal8MCcJv0fvvv/+rj5GSkkJAQIAnIVFQUMC4ceOYMmUKS5cuBcBisfD+++/jcDjw8nIPu5w5cyZ33XUXiqLw3HPP8fHHHzNx4kQAoqKi2LJlCy+88AIajYasrCyqqqouH8D5cz700EP85S9/8VzYFhYWsmvXrl/dvivp27cvffv29Tz/8MMP2bZtG4GBP292I0mS0GhqD8VLTk6ma9euJCcn10hIXMu3337LjBkzWLZsGa1btwYgNTWVwsLC605I/OMf/8BsNvPNN9+QkZHBuHHj+Pe//10rEbR161YOHz7Mxo0b0Wq1vP7667zzzju8/PLLnm0WL15M27ZtSUtLq7Fv9+7dPT8fwu1JJCQEQfjNUFwy5977AVeZuxeDLsSLgG4Gst95G8XpQBsYQPgjozCEhZGSksL3339/fkcZldoHvPQ4ZZC9CtDqNRgw4OdXjwEDBtTIst8sJ85ZWb45l2MZFwsUGvVq7u4YyJ+GNrjp5xduDKkkG2vKfKT8U2gjWmHsPhptozi0DVqg0t+6uiPCnUF22Sg4tYGSrG3ojAE0aj8Zc1BsXYcl/Eal7T1H2p6sm3LsFl0b0iL+lw0hvLTnRHp6OtOmTaO6upoWLVpw9uxZnnrqKc+d8X379rF8+XIKCgq49957eeGFF0hOTubo0aO8+uqrLFmyhKSkJPbu3Ut8fLznbj2A2Wzmueeeu2wMLpcLm82Gn9/F4U8mk4lmzZrx7bff0qtXL1JSUhg2bBhHjx697DFWr15NfHw8w4YN8ywLDg6u8fyCTZs2sXLlSpxOJwBJSUl069YNWZaZO3cue/bsQa/XYzKZWLt2LcXFxUyZMoXi4mIAunXrxvTp01m/fj3btm1j6dKljB07Frvdzh/+8AcSEhJISkoiJSWFNWvWIEkSZrOZl19+mejoaNavX8/GjRvx9vYmMzOThQsX0rJlyxoxlpaWsmfPHr744gsGDhxIYWEhwcHB1/OWsmzZMiZNmuRJRgC1jn8tX375JW+88QYAkZGRxMXFsWPHDu69995a2zocDux2O2q1GqvVSkTExZ/F/fv3k5GRwaOPPlorISHc/kRCQhCE254iKdiyyilIOYqrTEKlTUcqPYE1qxDr9zIqrZbI2bPQN27Mpk2bODh3Fg6bhL46Hy9rHmpDKGpTNAGB7WjRJpIufVui1d7agn1n8qpJev80iqIwtHs9Etv4UT/IgJ+3Vtw9r0OKooDThmKzINssKNUWFFslis3iflRXXlx3fr1UlIlKq8f7/pnoY7rVdROE3zBL8TFyU1fjtJUS2LA3IU2GotYa6zosQbipXnzxRf7whz8wdOhQjhw5wujRo2usz83NZfXq1VitVvr168f999/PyJEj+eyzz2oMIfjwww/p0aPHNc93IYmRm5tLVFQUw4cPr7F++PDhrF27lsTERDZv3szatWt59dVXL3us48ePX9c5ARISEhg0aBAqlYrTp08zfvx4duzYQVpaGnv37uWLL75ArVZTXl4OuBMYjRo1YsWKFQCe5Zdas2YNzZs3Z+3atXh7e7N//36+/PJLVq9ejV6vZ/v27UyfPp21a9cCcOjQITZs2HDFoTIbN26kd+/e1KtXj/79+5OSkuLpPXItx48fZ/bs2VdcP3nyZDIzMy+7bt26dRiNRnJycjzDKQDCw8PJy8urtX2fPn3Yt28fCQkJGI1GoqOjPeeuqqpi3rx5vPPOO2RkZNTad9++fQwdOhSz2cyECRPo3bv3dbVPuHVEQkIQhNtO9blCsv+5FZyBoNIAKlQqNYrsosiSxRFXIC66I/tI59epkP/+I2XaZKp0pWicFgwuBz5SKMFPdyOmfixt/Nvjpa2bu9cVVS7e+uwcAEv/FEOjUHHBcTPI1nKUqrJLkgtXTixElBVRttOFYrOA7LryQVVqVEbzxYeXGX2r3nglPIja9/ruIgnC/5KcVvJOfEp57h70pjAiO03B5N+0rsMS7gAt4iN+cS+GW8FisXDixAkGDx4MQOvWrWt17x8wYABqtRofHx+aNGnC2bNniYyMvOaxly9fzubNmykrK2Px4sV06OCuv3JhyIYkSbz00kssXLiQGTNmePaLj49nzpw5/Oc//yEmJoaAgIAb0tasrCymTJlCfn4+Wq2WoqIiCgsLadiwIS6XixkzZhAfH+9JsLRt25YVK1Ywf/58unTpQkJCwjXPsXXrVtLS0hg1ahTgTrJXVFwsmtyhQ4er1u1Yv349U6dOBdyJmRkzZlwzIXG9N1Fu5DCJY8eOcerUKXbs2IG3tzevvfYab7zxBrNnz2bBggWMHTuW0NDQWgmJ3r17c99992E0Gjl+/DgTJkxg5cqVNGnS5IbFJvx6IiEhCMJtQXY6Kf92PyXbM1DkUBRCyXO6sMkSABWSTInThUOphza8ktzqn3BJjlrH8bJkYjJq6dDtaRpE1SemTd0Ohfh8TxH//CqXarvM4K5BIhlxAymyhFJVjlSUhW3velyn9l9xW5XBG5XXxcSCwxyKObwhKi8zaqMZldHnkqSD+/9qoxkMXqhU6lvYKuFOV5H/A7k/rUVyWqgXeS/1ou5DrdHVdViCcEtd7aLWYDB4/q/RaJAk6bLbxcbGcuTIEc/ziRMnMnHiREaMGOEZJnEpjUZD//79WbBgQa1Y7r33XmbOnMnrr79+1bj/95xX8/zzzzN16lT69euHLMu0bdsWu91OcHAwmzdvZu/evezevZtFixaRkpJC+/btSUlJYffu3WzYsIHly5fz8ccfX/UciqIwcuRI/vznP192/eWKcl5w9OhR0tPTayRnCgoKOHDgAB07diQwMJCysrIa+5SWlnpqV8TGxnL48OErDtO4nh4S9evXJzs723PM3Nxc4uPja22fkpJC165d8fHxAWDIkCFMnz4dgAMHDrBjxw7efvtt7HY75eXlDB48mE2bNtWosxEbG0uHDh04fPiwSEjcZkRCQhCEOiVVO6k+VUrhpr1I1T4ohJNjd5FmLyUwzoTJ14AkSzirq1BVV2ApzabcYiHQ1xuTs5LSvCxUGg0BDSM5Z8rD1sTEi/3ewqA1XPvkN7Ndksy+nyr56N95+HhpmPpAYzrFmOs0pt8SxWlHrixGrixCrixGqSw+/7wY2VLsWYbinmVF5eWLMWEsmnoNayUWVAYTKnXNITonDhygYceOddE04XfKaS8nL20tlYU/YvRpSOP2z2D0aVjXYQnCLWU2m2nWrBmff/45gwcP5tixY5w4ceK69vX29qaystLzfOzYsQwfPpzPPvvMU8NBkqTLJiMu2Lt372V7W4wZMwYvLy8SExOvGsOFc27atMnTy6O4uJidO3fWqiNRWVnpqXOQnJyMw+G+iVJSUoJGo6Fnz550796dbdu2kZWVhc1mIywsjIEDB9KpUyf69++PLMtXjadPnz4kJSUxZswYwsLCkCSJ1NRU4uLirrrfhZieeOKJGjU33nvvPZKTk+nYsSM9evTgs88+47777sNgMJCWlsapU6do06YNAE899RSzZs0iLi6OVq1aAZCWlkZRUREJCQnX1UNiwIABrFu3jtatW5ORkcGRI0f461//Wmu7iIgIdu3axfjx49HpdGzfvp1mzZoB7qEuF+zdu5f58+d7ZtnIz8/31AjLzs7m4MGDPPXUU9eMS7i1REJCEIQ6ITslSv5zmtIfckirPotTsVPiysemV2OIUHE2L53M00qNfQxaNVJVGSZbIeTl4DJ40axZK/L6hHJUOgVo6Rd6T50nI/JK7Ly6OoNTOTZMBjUTB9anSwvfOo3pdqEoCkpVBbKl6GKSoaLIk2S4kGhQbJbaO+u9UPsEofaph65xG9Q+QaguPI9sKwpKCrclRVEoz/2OvBP/hyI7CGk6nKBG/WolyQThTvHoo4/WmM3h0gtGgPnz5zN9+nSWL19OTEwMMTExnjvfVzNmzBjeeOMN/vGPf5CUlET37t3517/+xd/+9jeWLl2Kv78/er2efv36eS6Q4WINCZfLRXh4OHPmzKl17NDQUCZMmHDNGEJDQ1m1ahWLFi1iyZIlmEwmTCbTZfedNm0akyZNws/Pj549e+Lv7w+4ewHMmjULl8uFJEkkJibSrl07UlJSWLFiBWq1GlmWmTNnDmr11Xvode7cmWeffZannnrKk4wZMGDANRMSdrudzZs31+qBMWjQIIYMGcLMmTO5//77ycnJ4f7770etVmM0Glm8eLFnSEtiYiJz585l7ty5lJWVodVqiYiIYMqUKdd8HS94/PHHmTp1Kv3790etVjN37lzMZvfNmzfffJOQkBAefPBBxo0bx8mTJxkyZAharZbw8HBeeeWVax5/9erVbNmyxfPz+PzzzxMbK4oG325UiqIo197s9me32zl69ChxcXE1unrdKhe6Nwl3LvEe/zKKorB7926KioouLpMUqn4qRLK6yJaKqFBV19ovNjaWvn37Upp3llM/7iD/5EHK888C0OqR8RT5VOPw1ZFqPU61VE3v4L50D04gzBiO+ld0sf+173NllYuJi3/CapMY1SuYYd2D8TH9PnK/isuJbPmf3gyex/kEhKUYpP+t2aBCZQ44n2wIQm0OQuVbz/P/C8tVBtMNi1X8Pt/5bof32FFdRG7qaqwlqZj8mxLe8iEM3mF1GtOdpq7e57r+3nlBamrqz57ZoK5ZrVZMJhMqlYr09HQefvhhvvrqqxqzXwiCcGe52t+q38e3ZEEQbqmzZ8+yc+dOzxRbmZmZGFQq3CNGVSiKBhXgUkBWFHSYiFDa06JLI+IHNac4N4PCjJ84sXM9+79yV4r2Mvvh36wZeYMi2OTYg06lw7vamxa+sXQJ7Epr/7Z11t4yi5Nth8o4fNrCj+kW7A6ZVx6NomPMndErQlEUd2HIS4ZQeHozXNqzoap2RXC0BtS+7sSCJqKlp0eDJ/ngUw+VOUDcLRbuKIoiU5q1nfxTn6ECwpo/QEBEoqhHIgjAjz/+yIIFC7hwT/SVV14RyQhB+B0TCQlBEG6oQ4cO8cknn+Dl5UVAQABSZSWtKitpo2jJM8URqo/CoFFz2GKjXrSG5gM6UL9pkGf/zOP7SXkzCcnpQK3RElC/MYW9QjkT7KJKqsJPqaapuRlPNn0GL03dd9Evt7iY+eFpTuXYCPHX0butP/EtfX8zyQhFllAsJZfvzXDJA5e91r4qk58nsaCp3/ySJMP5Xg4+Qe4aDmJaU+F3xG7NJef4KqrLT+Md1Ir6LcehMwZee0dB+J1ISEi4rhkkBEH4fRAJCUEQbpgzZ86wbt06z1zPquMn2PHxQYp0/dil1tJWY8RLoyE/zETvwe3x9nPfHSkuzSXHdo60E3s4uyYZyaSl8L4YXIFGZJWMrFho49OOYEMIQxoM/1VDMn4tSVZIz65i55FyDp6ykJFXjSRDTAMvlvyp2S29+FYUBRzVKNIVCngpoFSX1+zRcElRSLmiCMVa5t7wUhrt+aRCPTRhTdE1i6/Zo8EnCLU5EJVWzAwgCBcoskRR5tcUnf4CtdZA/Vbj8QuLFwk5QRAEQbgKkZAQBOGG2bf3ewB6tOqBPa2MtPWpVBmjadXUn+ByFxqHjFeTAHo83JpNb8/i5IHttY7hCjAS+NC9NA2+WH2+nj6YxJDet6oZtTicMgdPWfj2aBk/nrRQVOFOAIQGuHtEdG7uS5fmvqhUKhRFQSrMuEydhJ9JciJXlbsLQFrLUKrKUarKzy+7+O/PPY/KaL5YCDIkqkavBtX5JITK5CsuogThZ6iuOEvO8ZXYLefwDe1IWMwYtIbfRi8pQRAEQahLIiEhCMIvkpOTQ0lJCU67i+Jc9zRchw4fxE/2J+IHsJNDlH8zogAKHai0avzuakS+4wQHP9jAyQPb8W3fitPmIjQqNfFB3QkxhdOobTzh9SLrrF0OF5zMtnIqu5rTuTYyC2ycyqnGapNRq6BJfS/u6RRIzzZ+NAw2olbXvHB3/PgVVV+9deMD03uhNvmiMvmj9qmHJrQJKm8/1F6+oNVfcTe1lw8qT2HIQFQ6442PTRB+p2TJQeHpzRSf/QatzkxEmyfxDWlX12EJgiAIwm+GSEgIgnBNiqJw+lAelcXVaAqsVFSUsDnz68tu20gVgCQ5KK84jm/nTgQ3q48hwAe1l46Dxzay+9PlaIxGTG2bczDeRcd6dzOswf0EGupmjLXDKXM0w8L3P1Vy+LSFzHwNkpzuWR/ooyUmwkTXlr70iPMjyPfKF/+K00b1t2vQNGiBsfvoXxWXSq1x12gw+aEy+aHS1e1UpoIg1GQtPUlu6r9wVOXjX78Hoc1GoNF513VYgiAIgvCbIhISgiBc077NJzjwdTqoqwn2KeeonAHAvZpOGFUXL9BVqPDDG1OkN9GjHkPWaThzZC9Zew+QnZVGYVY6slHD6ccag0qFXq1nRMQo/PUBt6wtDqfEmXwbB36q5OApCyezq7E5ZADCA/W0aqDQsVU4TcK9aFLfCx+TFo36+oYv2A9sRrGU4DUsCV2jq88BLgjCb5PkslGQnkLpue3ovOrRqMOzmANb1HVYgnBb6NOnD3q9Hr1ej9Pp5LHHHmPUqFE3/BzvvvsuMTExN/S4Fzz88MPk5ORgNpsBiIqKYsmSJTflXBdUVFSwbt06JkyYcMVtCgsLWbhwIQcOHMDX1xeNRsPo0aMZPXo0U6dOJS4ujoceeuiGxXTkyBFWrFjBX//6VwDWrFnDypUrMRqNLFu2jKlTp7Jq1aobcq5nnnmGCRMm0KZNG/7zn/8QEhJCmzZtbsixAUpLS3nxxRc5e/Yser2exo0bM3fuXAID3TfDmjdvTkxMDGq1u0bZggULaN68OQBbt25lwYIFSJJEq1ateP311/Hy8uLkyZPMnz+fDz744KrnXrhwIR999BHbt28nKOhiEffmzZvzww8/4O19MZEdHx9PcnIyERERKIrCypUr+eSTTwDQ6XTExcXx4osv4ut7fUMCz5w5w9SpUykrK8Pf35/58+cTGRlZa7vi4mKmTZtGbm4uLpeL+Ph4Zs6ciVarJTk5mRUrVqBWq5FlmVGjRvHII4/U2P/06dMMHz6csWPHkpSUdF2xXUokJARBuKqS3EoOfJ2OuZGTtNIfOCdDmF8wCb17EdsilpPPPYdGceLdsiV+iYn4tm9PUUEGX65awIn92wDQGb2oCtRS3dSH1m37Mrn9kwBoVVq06pv7Z6iyysXWH0s5mV3FmTwb2UV27E53EUd/bw1to820b2amc3NfwgL0/PjjD3TsGPqzz+PKTsP23adom3QSyQhBuENVFh0lN3U1LnsZgQ37ENJ0KGqN6L0kCJdaunQpMTExnDhxghEjRpCYmEho6M//XK1LM2fO5K677vpF+7pcLrTan/fdpqKigg8++OCKCYnq6moeeughRowYwRtvvIFaraayspLNmzf/ohivR+vWrT3JCIBVq1axYMECT6LglyQjLvfaHDp0iKqqKs9x//Of/xAXF3dDExIqlYonnniC+Ph4AObPn8+iRYuYN2+eZ5u1a9fWSA4AWK1WZs2axerVq4mMjGTGjBn84x//4Omnn6ZZs2bo9Xr27NlD165dL3teSZLYsGEDHTp0YMOGDTz22GPXHfOSJUv4/vvv+eijj6hXrx6KovDNN99QXl5+3QmJl156ibFjxzJ06FA2bNjA7NmzWblyZa3t3n33XZo0acLy5ctxOp2MHTuWf//739x3333cc889jBgxApVKhcViYfDgwXTp0oUWLVp42vjSSy/Rr1+/627b/xIJCUEQruj48eN8vnYTio/EuVILAEOMXen0zABcRXnkv7cUrVyNX2Ii9c//kS2qymfV3ycjV1ThDDZiaeZDedtAUKv4U9PJxPq1vqVteOVfGRw5Y0WjhvBAA51ifGnawIuOMWYiQ73QaX/djB1SaS7V21biTN2Bytsf012P3qDIBUG4XbgcFvJPfEp53l4M3uFEtPkLJr/oug5LEGo5tusrjn57cy5S4xIG0qrHgOvePiYmBl9fX/Lz8wkNDWXTpk2sXLkSp9NdGDopKYlu3boB7l4PQ4cOZffu3RQWFvLYY4957vbv37+fOXPmANC5c2f3DFPnHT58mNdee42qqipMJhMzZsygTZs2nDt3jpEjRzJ69Gh27tyJzWZj0aJFrF27lkOHDmE0Gnn77bcJDg6+7vZc61wjRoxgz549jB49mr59+/Lqq6+Sk5OD3W5n4MCBPPnkk8iyzNy5c9mzZw96vR6TycTatWuZO3culZWVDB06FC8vL9auXVvj3J9//jn+/v788Y9/9Czz8fHhgQceqBXnd999x5IlS7Db7UiSxJNPPsnAgQMBeOutt/j8888xGAyoVCpWrlyJTqcjKSmJ9PR0tFotUVFRvPnmm+zdu5f58+ezfv16nn32WbKysnjxxRdp1aoVzz33HCNHjmTv3r2AO6mwaNEirFYrAJMnT6Z3796XfW0efPDBGvGuW7eOQYMGAbBz5062bt3K7t27+fTTT3n00UcZNmwYy5cvZ+PGjYA7UTJz5ky8vb35+9//Tnp6OqWlpRQUFNCsWTPmzZuHj49PjXP4+/t7khEA7dq14+OPP77me75jxw7i4uI8vQoeeOABpk6dytNPPw3AoEGD+PTTT6+YkNi+fTuNGjVi8uTJvPzyy9edkLBarXz44Yd89tln1KtXD3AnVe6+++7r2h/cvR6OHz/Ohx9+6In1lVdeoaSkxNMz5AKVSoXVakWWZRwOB06n05NEvNBTCMBms+F0OmsUPl++fDm9e/emqqqKqqqq647vUiIhIQhCDYokIzskPv2/9RxOPYQvJnzVZoLM/jTyCaPjg71RKU5yP/wQe3Y2Pp07EzhgAD8d3MHOtW9TXpoHTgn/MX35f/buOzqqan34+HdaZtJ7QiChkwKEUAIRCMVQpCWRrqBevAj3B76gEBQIRZogiKBcQcWrRBDEiyFCaKJICDUISk1CCWmQ3vv0949cRmIChI66P2u5ljlln33OmRlmnvPsZzdo64e1whYbuQ32SgeaWT6+L/BqrYEj54s4n1zO4ABHXh3ghoVS+lBnj9BeO03ZtkUglaEKfBFVwDAkSouH1r4gCE+W0WikJOc0WYnfoteV49RsEE7NBiKViilvBeFuTp8+jb29velJamBgIEOGDEEikXDt2jXGjRtHbGysafuqqiq+/fZbrl+/TnBwMEOHDkWhUDBt2jRWrlxJQEAAe/bsYfPmzQBoNBqmTp3KsmXL6Nq1K8eOHWPq1Kns378fgKKiIjp16kRYWBj/+c9/GDduHJs2bWLJkiUsWLCAr7/+mmnTptXZ9yVLlpiGabzyyisEBwff9Vi+vr6mdPVXX32VyZMn07lzZzQaDePGjcPX1xd7e3vi4uLYs2cPUqmU4uJiAObPn8/w4cPZsWNHnf25ePFivTMGWrduzZYtW5DJZOTl5TFs2DACAwMxGo1ERERw5MgRVCoVZWVlqFQqDh48SHl5OXv27AEw9elWH374IUFBQabsl+vXr5vWlZSU8M4777B+/XpcXFzIyclhxIgR7Nq1q85r80cnT55k/PjxAPTo0YOgoKAaw08OHTrEzp07TRkMM2fOZN26dbz11ltA9evs5g/32bNns27dujsOGzAYDHzzzTcEBQXVWP7yyy+j1+vp2bMnU6ZMwczMjMzMTBo2bGjapmHDhmRmZpr+bt++Pe++++5tjxUZGcmwYcPw9/dHq9Vy9uxZ/Pz8brv9TUlJSZiZmdG8ed3fm0tKSnj55ZfrXNeyZUs++OADMjMzcXV1RSaTASCTyXBxcSEzM7NWQGLy5MlMmTKFwMBAKisrGTt2LJ06dTKtP3DgAKtWrSItLY2wsDDTcJbExESOHDnCxo0bWbdu3V3P63ZEQEIQBJOiwiKSI06RnJ/OOcMVnCW29JN1wCnIDesGCm6sW8fV6dURai16yt0dOWR1heKPX0OWVYpeKaXM2wZfz56MfO6tJ3IOer2Bz/dk8vNvhZRW6rE2lzG2jyuWKtkDt200GjDkX8eo14JeT/muj5A6NML6hUVIrZ0eQu8FQXhaaNVFZCV+Q2nuWVTWjWnS+g1U1u5PuluCcEdtug+4pyyGR2Hq1KkYjUbS0tL46KOPMDOrrjWVnp5OWFgY2dnZyOVy8vLyyM3NNWUpDBo0CAB3d3dsbGzIyspCq9Vibm5uero9aNAg5s+fD1SPj1coFKYsi27duqFQKEhOTsbS0hILCwt69+4NQJs2bWjQoAE+Pj6mv48dO3bbc/jjkI1Lly7d8VhKpZKBAwcCUFFRwcmTJykoKDDtX15eTlJSEkOHDkWn0zFnzhwCAgLue1jInRQUFBAeHk5qaioymYzi4mKSk5Px9fWlcePGvP322wQGBtK7d2+srKzw9vYmKSmJhQsX0qVLF9M1q6/ffvuN69ev1xhuIpFISE1Nxd7evsa1qUtWVpYpC6Aux48fZ9CgQaYn9aNGjaox1KJ375VjJBAAACAASURBVN6m/UeMGMGSJUvu2N/FixdjYWFRo95GTEwMbm5ulJWV8dZbb7F27drbBqtu5eTkRF5eHlqtFoWiZqA6Pz+fkydPsnz5cgCef/55IiMj7xqQqM+DMxsbm9sGr+7Hvn378PLy4quvvqK8vJwJEyawb98+Bgyo/izp06cPffr0ISMjg9dff52ePXvi4eHBvHnzWLZsmSnocb9EQEIQBCoqKjh16hQ//PCDKRVSKVHiWeWDvOIIhRFpFAK5kkpyXJTkluVSUVUGGSmQARIbFQ2efw57H29cHD3oaN/5iZ3LB9+lc/BMEd6NLejT3p5efnZYW9z/R52+KAtd8hm0yb+hSz2LsbL095USKVYj54lghCD8hRiNRooyjpF95TuMBh0uLYfh2LgPEumDBzUF4e/g5lP0vXv3Mnv2bDp27IiTkxPTp09n1qxZ9O3bF4PBgJ+fH2q12rSfUvl7PRaZTIZer6+z/fpmOt4MhABIpdIaf9+p/fthbm5u6pfBYEAikfDdd9/V+pEKsHv3buLi4jh27BgrV64kKirqru23adOGyMjIevVlwYIFBAUF8fHHHyORSHjuuedQq9XIZDL++9//8uuvv3LixAmGDRvGf/7zH7y9vdm1axcnTpwgNjaW1atXEx0dXe9zNxqNeHl5mTJXbnX9+vUa16YuKpWqxuvgUVq+fDmpqal8+umnpgKWAG5ubkD18ISRI0eahjm4ubmZhqVA9ZT3N7eF6iwdhUJR533esWMHOp2OkJAQoLp+RmVlJeHh4ahUKhwcHCgqKjLVrdDpdJSVleHg4ICdnR1qtZrk5GSaNWtWq+36ZEi4ubmRnZ2NXq83vd5zcnJq9P+mr7/+mqVLlyKVSrG2tiYoKIi4uDhTQOKmhg0b4uvrS0xMDAMGDCAtLY2JEyea+mQ0GikrK2Px4sV19u12REBCEP6GjEYj8fHxFBUVAXDixAny8/NxktvipW9CYoUGM701lZobqBqrcHxuCrk56fyybRXSMjOMHg6UOFljcLFimN8/8WvcFekT+LJuNBpRaw1cTq/g16tlnL9WRnxaBf062jNthMd9Dc+QairQJB6tDkCknMFQWJ2aJ7FyRNGyC/ImvkiU1VF6mUNDZM5NHuo5CYLw5GgqcslM3Ex5QSIWdq1wa/0SSos/VzE+QXhaDBw4kL179/LZZ58xZ84cSktLcXevzjKKjIxEo9HctY3mzZtTVVXFqVOn8Pf3Z9++fZSUlADVs19otVpTUcHjx4+j0+lo1qwZOTk5D/Vc7uVYVlZWdOrUifXr1/P6668DkJmZiVwuRyaTIZPJ6NGjB926dSMmJob09HTTed6uIObgwYP5/PPP+c9//sP48eNNBQZ3797N6NGja2xbWlpKo0aNkEgkHD16lNTUVADKysqoqKigS5cudOnShTNnznDlyhXs7OywtbWlb9++dO/enR49epi+H9ZHhw4dSE1NrVHc8dy5c/j61q9mmKenJ8nJybi4uJiuX2np7w9/unbtysqVK3nllVewtLTku+++o1u3bqb1MTExproI27dvv209h1WrVnHhwgXWr19fIzhVXFyMUqlEpVKh0+n44YcfTJk0PXr0YPHixaSkpNC0aVO2bt1aI9sjKSnptrO9bN++nbVr19bo6/jx49m3bx/PP/883bp149tvv2X69OlAdS0NPz8/zM3NARg3bhzz58/nww8/xNHREaPRyIEDB/Dy8sLDw+OuGRKOjo74+Piwa9cuQkND2bVrFz4+PrWGa0B1VlJsbCzt2rVDo9Fw/Phx+vXrZzrHFi1aANXZN3FxcfTv35+GDRvWCNb8+9//pqKiQsyyIQhC/Rw9etQ0VhBAJpHR3awdnoYGxFep8TKm4FJ2GlsHFU1mLaOqopTdX4ZjlEHyKHecnD1wVNjT1bEbHRzr/uB/FFKzK7lyvZLk7ErSc9Rk5KvJLdKi0VVndThYywlqb8eUoe71DkYYdVp01+PRpZxBm3yGJplXKMcIZuYomvii9A9B0aw9Usf7C3AIgvD0MxoNFKQfJOfqDiQSKW7eY7BrFIhE8mBFbwXh7y4sLIxhw4YxYcIEZs+ezeTJk7G1taVHjx7Y2dnddX8zMzNWrVpVo6jlzTH9ZmZmrFmzpkahyVuHiDxM93qslStXsmzZMoKDgwGwtLTk3Xffpaqqinnz5qHT6Uz1Ctq3b49UKiU4OJjg4GBsbW1rFbW0sLBg06ZNvP/++/Tp0wdLS0vkcjljxoypdeywsDAWLlzIv//9b3x9fU3j/cvKypgyZQpVVVUYjUZat25N//79OXHihGk2DYPBwMSJE3F1dSUlJaVe18bW1pZ169bx/vvvs3TpUrRaLR4eHnz66af12r9///4cOXLENCwnJCSE2bNns2/fPlNRy0uXLpkKeLZt25ZJkyaZ9vf392fatGlkZ2fTsmVLZs2aVesYV65c4bPPPqNp06amdtzd3Vm7di3Xrl1j/vz5SCQSdDodHTp04I033gCqgyOLFi3iX//6FwaDAR8fH+bMmWNq9/Dhwzz33HO1jnf27FmKiopqBUeCg4OJjIzk+eefZ86cObz77rsEBwcjlUpxc3NjxYoVpm2nT59ORESEKRPCaDTi7+9Ply5d6nVdoTpbZtasWaxbtw4bGxvT8BGACRMmMHXqVHx9fQkPD+edd94hODgYvV5PQEAAo0aNAqoDJUePHkUul2M0GnnppZcIDAysdx/qQ2K8tVTtn5harebChQu0bdu2RsrX43L69OkaxT+Ev56/wj0uLi5m29bvuJaahLnWEfuqFtiZK+isMkcukZGm0VJemkCLstNoxwdzzDUD/dUM1EcvYswsJHOwOy88Ow1/h/p/GD4sm37MYsvP2aa/rc1luNqb0cjJDC8PCzq2ssbdWYVMeueggdFoQJ+T8r8AxG/o0i6CTg0SKbJG3uQpXWjcbTCyhp5IZCJm+1f1V3g/C3dW33tcVZZBZvwmKkuSsXLyxc17DAqV/WPoofAwPKn38pP+3nlTQkKC6WmuIPwZlZWV8eKLL7Jt2zZUKtU97fsgT+UflEajYeTIkURERGBvL/7NuJs7fVaJb9uC8DexZ+d+jpyIAUCldcDTqROBw33RRv6MvsIK9CdwKEvBUJbDtQ6enN/7OQadDrNiHRKtHstnOzKuz2u0tXu803YCZBdq2PJzNr7NLBnR04VmDVQ4WCuQyeqXsWAoyUOb/Nv/hmGcxVhRnYoodfRA2b4/8mYdUDT2RaK04Mrp0zT3aP0oT0cQhKeA0aAjL+UHcpP3IJOraNT2n9i4dhaZUIIgCI+RlZUVs2bN4vr167Rs2fJJd6feMjIymD59ughGPAQiICEIfwMVpWpOnvgFGWa42TSlZ5fueHV0ozw+nbwyOwzaPA4Vx1KiLgEpEF+AzsWMxo6e2HrY0a5XKM3aBdz1OA+L0WhEozOQU6QlPaeKXxKrxxK+0r8BbZta3WVvMFaVo007979ilGcwFFRPTyWxtEPRrH11AKJpe6Q2ohilIPwdVRankJGwCXXZDWxc/WngNRq5mfXddxQEQRAeuu7du9/XflOmTHnIPam/pk2b0rRp0yd2/L8SEZAQhL+wkvwK4qLOcvzKETSycpo7NqBH8lVkpa6kHcqo3sho4JRuHyXqEjoPHIu9bxs+Lf2Kjo6deaX5vx5rfxPSyvlqfxbZBRryS7Vodb+PKHOyUeDjYVnnfka9Dn3GJVMWhD7jMhgNoFAib+yLssMAFM06IHVuIp5+CsLfmEGvIfdaNPmpPyFX2uDhNwlr57vPCS8IgiAIwqMhAhKC8BdVWlDJ1wsOUiUrosqiEI9KPR3SKpE59EGqNKJsoKFSU8r+s1+g1lcwYMJcWj/Tn0UX54JUQmij4Y+lnwaDkbwSLTfy1Gzcn0ViegXeHhb4NLagoZOSxi5KPJxVuDspTUM0jEYjhrx000wY2rTzoKmsrgPh1gpV15HIm3VA3sgbibz2VEyCIPz9lBdeJjN+E5rKXOwa9cC11TBkcvMn3S1BEARB+Ft7bAGJ5ORkZs2aRVFREXZ2dixfvrxWmkt+fj6zZ88mMzMTnU5HQEAAc+fOrXP6G0EQajMajVQlJ5NwLJ7j8VmUK8DMIgcM0LdNH6TpShQuKhqO64jEzMjOtfOo1JTy4pxPkDS05+vUCHLUOQxsMAQn5aMbznAjr4qDvxVxPKGYG3lq1NrfMyG6tbZh3su/z7msL8pGm3QU/Q0jeoMBfXYS2uQzGMvyAZA6NETZ9tnqAERjX6TmIu1aEITf6XWV5FyJovBGLApzZ5p0nIalg9eT7pYgCIIgCDzGgMQ777zDmDFjCA0NZceOHcyfP5+NGzfW2ObTTz+lRYsWrF+/Hq1Wy5gxY9i/fz+DBg16XN0UhD+lilI1Z364RPGp01yrLOOGZWb1ChVgADdrZ2TZFsidlXhM8kcilbD3i6VcO3eM1t0Hkm5Xwqb4jwDwtPKib4PaUxjdC4PBQHqumqwCLRn5ajIL1OQVa03/FZbpAHCxU9C3owPN3cxp5KTEylxGswbVFZaNRiOaCwep+GFddfbD/0jMbZA3bW+qBSGzdXmgvgqC8NdVmneezIQt6NRFODTui0uLEKSyhz8loCAIgiAI9+exBCTy8/OJj49nw4YNAAwZMoTFixdTUFCAg4ODaTuJREJ5eTkGgwGNRoNWq8XV1fVxdFEQ/nSKc8s5ufsyBr2BnJQCSos0VMqk5P8vGDGgQVcaSB3Q5lRgViXHrKEFNv0bcmLXV+RdT+LyqRjsm7UgtZcN+1I34mjmxMSWr9PIvNF998loNFJSrmftzuscPl9sWm6pkuJsa4ajjYJmDVRYWcgJbGuDl7slVJWiObMP3dmrAFTcbKuiBF3aeeQebbAYOAWJhQ0AEnNrJBLpffdREIS/Pp2mDEX5z6SfSUJp2RCPdhMxt2129x0FQRAEQXisHktAIjMzE1dXV2QyGQAymQwXFxcyMzNrBCQmT57MlClTCAwMpLKykrFjx97z3NIXLlx4qH2/F6dPn35ixxYej6flHpflabm4Ox+MoLKSIC0toY2ikjjLDDDAKHkPLDRWgAGJhwXXXbLITYshe9VpNOXF6O3MMTpYEN9Wg6b4Au540EkTQFZ8Fllk3VNfDAY4ly7hSpaEG4US8svBaJTQzsPAMy0NOFmBhRJAU2M/bfJZ0mNOYJVxFqlBi8bCEW4NNEgklLXqQ1GzHpCaA+Q86GWrt6flPguPlrjPf0FGIzLtNRQVx5AZNWhVHalUtKfoagFQ8KR7Jzwi4r38dCgqKiIkJISPP/6Ydu3aAdXZzxcvXuTf//43AOfPn+fDDz8kOTkZOzs7jEYjQ4YMYfz48QAEBQVhZmaGUqlErVbj7+/PO++8g0LxcOtBXb9+naNHjzJ69OjbbpOcnMzKlStJTEzE1tYWMzMzXnvtNfr27cvLL7/MP//5T5599tmH1qcDBw5w6tQpZs6cCcDq1av54YcfcHBwYMuWLffc3saNG9FqtYwfP56EhASSk5Mfetb5woULOX78OGZmZlhYWDBnzhx8faunh3/55ZfJyMjAyqp6drRXXnmF4cOr65Pdbii/RqPhhRde4KuvvsLa+vZDcA8dOsTEiRP5+OOP6devn2l5Xfdl6tSp9O7dm2HDhgFw+PBh1q5dS0FBAVZWVjg7OzN9+nS8vOo3lK+yspLZs2dz8eJFZDIZM2fOrPN1YDAYWLp0KceOHTP99l26dGmNh+1qtZphw4ahVCrZvn07ACtXruTw4cOmba5du8Zbb73FK6+8Uq/+/Rk9VcUZ9u3bh5eXF1999RXl5eVMmDCBffv2MWDAgHq30bZtW5RK5SPsZd1Onz59z8ET4c/lSd5jo9GIUaNBrzeQdCabuF0XkBk1uEtS0JSnEG9lw2WpAQwwsE1vvJ/tiqqhNVqNmj3rF3N1ZywSiQS3Fm241AFynbW0sfGlhcqV/g0GYKOwva9+FZZqWfpNKheSy5HLJLg7K+noZY5fCyt6tbNDJvs9wGA0GtEmHKHyyBaM5UUYK0tAboZZuyBU/iHIXJrWav9J5EeJ9/Lfg7jPfz3aqkIyE7+hrOgcKpumFNGJjl363X1H4U/tSb2X1Wr1E30I9jSys7Nj/vz5zJ49m6ioKJKTk9m8eTPff/89AJcuXWLChAksX76cXr16AdVZ1BERETXaWbNmDZ6enuj1esaOHcuPP/740H9I37hxg2+//fa2AYmcnBxeeukl3nrrLdauXQtAbm4uR48efaj9uFWfPn3o06eP6e8NGzYQExNT4+Ftfej1ejQaDRs3bmTXrl0AJCQkEBMT89CvY8+ePQkPD0ehUHDw4EGmTZvGTz/9ZFo/d+7cOn+s324ov5mZGSEhIWzYsIGpU6fe9riRkZE888wzREZG1ghI3M2RI0eYM2cOa9euNQVOEhISyM3NrXdA4osvvsDKyooff/yRlJQUxo4dy/79+7G0rDkT3M8//8y5c+fYuXMncrmcZcuW8cknn7BgwQLTNqtXr8bPz4/ExETTshkzZjBjxgwACgoKePbZZxk4cGC9z/HP6LEEJNzc3MjOzkav1yOTydDr9eTk5ODm5lZju6+//pqlS5cilUqxtrYmKCiIuLi4ewpICMJfjdFoJOPTT8n8JZ6TltXvBUt9CcXmZzmqlAFWOMmsaCK3p0GX5vQY2I/SwlzKb+Rx6dRBrv56CO+AvnQZ8SrbC6PJLT7DKI8x9HK5/6i+Xm8gLrGU9btvkFesZVQvZ0b2csHKvO6PFF3mFSp+XI/+ejwyl2YofAKR2rpi5tcPqcX9BUMEQRCg+jOyKOMI2ZcjMRr1uLYagUPjIH799bcn3TVBeOxKz2RR8tu9ZTrWl02HBli3b3DHbfr27cu+fftYuXIlv/zyC7Nnz8bR0RGAzz//nJEjR5qCEQCOjo6EhYXV2ZZarUatVmNjUz1ks7y8nCVLlnD+/HkAQkNDmTBhAgCpqanMnz+fgoIC5HI506ZNo2fPnlRWVjJz5kyuXr2KXC6nWbNmfPTRRyxatIjr168TGhpKkyZNWLNmTY1jb968mYCAAJ5//nnTMmdn5xp/3xQdHW3KRgCYOXMmXbt2xWAwsGjRIk6cOGHKINi6dSv5+fmEhYWRn19dnLtr166Eh4ezfft2YmJiWLNmDWPGjEGtVvOPf/yDwMBAZs6cSVRUFFu2bEGv12NlZcWCBQto3rw527dvZ+fOnVhaWpKamsr777/PpUuX6Ny5MyqVisLCQtasWUNZWRmhoaF07tyZuXPnEhsby6pVq9Dr9Tg4OLBo0SKaNGlCXFwc7777Lt7e3ly8eBFzc3Pee+89WrZsWevcbw02tG/fnqysLAwGA1Lp7YfW3m0o/+DBgxk+fPhtAxKFhYWcOHGCPXv2MHjwYHJzc3F2dr7t8W61du1aJk+ebApGAPj4+NRr35v27t3Le++9B0DTpk1p27YtsbGxdQYNNBoNarUaqVRKeXk57u7upnWnTp0iJSWFV199tUZA4lY7duyga9eu9T6/P6vHEpBwdHTEx8eHXbt2ERoayq5du/Dx8akV8XN3dyc2NpZ27dqh0Wg4fvz4PUW9BOGvRq83cO2HX0g/nUXi/4IR9vZqsoxp5OplSJDQRepJK0kjLJpZYWwEvx3Yzs+bV5vasHZwpc3YV1iV9BEluhJ6OPV6oGBEem4V8yOSySrQYKGUMvvFJnRva1fntobSfCoPbURz7ickFnZYDJqKWbu+SKSy+z6+IAjCTZqKHDISNlNReAkLey8a+ryEmcVf+4ubIDzt5s2bx7PPPku3bt1qPJGPj4+v10PGqVOnolQqSUtLIzAwkMDAQADWrVuHwWAgOjqa8vJyRo8ejaenJ7169WLGjBmMGjWKkSNHcvXqVcaOHcvevXs5ffo05eXl7NmzB4Di4ur6VvPnz2f58uWmNPk/io+Pp3v37vU638DAQIYMGYJEIuHatWuMGzeO2NhYEhMTiYuLY8+ePUilUtOxo6Ojady4sSkz5ObyW23ZsgUvLy+2bt2KpaUlp06dYu/evWzevBkzMzMOHTpEeHg4W7duBeDs2bPs2LGDxo0bA7Bp0ybTsBl7e3umTp1qCnZAdVDg7bff5uuvv6Zly5Zs27aNGTNmsG3bNqA6m2Xu3LmsWLGCqKgo3n777dteq5s2b95M7969awQjVqxYwapVq/Dy8uKtt97C1dX1rkP5nZ2dUSgUJCUl0aJFi1rH2blzJ71798bJyYl+/foRFRXFxIkT63Wv4uPjmT9//m3XT506ldTU1DrXffvtt6hUKjIyMmjU6Pd6a25ubmRl1Q4CBgUFcfLkSQIDA1GpVDRv3tx07IqKCpYuXconn3xCSkrKbfuzfft23njjjXqd25/ZYxuysWDBAmbNmsW6deuwsbFh+fLlAEyYMIGpU6fi6+tLeHg477zzDsHBwej1egICAhg1atTj6qIgPHUuRMXx05Ek9DYeKM3zaNTchgvJZ9EYtHRwa02PZ7pjZW7JqR+38mvMdowxetO+jfoFoXJ2Qt7AkU+S/o3BqGdii0n42XV8oD5F7MukoETLqwMaMLCLI9Z1ZEUYtWqqTkZRdWwbGHQonxmOefcXkCgtHujYgiAIAEajgYK0A+Qk7UQikeHm8xJ2DbsjkUiedNcE4Ymybn/3LIZH7fjx41hZWXHt2jU0Gg1mZnXPbLNkyRJ++eUX8vPz2bZtmylz+uaQDbVazZQpU4iIiGDcuHEcP36c8PBwJBIJVlZWDB48mOPHj9OpUycSEhJM9QlatmyJj48PZ86cwdvbm6SkJBYuXEiXLl3o3bv3Qz/f9PR0wsLCyM7ORi6Xk5eXR25uLh4eHuh0OubMmUNAQIApm8DPz4+IiAiWL19Oly5dTAGXO/n5559JTExk5MiRwP+KiJeUmNZ37NjRFIwAyM7OvmN9i7Nnz+Lt7W3Kehg+fDgLFy6krKwMgCZNmtClSxegOhNl3rx5lJWVmepB/NHu3buJjo5m8+bNpmUrVqzAzc0NvV7PZ599xptvvsk333xz13OF6myU7OzsOgMS27dvZ9asWQAMHTqUOXPm3DUgUd9/G/6YKfMgLl68SFJSErGxsVhaWvLuu+/y3nvvMX/+fFasWMGYMWNwdXW9bUDi3Llz5OfnP5LX7NPmsQUkWrRoYYq63erzzz83/X/jxo1N6TuC8HcUfyyNhMR4isvz0Wl05GcUU2rxv6irEbKTwEliQ5+Oz+E3ohtGg4FD/13L6fPbaN31OdoHDUVt0LAubS3X7DKADFBDA5UbLzQeSyvr+o2Pu50qjYGE9ApaNTJnVK/aFR6MOi2a+ENUHd6MoTgHhVc3zIP+iczerY7WBEEQ7l1V2Q0y4jdSVZKKlVM73LxfRKGyf9LdEgSB6jHvS5cuZf369Xz55ZesWbPGNB7ex8eH8+fP07dvX6C6vgBAQEAAer2+VltKpZLevXsTExPDuHHj7qs/Hh4e7Nq1ixMnThAbG8vq1auJjo6+636tW7c2DQ25m+nTpzNr1iz69u2LwWDAz88PtVqNs7Mzu3fvJi4ujmPHjrFy5UqioqLo0KEDUVFRHDt2jB07drB+/fq7/lA3Go0MHz78tk/L/1i/4GZR0Mfhxx9/ZPXq1URERODk5GRafjPAJJPJeOWVV/j4448xGAz1GsqvVqvrrAl44cIFrl69ypw5c0zLcnJyTLVkHBwcKCoqqrFPYWGhKSu/devWnDt37rbDNOqTIdGwYUNu3LhhajMzM5OAgIBa20dFRfHMM8+YinOGhIQQHh4OVNe+iY2NZd26dajVaoqLiwkODq7x2vzuu+8IDQ1FLn+qSj4+En/9MxSEp5jBYOT49wmUFJZzPS+JrLwbVCqqxxRilCAxA0upkgFSf5QoQAoNurXAqX91xDjx5AFO7/8vZipL+rwchpnKnDOFv1JeImG0xxja23dCggQrudUDPTk0GIzEnC1iww+ZFJbqGNXLpeb6skLUv+1F/etujOVFyFybYzVkGoom7e77mIIgCLcyGnTkJu8lL2UfMrkFjdq+ho1rJ5EVIQhPkYULFzJq1Ci8vb2ZM2cOoaGhPPfcc/j6+jJhwgTGjRtHp06d6NmzJ1A9xt5gMNTZlsFg4JdffqFp06ZAda2FyMhIOnXqZBqG8fbbb2NlZYWPjw9RUVEMHz6cpKQkEhMTTTUNbG1t6du3L927d6dHjx4UFRVhZWVlygaoy5gxYxg6dCjR0dEEBwcD1cMcDh8+XKuORGlpqak2QGRkJBpN9axiBQUFyGQyevToQbdu3YiJiSE9PZ2qqioaNGjA4MGD8ff3p1+/fre9BjcFBQUxc+ZMRo8eTYMGDdDr9SQkJNC2bds6t/f09CQ5Odn0t5WVFaWlpaa/27dvT3h4uGlYRFRUFK1btzZlQKSlpXHq1Cn8/f2Jjo7G09OzzuyIgwcPsmzZMjZs2FCjPoJOp6OoqMgUoNi9ezeenp5IpdK7DuXX6/Wkp6fj6elZ63iRkZG89tprTJs2zbTss88+M70uunfvzvfff8+gQYNQKpUkJiaSlJRkGr4yadIk5s2bR9u2bWnTpg0AiYmJ5OXlERgYWK8MiQEDBvDtt9/i6+tLSkoK58+f54MPPqi1nbu7O0ePHmXcuHEoFAoOHTpEq1atAGoEHuLi4moNH6qqqmLPnj2mITl/dSIgIQhPSFlRFTvWnKA4t5wS2yRKDFkozFQ4O7rwXBsfJN9sxazpK0hkcix9nHAc2JyzR79n/8kVcLK6jYKsNMxUFvxjUQRSMwU3Kq/z/Y1IrOXWPOPUHTNp3WmS9yIhrZzPdmVw6X+ZETNfaEzbptX/KOmyklD/sgNN/CHQ65C38EfVORR5sw7iR4IgCA9NRXEymfGbUJdnYNugC66eo5Cb1Z06LAjCk7Fnzx5SUlJYuXIlALa2tsyfP5/w8HAiIyPx9vbmNzArBQAAIABJREFUs88+46OPPmLBggU4ODigUCj4v//7P1xcfn/QcbOGhFarpVWrVrz++usATJ48mcWLF5sCBCEhIabAxsqVK5k/fz4RERHI5XJWrFiBg4MDhw4dMv1YNBgMTJw4EVdXVxwdHWnWrBlDhgyhefPmtX6Iurq6smnTJlauXMmHH36IhYUFFhYWpiKat5o9ezaTJ0/G1taWHj16YGdXXVcrMzOTefPmodPp0Ov19OzZk/bt2xMVFUVERARSqRSDwcDChQvvWAQSoHPnzrz55ptMmjQJvV6PVqtlwIABtw1I9O/fn4ULFzJlyhSgOpjz5ZdfEhISQpcuXUz1IWbMmIFOp8PBwYH333/ftL+npyfbtm1jwYIFqFQqVqxYUedxZs+ejUKhqFGAMiIiAqVSycSJE02FPl1cXFi1apVpm9sN5Qf49ddf8fPzqzXtp1qtZvfu3bWySYYMGUJISAhz585lxIgRZGRkMGLECKRSKSqVitWrV2NvX51F17NnTxYtWsSiRYsoKipCLpfj7u5+28KqdRk/fjyzZs2iX79+SKVSFi1aZArWfPTRR7i4uPDiiy8yduxYrly5QkhICHK5HDc3NxYvXlyvY+zfv5/mzZvXWUj0r0hiNBqNT7oTD8PN6ZfEtJ/Co/Iw77Fer+erz77hWtoVpBjRSXU012joodUiAfTFxciceiGzbI7ts41JUccRt/trqsqqCx81bdsFvdFAijqVogAXDE6WlOvKMVCd8vhy03E841i/Yky3k1ukYcMPmRw8U4SDtZxxz7nRp4M9EgxoL59A/csOdOkXQaFC2a4vSv9gZI7ud2/4KSfey38P4j7/ORj0anKSdlKQ9jNypR1uPmOwdvK9+46Ie/x38aSn/XxS3ztvSkhIuOdZAoS/h/HjxxMWFkbr1q3vab+6ntg/TmFhYYwYMYKuXbs+keMLj8adPqtEhoQgPGYVpWr+/dFaiivysDTa4VaVgZOrC60dHLBQqjColWhynTHqFGgNlWzZNBGtsQpnj5YEjXkD1ybeZFoUs+7qR4AjNnIr2tn5YSW3oqF5I1yVDfCwbHzXftxOlcZAZGwO22JzMBjhhWddGNXLBaWhAs3J7ahP78JQnIPU1hXzPq9VT92pEk8qBUF4uMoLEslI+BptZR727j1xaTkUmdz8SXdLEAThT2HevHm3rYfwtNJoNPj7+4tgxN+MCEgIwmOUnpjLzk+PUmyRh6XWlpDcS5i7OtFi9mwkEglZe+KpjMtFL9GRULSP5LKTdB02DqlMTrO2z1BmD0eLTvPD1erpqwa7hRDk2g+VTHVf/dEbjJRW6LiWWUVKViUp2VX8drWMvGItPdvZ8c8Bbjjps1Ef/JTi8wdAq0be2BfzvhNQtAoQ03cKgvDQ6bUVZF+JpCjjKGbmzjTpNB1L+9pjiQVBEITba9q0qan+xr0ICAh4YtkRZmZmvPjii0/k2MKTIwISgvAY6MvLSdxzhJgzN8hW3gCgc3EWjV4che3/xj+m7TmNNq66wNKBzI/w7PEs457bRJmlnrSKVH6qOMbhxBgkSPCxaUNP59742vndc1+MRiPJWVV8fzSXU5dKKSzTmdbJpOBiZ8b8lxrjr0qi6ocvKLl2GmRyzNr0Rtk5BLlr7SmYBEEQHobS3LNkJm5Bpy7BsUl/nJsPQSp78Fo4giAIgiA8nURAQhAesZy0Yn7Z+iNx+fHoZFUgA/fKSlo1ssO+f39K4jIovpSJ9lo5eeoU9P1scHUdRomdFSsyV1GsLTa11d6uIyM9XsDO7N6nuMvIq+Ln34o4Fl9MclYVEgn4eFjQw9eWpg1UNHczx90OZJdjUB/5kLL860gs7VH1fAllh4FILe0e5mURBEEw0WlKyLr0X0qyT6G0aoSH32TMbZo86W4JgiAIgvCIiYCEIDxCZ366wvGdl8hRJaGXaejr3x/fDs2QXteguaEm9YMT6Eurp4dKLDnIhc4ZpNkZkaglSLKlNLVsRj/XgbSx9cVcpsJaYXNf/cgsUDPpo8todEZc7BQEd3UkuKsTHs7VQz30xTmoT21Fc/YHNFXlyNxaYREShplPDyQyxUO7HoIgCLcyGo0UZ50k+/J/MejUODcPwanpc2I4mCAIgiD8TYiAhCA8REadjrydOylPSKBUreBAiRPlqhuo5ZX4u7vj39uHpKNHsDnlhB4dxdIscrVJJGT9REWoD1nuRno49WaEx2jk0gd/e/5yqYQDvxVyMqEEncHI0vHN8GtujVQqwWg0oku/SNXJHWgvHwdA4d0dVecQZI18xLSdgiA8UtqqAjITtlCWfwFz22Y09HkZpVXDJ90tQRAEQRAeoztPeisIQr3lfv89l//f/yNv504qy7TEFTelwDwJibkeT09Pug4dyo9fvU/ivv0AnDP+wHHpDi6ofiWvXyOy3A342XVgVOMXHzgYYTAYiTlbyDtfJXP6Uim9/OyYN7YpHVraIDHoUJ8/QOmGNyjd9Da61LMonxmG7etfYjV0FnL31iIYIQjCI2M0Gii4foik44soL7yMq+dImvq/JYIRgvAXEBQUxOXLl2stnzBhAmlpaXfd/+WXX+bgwYN1rtu+fTvJyck1lqWkpDB16lSCgoJ4/vnnCQkJ4YMPPkCv15va69OnD6GhoQwaNIgpU6ZQWloKVE9v6eXlxfLly2v1wcvLi/Ly8jr7kZyczOuvv06fPn0YNmwYL7zwAj/99NNd+3+/Dhw4UKOPq1evZsCAAYwZM+a+2tu4cSNffPEFUD0V4549ex5KP2+1cOFCBgwYQEhICC+88ALnz583rbv1noSGhhIZGWlal5yczOjRo3nuuecYPXo0KSkpQPXsG8OGDTPdu9s5dOgQXl5e/PjjjzWW13Vfpk6dWqN45+HDh3nhhRfo378/w4YN41//+heXLl2q9zlXVlby5ptv0q9fPwYMGHDb14HBYGDJkiUMGjSI4OBgxo8fT3Z2do1t1Go1gwcPZtiwYaZl27dvx9/f33TdXn/99Xr37WknMiQE4SHQFhaSv3s3MktL7Mf9i7iTFRSmJmCUGHnhpbG0atWKkjNZtMgLwNrJBaQSEvsoKLB0oYVVK5oonfGzbY+37b3NFV0Xo9HIsm9SOXKhGGdbBeumemJlIcdQVkjl4S2of92NsbwIqVNjLAb+P8zaPotEcX+zdAiCINwLdXk2mQlfU1F0BUsHb9x8XsLM3OlJd0sQhEfs888/f+A2oqKisLe3p1mzZgDk5OQwduxYwsLCWLNmDQBlZWV8/vnnaDQazM2rpwmeO3cuzz77LEajkWnTpvHNN98wceJEAJo1a8aBAweYMWMGMpmM9PR0KioqbtuHnJwcXnrpJd566y3Wrl0LQG5uLkePHn3g87udPn360KdPH9PfGzZsICYmBgcHh3tqR6/Xo9Fo2LhxI7t27QKqAxIxMTEMGjToofa5Z8+ehIeHo1AoOHjwINOmTTMFbeD3e/JH77zzDmPGjCE0NJQdO3Ywf/58Nm7ciJmZGSEhIWzYsIGpU6fe9riRkZE888wzREZG0q9fv3r398iRI8yZM4e1a9fi6+sLVF+b3NxcvLy86tXGF198gZWVFT/++CMpKSmMHTuW/fv3Y2lpWWO7n3/+mXPnzrFz507kcjnLli3jk08+YcGCBaZtVq9ejZ+fH4mJiTX27datm+m1/lciAhKC8BDkRkai0xu51mIEyZtvUCUrpMImE1dHVxpbu1F+KY/c6MtIjFJybZORDfUisyCHEQ1H86xr34fSB7XWQGGplt+ulnHkQjEjeznzwrOumBUmU35gJ5r4Q6DXoWjZGWXnUORN24tMCEEQHgujQU9+2k/kXtuFRKqgYetXsHXrKj6DBOEh+/XXXzl9+vQjabtTp0507NjxvvYNCgri008/xdPTk6tXrzJ79mwqKyvx9vYmLS2NSZMmmX6gnjx5kvXr15OTk8PAgQOZMWMGkZGRXLhwgSVLlvDhhx8yc+ZM4uLiCAgIqPEU2crKimnTptXZB51OR1VVFba2tqZlFhYWtGrViiNHjtCrVy+ioqJ4/vnnuXDhQp1tbN68mYCAAJ5//nnTMmdn5xp/3xQdHc3GjRvRarUAzJw5k65du2IwGFi0aBEnTpzAzMwMCwsLtm7dSn5+PmFhYeTn5wPQtWtXwsPD2b59OzExMaxZs4YxY8agVqv5xz/+QWBgIDNnziQqKootW7ag1+uxsrJiwYIFNG/enO3bt7Nz504sLS1JTU3l/fff59KlS3Tu3BmVSkVhYSFr1qyhrKyM0NBQOnfuzNy5c4mNjWXVqlXo9XocHBxYtGgRTZo0IS4ujnfffRdvb28uXryIubk57733Hi1btqx17rcGG9q3b09WVhYGgwGp9PbJ+fn5+cTHx7NhwwYAhgwZwuLFiykoKMDBwYHBgwczfPjw2wYkCgsLOXHiBHv27GHw4MHk5ubi7Ox82+Pdau3atUyePNkUjADw8fGp17437d27l/feew+onnK1bdu2xMbGMnDgwFrbajQa1Go1UqmU8vJy3N3dTetOnTpFSkoKr776aq2AxF+VCEgIwgMqL67i6pUKrtv2o+hyMRaNddwoTMBSZkFQ4wCuf1L9xUBv1PJr8XZaDxtNRMGXAHhZ39uH3R9dzajgvW/SKCrTUl5lMC1v7qrgxcbX0P33Q0rTL4JChbL9AJT+IcgcGz3QMQVBEO5FVel1MuI3UlWahrVzexp4v4hCaXv3HQVB+Et6++23+cc//kFoaCjnz59n1KhRNdZnZmayefNmysvL6du3LyNGjGD48OF8//33/POf/zT92N2wYQPdu3e/6/FuBjEyMzNp1qwZQ4cOrbF+6NChbN26lZ49e7J79262bt3KkiVL6mwrPj6+XscECAwMZMiQIUgkEq5du8a4ceOIjY0lMTGRuLg49uzZg1Qqpbi4eja16OhoGjduTEREBIBp+a22bNmCl5cXW7duxdLSklOnTrF37142b96MmZkZhw4dIjw8nK1btwJw9uxZduzYQePGjQHYtGkT7dq1A8De3p6pU6eagh1QHRR4++23+frrr2nZsiXbtm1jxowZbNu2DYBLly4xd+5cVqxYQVRUFG+//XaNYQ912bx5M717964RjFixYgWrVq3Cy8uLt956C1dXVzIzM3F1dUUmqy5qLJPJcHFxITMzEwcHB5ydnVEoFCQlJdGiRe0p6Hfu3Env3r1xcnKiX79+REVFmTJh7iY+Pp758+ffdv3UqVNJTU2tc923336LSqUiIyODRo1+/47t5uZGVlZWre2DgoI4efIkgYGBqFQqmjdvbjp2RUUFS5cu5ZNPPjENV7nVyZMnCQ0NxcrKigkTJtC7d+96nd/TTgQkBOEBHd+RwOWKFoARNy8tifkXkSFlqD4A5Rk1AAklByg2zyQwLIyLZtcgGyY0n0RDiwcLDuw7WcCNPDXP+TvQwMEMJ5WaJnlHcErZT9X3uUhtXTHv+xrKdv2RqCzv3qAgCMJDYjBoyUveS17KPmQKS9x9J2Dt0lFkRQjCI9SxY8f7zmJ4HMrKyrh8+TLBwcEA+Pr61kqJHzBgAFKpFGtra1q0aEFaWhpNmza9a9vr169n9+7dFBUVsXr1atN1uDk8QK/X88477/D+++8zZ84c034BAQEsXLiQn376CU9PT+zt731q9bqkp6cTFhZGdnY2crmcvLw8cnNz8fDwQKfTMWfOHAICAkwBFj8/PyIiIli+fDldunQhMDDwrsf4+eefSUxMZOTIkUD1sN2SkhLT+o4dO5qCEQDZ2dl1DpW46ezZs3h7e5uyHoYPH87ChQspKysDoEmTJnTp0gWA0NBQ5s2bR1lZGVZWVnW2t3v3bqKjo9m8ebNp2YoVK3Bzc0Ov1/PZZ5/x5ptv8s0339z1XKE6GyU7O7vOgMT27duZNWsWUB1kmjNnzl0DEvX99+hhDpO4ePEiSUlJxMbGYmlpybvvvst7773H/PnzWbFiBWPGjMHV1bVWQKJ3794MGjQIlUpFfHw8EyZMYOPGjXVeiz8bEZAQhPukya/gyjcXUN0owdu8lPPmqfySUYwV5vQ396f5qwHIrMxIOneUSxEHMYR248PCTwCwU9jT1rbdA/chLrEEbw8LpvQC9anvUJ86AFo18sa+KPv/C0WrLmL6PEEQHruKomtkJGxEU56FrdszuLYagdys7i+sgiD8/dzph6BSqTT9v0wmMxWo/KPWrVvXKJY4ceJEJk6cyLBhw0zDJG4lk8no168fK1asqNWXgQMHMnfuXJYtW3bHfv/xmHcyffp0Zs2aRd++fTEYDPj5+aFWq3F2dmb37t3ExcVx7NgxVq5cSVRUFB06dCAqKopjx46xY8cO1q9ff9cf6kajkeHDh/PGG2/Uuf6P9QuUSiVqtbpe/X9QP/74I6tXryYiIgInp99rBbm5uQHV9+OVV17h448/xmAw4ObmRnZ2Nnq93nTfc3JyTNtDdbHHW18fN124cIGrV6/WCDTl5ORw+vRpOnXqhIODA0VFRTX2KSwsNNXhaN26NefOnbvtMI36ZEg0bNiQGzdumNrMzMwkICCg1vZRUVE888wzWFtbAxASEkJ4eDgAp0+fJjY2lnXr1qFWqykuLiY4OJjo6OgaNUNat25Nx44dOXfunAhICMLfUfK5LBLjruNwoxwHgxE7mYzjkkwKtOUEWLam7wuDsGzx+4dGcmV1tek8iujh1JtnnLrhYdEYmeTBAgUnE4upLC5mos0uStafAZkCsza9UXYORu765/9wEgThz8egqyInaScF6QdRqOxp3GEKVo5tnnS3BEF4SlhZWdGqVSt27dpFcHAwFy9erHNWjrpYWlrWmGVhzJgxDB06lO+//95Uw0Gv19cZjLgpLi6uzmyL0aNHY25uTs+ePe/Yh5vHjI6ONmV55Ofnc/jw4Vp1JEpLS021ASIjI9FoNAAUFBQgk8no0aMH3bp1IyYmhvT0dKqqqmjQoAGDBw/G39+ffv36YTAYuJOgoCBmzpzJ6NGjadCgAXq9noSEBNq2bVvn9p6enjVmKrGysqpxTdu3b094eLhpWERUVBStW7c2ZUCkpaVx6tQp/P39iY6OxtPTs87siIMHD7Js2TI2bNhQoz6CTqejqKjIFKDYvXs3np6eSKVSHB0d8fHxYdeuXYSGhrJr1y58fHxMP8T1ej3p6el4enrWOl5kZCSvvfZajfohn332GZGRkXTq1Inu3bvz/fffM2jQIJRKJYmJiSQlJZmGr0yaNIl58+bRtm1b2rSp/jcrMTGRvLw8AgMD65UhMWDAAL799lt8fX1JSUnh/PnzfPDBB7W2c3d35+jRo4wbNw6FQsGhQ4do1aoVUD1s56a4uDiWL19uGhKTnZ2Nq6srADdu3ODMmTNMmjTprv36MxABCUG4B0WXk9n3+QVaqGQ4WViSU1mEde5Oin28aN6gBaH/eMm07YWic0Se/g9mP8WjkEqY2WERDrauD9wHo9HI9Tw1W3ecZYn1l7iWFqLqMRZlx0FILe0euH1BEIT7UZYfT2bCZrRV+di798al5fPI5GIGH0H4u3n11VdNdQCg5o8sgOXLlxMeHs769evx9PTE09PT9LT4TkaPHs17773HF198wcyZM+nWrRtff/01q1atYs2aNdjZ2WFmZkbfvn1NPyrh9xoSOp0ONzc3Fi5cWKttV1dXJkyYcNc+uLq6smnTJlauXMmHH36IhYUFFhYWde47e/ZsJk+ejK2tLT169MDOrvo7WmZmJvPmzUOn06HX6+nZsyft27cnKiqKiIgIpFIpBoOBhQsX3rEIJEDnzp158803mTRpkikYM2DAgNsGJPr378/ChQuZMmUKUF0488svvyQkJIQuXbqY6kPMmDEDnU6Hg4MD77//vml/T09Ptm3bxoIFC1CpVLWyTW49d4VCUaMAZUREBEqlkokTJ5qCRi4uLqxatcq0zYIFC5g1axbr1q3DxsamxnSnv/76K35+frVeK2q1mt27d9fKJhkyZAghISHMnTuXESNGkJGRwYgRI5BKpahUKlavXm0antOzZ08WLVrEokWLKCoqQi6X4+7uTlhY2B2v/63Gjx/PrFmz6NevH1KplEWLFpmCNR999BEuLi68+OKLjB07litXrhASEoJcLsfNzY3Fixfftf3Nmzdz4MAB03tr+vTptG794LPzPQ0kRqPR+KQ78TCo1WouXLhA27Zt60zledRupgQJf11n957A7FQpaCWo0SFVlGDmUoJVuzasi46me/fuDB48mDJdGeevx3EwYhXKtBIkSgXPjJ1Et8CRD3T8onItkbG5HI8vxqLgCmHmmzA3k+Dw4jvIPcQTyIdFvJf/HsR9fnj02nKyL0dSlHkMMwtXGrZ+GQu72lXXHzdxj/8entR9ftLfO29KSEi459kAnrTy8nIsLCyQSCRcvXqVl19+mX379tWY/UJ4dMaPH09YWNg9/5j94xP7xy0sLIwRI0bQtWvXJ3J84cHc6bNKZEgIQj0YNFqUxwo5bLhMiiEHo8QIOiAVSL0GgJm1gp+T97D/4n+xO5CGsliLe69eDAp5E2t7pzu2fzfH44tZ+d80UJczxP4CwVbRGC2dcHhpETIHMWuGIAhPRknOb2QlfoNOW4Zj0wE4NxuMVKZ40t0SBOEp9ttvv7FixQpuPhNdvHixCEY8RvPmzbttPYSnlUajwd/fXwQj/qJEQEIQbmE0Gim/eBHdLYVvjFoDFdfKuGLMI9mYja2xIX1HVH8g5uekc+7YHnTlhZzZeBipAZwBJBJ6j32TTn2GP3iftGou/vT/2bvvMKvqQ93j37X7zJ7eKzDUofcmIiooKKKgYlQwMR411pPEelLU5BgTk2Pu0aMeY0ksoNhQiYMdS1SU3mdowwxM773ttu4fXLkhUjbKzJ4Z3s/z+Dwza//WWu9iGNzzzlq/3wfcbFvHGNcuDJ8Pa8YwIi79NZZw/Q9cRLqer6OBsl2v0FS5EVdEJpljbiEsqs/xdxSRU97pp58e1AoS0jn69esX1Kol/2ry5MkhuzvC4XBwxRVXhOTc0vlUSIj8k4bPP6fsb387+IklDGvMWCwRAzEMN2XmLhy46BszkuyBWexY/yFblz9OwGGhIyuK+GHDGZo2jvSEAdjjY0mO++5vzqtq29i5+mtchZ+T2biZS+ig1R6Fa9xc7AMnYuszAkO/hRSRLmaaJg1lX1O++zXMgIekAfOJ73uOVvMRERGR70SFhMj/075/PxUvvUR4djZR0xZQ92kFpt+kI85OlTXAgZJaXN54bJl1PHnnxZg+P6bDQsrVF3PpuGsJt4V/v/N7/Hz5eS6BTX9nUPsWRlhaaMXFLtdYyhOncMElMwgPd5ykqxUROTGethrKdr5IS00uYdEDSBt2FU53SqhjiYiISA+mQkIECHg8HPjTnzDsdmLOWkj1O8UQ5WB1aQONNU20uIsJWPx4M9tZHf066T4/TTP60O+0M1g88DosxrFnQT6WvAMtrFi5hVHV7zDZshk/VoqjRmOdNJOMCaeRbtOdECISOqYZoK74Myr2vgVAypAfEJsxA+N7/LsnIiIiAiokRACoWbkSf0sLyYsX07C+DsLtfFbRCHYLSZM9bNp8AH+KlwuvOIOqd3zsNg7w87l/Jjrh+/120FtTQunLT3GddwMBm43a/ueTOftyxsXEnaQrExH57jpayinNXUJbQz7u+GGkZi/CERYf6lgiIiLSS6iQkFNewOOhYfVanFlDsSUNpnH1bgr8jRyw5uG1tGJuNvGmdjB66mjGuEfx5Ff30WfouO9VRnj3baDlnUcxG6sYYdoozTyX4RdfRWJE7Em8MhGR78YM+KnZ/yFVBTkYFgdpw64mOnUyhmGEOpqIiIj0Iiok5JRm+gMU/v5JjPA5mAHY/Np63vOvB8BpD2Pg2P5sa97KBafNI6LKzWevPI7P087Ui3783c5nmrSvfpX2z17AE5lGjn8228Om8OfFE7FY9EZfREKvrfEAZblLaG8uIjJpHKlDLsfmjAp1LBHpIRoaGpg+fTqXXXYZv/71rwF49NFHaW1t5e677z40bunSpWzfvp0HH3wQgIKCAh566CF27txJdHQ0DoeDa6+9llmzZgV97scff5w333wTgAULFnDzzTcfcdyQIUMYPHgwFsvBR8/+9Kc/MWTIkO90vSLy/aiQkFNa8dPrCQRGUmiWs8NXQqVRA8BZU2czZsow1nvXsL6slcGuDN5+7R5a68rpP3oqGYNGn/C5TL+X1ncfx7P1Q0riJ/PLwvNxhrn4/RX9VUaISMgF/F6qCnKo2f8hNnsEGaN+QlTS2FDHEpEeJicnh9GjR7Ny5UruuusuHI7jT8hdWVnJ4sWLufPOO3n88ccBqKqq4ssvvwz6vOvWreO9994jJycHgIULFzJp0iQmTpx4xPEvv/wybrc76OOLSOdQISGnpIAvwO6vD2Ava2O9p4itljwwYOrk05g6bQoJCQkAVBRUYDWsbHxrKR0t9Vxy25/pN2LSiZ+vvZmW5b/Ht38L1imX84evx5KSaONP1w8gJkKTVopIaLXW76U0dwme1gpiUk8jefAlWO16oy7S09SXfk19afA/xJ+ImLRpxKRNOe645cuXc+edd/Lkk0+yatUqzjvvvOPu8+KLLzJ58mTmz59/aFtiYuJhnx/PO++8w/z583G5XADMnz+fd95556iFhIh0Dyok5JS0JmcXuZ/l4XBXsMtSTFRkFDffcguRkRGHxvhNHzsattHX1Y/i3V8Sndr/O5UR/vpyml/5DYG6MtrOuIUHNvantqmd6+amq4wQkZDy+9qp3PsmdcWfYXfF02fsvxMRPyzUsUSkh9q5cyf19fVMmTKFqqoqli9fHlQhkZuby7Rp0476+lNPPcXKlSuP+No999zDhAkTKCsrY9Kk//8+LTU1lXXr1h31mFdddRV+v58zzjiDW2+9Nag7OUTk5FPButyBAAAgAElEQVQhIaecTevX8+maD2gKa8YSMBg8cBAXXjz/sDICoKajhlZ/C8PoT27TSjLGBv8M4zc6ivJofOU/8fu8PGNcy5cr07DbOrjpwjTOHK0JLEUkdJqrd1C280W87XXEZZ5N0oALsdhcoY4lIt9DTNqUoO5i6Cyvv/46F110EYZhcO655/K73/2OioqKo06IG+xEuddffz3XX3/9Scv56aefkpqaSnNz86HHRH7+85+ftOOLSPBUSMgpw9fSSslflvFpdRVNRjOjjSzGjRnLoIUTjjj+iyWPkZmbz55ACQBhMYlBn8s0TSreX4J946vUBmJ4qP0aHEl9uGhYBOeMj2VAWvhJuSYRkRPl8zRTsed1Gsq+xuFOod+EOwmP6R/qWCLSw3k8HnJycnA4HKxYsQIAr9fLG2+8QWxsLCUlJYeNr6urIy7u4DLnw4YNY9u2bUc9djB3SKSmplJaWnpoe1lZGampqUfc55vtERERLFy4kGeffTb4CxWRk0qFhPRK/jYfps9P6+5aWvfU0nagAV9zBwWmm2pjHwOtfbjg5stxp0Qecf+25gYK136O3e8nYfwYxoybTZsjIahzm6ZJ1cqncW5dwRb/EKom3Mifz8giyq1vNxEJHdM0aarcSNmul/F7W0jIOp+ErPOwWPTomIh8f6tWrSIrK4tly5Yd2rZp0ybuvvtunnjiCZ544gnKy8tJSUmhvr6ed955h3vvvReAK6+8kgULFvD2228zb948AGpqavj888+ZP39+UHdIzJkzh9/97ncsWrQIgLfeeot77rnnW+MaGhpwOp24XC58Ph/vv/8+Q4cOPVl/DCJygvQTkvQ6nsoWiv53PZgHP6+ztLG7rZJ8Yz9tRhsuIljw88txxx25jDBNk3ee+k9M06R4YT+umHE7aeHpbNiw4ZjnbWr1sXlPI4HVLzCq7kNWeSeTedm/s3BozMm+RBGRE+LtaKB85zKaqjbjiuxD37E/xRWZEepYItKLLF++/FCZ8I2xY8cSCASoqanhl7/8JTfddBN+vx/TNFm8eDFTp04FIDk5mSVLlvDQQw/x8MMPEx4eTnh4ONddd13Q5588eTLnnnsuc+fOBQ5OavnNnBKrVq3i448/5oEHHmDfvn3ce++9GIaBz+dj7Nix/PSnPz1JfwoicqJUSEiv461tBxNiTs9kv7+CN/7xPljAME3OnX42YyZOIeYIZYRpmuzZ+A+2f55D4fa1NA2LYfCgSaSFpx/1XB5vgC+21bNybQ27i1tZaHuXeY7P2eCcxrirf0r/NM1SLyKhY5om9aWrqdizHDPgJWngxcT3mYlhsYY6moj0Ms8888wRt3/00UeHPj7//POPuv+AAQN44oknvleGW2+9lVtvvfVb22fOnMnMmTOBgyXJ22+//b3OIyInjwoJ6XW8dW0ARE1IY+vLHwAwuCWGcUleRp13zlH3++KNp1m7cgnO8AhaB0ZTf3oqt2ReftTx+8pa+e0LhVTWe4kIs3B76seMbvgcRp7H2XNvwmKxnNwLExE5AZ62asryltJSu5PwmEGkDl2M050c6lgiIiIih6iQkF7FNE3aixvBalBRUU/+/mLC/PH083sYePnRy4Xd6z9j++c5RCWkMPauO3h2/zNc3udKEpxHnshyy75m7l9SgGnCTRekMqPl7/jXf4Rz3FzCZt8Y9KzRIiInm2kGqC36hMq9KzAMCynZVxKbfjqGoZJUREREuhcVEtIrmKZJ655aSr/cy8d7v8Jn8VLzty/w2wJE+2IZ+ctLCE+LPuK+RTs38fb//pqYpHTOuvJnvFhxcBbnEVGjjzg+70Az9z23j3CnlfuvziJ15zI61r+pMkJEQq6juZTSvKW0NewjIn44qUMXYXfFhTqWiHQi0zT13kNEui3TNI/5ugoJ6RU8FS2Uv7idvZSw36zE7nFjGAaphp1r7p6HO/HIZQRA5YG9AGRfdw0vd7xPeVsZ42InEOuM/dZYry/An18rxmm38ND1A4jZ+iIda95QGSEiIWUG/FQXvk91wTtYbE7Sh/+YqJRJ+jdJpJdzuVzU1NQQHx+v73cR6XZM06SmpgaXy3XUMSokpMczTZOSv24GYGN7KYbFwsS0aaSPyGTszAHH3b++sgiAZTWvEeGMZFbybM5JmfOtcU3tcOtjeyip7uC681KI2bqUjjVv4hx/AWHn3qA3AiISEm2N+ynNfYGO5hKikieQMuQybI6oUMcSkS6QkZFBcXExVVVVoY4iInJELpeLjIyjr+ylQkJ6NF+rl4p/7Ke6o459vgparLWMzh7PBT88M+hjFO5YhzcpnJiwOH459B7CbRFHHPfFLoP9Fe1cd14Ksz1/p2PtWyojRCRkAn4PVftyqNn/ITZnFJmjbyQy8ciPmolI72S328nKygp1DBGR70yFhPRYgQ4fex/6impvDe/41wHgtoZx/oJzgz5GU10V9RXFNI+P54LUC49aRqzaVMvqPRb6JTuZ43lbZYSIhFRL3W7K8pbiaa0kJv10kgdejNUeHupYIiIiIidEhYT0OH6vn9yviqhfU0Ksr50v/DsAuHD0DCZePAur3Rr0sT547o8A+GKcZEcNO7T969wGduxvoabRS22jjy37molwmvwh+3OVESISMn5fG5V73qSu5B/YwxLoO+5nuOOyQx1LRERE5DtRISE9Tt6aYra+kcfkqDDe9K2nkVbGtbUx6aIZWE6gjKitKOJA3kaaB0Ry5lk/JMZxcBLL5lYfD7y0H0yIj7YTH2VjcnYkc1qXYW5crTJCREKiqXobZXkv4euoJ67PLJIGzMNidYY6loiIiMh3pkJCehS/188Xr+1gZKSLrwM7qaOZ6enpzPnJTzBswf91rinbz7Lf3QAG1I+PP3R3xOa9TfwlpwSf3+SP1/VnVP9ITNOkbdVf6VirMkJEup7P00zF7ldpKF+L051Kxqi7CI/WM+MiIiLS86mQkB6lvLCeQMAkz8hln7+UwW1tzLzqqhMqIwC+fONpfD4Pw2++gb3eDwm3hbO3pJV7nivAZbfw4zkp/1JGvElDn8n0VRkhIl3ENE0aKzZQvusV/L4WErLmkpA1B4vFHupoIiIiIieFCgnpEUzT5JNPPuGTjz/FjAgQ8AeINq2clZmJI+rElrerLt7Hng2fkdp/GA2xAaiEddu9/C0nH7vV4NF/H0RKrPOwMsI5YR41cRPopzJCRLqAt6OesryXaK7eiiuqL32H/QxXRHqoY4mIiIicVCokpNszTZNXlr3K1u1bABhh7YsFC4NK1+AcceKTuf3j9b8AcMYPf8ajtU9DYzqPfVBDapyDX17Z92AZEfDT9uFTdGzIwTlhHmHn/AQ2bjyp1yUi8q9M06S+9Esq9izHDPhIHnQJcX1mYhiWUEcTEREROelUSEi31t7i4a1nPmFrxRYiAhGcbR9BYno6sUOh4tkvcfXte0LHKy/YSeG2NfQdPolt1n00+5oo2XAOF5+ewA/PScXpsOAr3U3ru4/hr8g/VEboMQ0R6Wye1ipK85bSWreL8NjBpA1djCM8KdSxRERERDqNCgnptqqqqnj1L+9R01QONpiTOIWYdgcZCwdT8uifsbhcRE2cGPTx2lubePt/78ERFs651/yCX6x/Bq8zkr7ufvzbeWlYLAaevM9p+ftDGGHRuC+6E/uwGSojRKRTmWaA2gMfU5m/AsOwkpq9iJj0aborQkRERHo9FRLSLZWWlvLUU0/j8XRg2C3MmDKd6A1WwgY6Kfo/D+IpKSHxssuwOINf8u7z15+ksbaCs358Pw+uPABD80loHcn91w/BYjFoX/82bR88iS1jGO6F92AJi+zEKxQRgfbmEkpzl9DeWEhEwkhSs6/E7ooNdSwRERGRLqFCQrqljz/+GI+ng8TWkVz9qwuwlrRQtX4XjV8uB7Oe5MWLiZs164SOWbxrM0l9B/PoejfW4W/gBH48eg4Wi0Hbp8/TvvpV7IOn4L7oLgx78EWHiMiJMgM+qgvfo6rgXay2MNJH/BtRyRN0R5aIiIicUlRISLdUUVaFwxfFgIFZREeFUf7OPjACBNoq6fMfd+HOPrHJLBtrK6mrKKYl5SysI1/HGV3H7JTz6e9x0bL8Aby7v8IxZg7hc27CsFg76apERKCtoZDS3BfoaCklKmUiKYMvw+bQHVkiIiJy6lEhId1OIBCgrq6OCCOBUZFhFP75a8wOP/6mnYQNHkj4kCEndDyvL8ALf12GGfCzOz6NqJi1XBU7n5Hrc2nc8VewOwk762qcUy7VbydFpNME/B4q8/9O7YFV2JzRZI6+icjEUaGOJSIiIhIyKiSk2/n6g60E8DEoMRn/vnpc/aLpOLAaf+MmUu+6P6jSwDRNquo9rN5ex6df5hKf/yk2ZwpDzyijIgBZrz+Jx7DjnHoJrskXYwmP7oIrE5FTVUvtLkrzluJtqyI2fTpJgy7GagsLdSwRERGRkFIhId1KQ3UrGz7bDgb09UbjSHFjjymg8bO1xJ13Hs6UlOMeo7rBy33P76OwtIkpNb8j018DQHRWFpsoYmJpG5ET5uOafAkWt4oIEek8fl8bFXveoL7kc+xhifQd93PccSd2l5eIiIhIb6VCQroF0zTZuf4AH772CT5bJQ5sRLY5sKSUUbNiBa6BA0mcPz+oY738wT5q9u9gVthqAv4aRp25gL7Dx5NT8QKWgMnF035LeMKATr4iETnVNVVtpWzni/g6Gonvew6J/edhsTpCHUtERESk21AhId3Cx0u38OW2D2m31QEwMXkEaTMyKH749wBk/vznQS3x+eZLy2j/6AkmYhKogyGTZjL9ypvY9uWj7EsymGEOIEplhIh0Ip+nifJdr9JYsQ5nRDqZo24kLLpfqGOJiIiIdDsqJCTkfF4/mzdtod1VxyTLEMZkjyR9Th8K7rsPgMzbbsPmdh/3OO+v3sf2T5djscVzzuKfMXTEUHZRyO823kZdhBe3z8KM0T/u7MsRkVOUaZo0VqyjfNcr+H3tJPafR0K/2RgW/a9WRERE5Ej0LklCrrGmFcPegAMbY5OyST5zIOXLnifQ0UHf++4jPCvruMfo8AZY9eL/kOSrYvoP72PSGTOo89Tx3NaniW/zcqWnHxPP/gUO6/HvshAROVHe9jrKdr5Ec/U2wqKySB12Fa6ItFDHEhEREenWVEhISHk9ft59aj1eawvx1mj63DKJltwdNG/YQMS4cUGVEQCllc0kt20gvM8kJsw4g7q6ApbueBgzDC5pzWLk7F9jWKydfDUicqoxzQD1JV9QsecNTDNA8uCFxGWehWFYQh1NREREpNtTISEh9dWbmyhs/ZoWSxMDIvpT+Lv7ad+3D1tMDEkLFwZ9nB3rvgSgJcvL7RtuxmsEMFwmc1rSVUaISKfoaK2gLHcprfV7cMdlk5q9CEd4YqhjiYiIiPQYQRcSX375JStXrqS2tpa//OUvbNu2jebmZqZOndqZ+aQXa8mrpnrbdlosTZxhHUnfwg14bR3Ez5tH7NlnY4+NDeo4DdXl7Pp0GT5rGAeGtjKssp0sZyYjhl9Gn5SxnXwVInKqMQN+aopWUZX/NobFRurQq4hJOw3DMEIdTURERKRHCaqQWLJkCS+88AILFy7k/fffB8DlcvHAAw+okJAT1lzfRsHaUnxf7SU3cIBIayT9KjdjdlSTsHgxcbNmBX2s0vwdvP7Qz6Gjg+Th0RywGlw67GaS++nvpYicfO1NxZTmLaG9cT+RiaNJyb4CuzMm1LFEREREeqSgConnn3+e5557joyMDJ5++mkA+vfvT0FBQaeGk97HNE22PL0Ro6GWVYENdJh+pjgSMTt2kXnbbbhHjAj6WB1tzbz3zANYTD+zBkWwfEIyUdZIlREictIFAl6qC96juvBdrDY36SOvJSppvO6KEBEREfkegiokWlpaSE1NBTj05svn82G32zsvmfRKRSt2kdTkYVlgPQHD5AK7g+j8NURNmULEqFEndKz3//YgdZVFcHoST4+IwzDs/CD9ok5KLiKnqtaGfZTlLqGjpYzolMkkD16IzRER6lgiIiIiPV5QhcTEiRN56qmnuPHGGw9te+GFF5g8eXKnBZPeqWlTBU1mI+14yWppJbq+nqjTTiP13/4t6GP4fV5ef+g2indvpn5MHK1D4nBWT+Sn086nT1RGJ6YXkVNJwN9BZf7fqT3wMTZnDJljbiEyIfi7uERERETk2IIqJH79619zww038Nprr9HS0sLs2bNxu908+eSTnZ1PepHGwnpcwI72YrDD1PR0EmbMIHHBgqCP4WlrZeuX71O8ezPWtEhqpyTi+8dCHv/5LGxWLbMnIidHc+1OyvKW4m2rJjZjBkkD52O1hYU6loiIiEivElQhkZSUxPLly9m2bRslJSWkpqYyatQoLBb9ACjB8VQ2UbF0C2YgwAGjDLfbzfA77gj6+evWxjq++uAVNr//Kvi9BOwWCmcnk9yczrU/nqEyQkROCr+3lYo9y6kv/RJHeBJ9x9+OO3ZQqGOJiIiI9EpBFRI33ngjTzzxBKNGjWLUPz3nf8stt/DYY491WjjpHVobG1nz389i2NLZ3F5JrT3AnOnTgyojfB4P7z/7B3at/xTT76OlfwSNw1NJdcAi70AmnHk3NkvQq9eKiBxVU+VmynYuw+dtIr7vbBL7z8VidYQ6loiIiEivFdRPcmvWrDni9rVr157UMNL7tLa28sT//A81Riv4K8AOfTMyOO200467b31VKSv/8hvKC/JoyI6jLjudSeHnM3vbElzRKUT+6C4MlREi8j35Ohop3/0KjRUbcEZkkDnmJsKi+oY6loiIiEivd8yf5h555BEAvF7voY+/UVRURFpaWtAnKigo4D/+4z+or68nJiaGP/7xj/Tr1+9b49555x2eeOIJTNPEMAyeffZZEhISgj6PdB+mafLC889T09rKeHMoMdYIxt9xFjExMcfd1+vp4KX7r6etrYmqs1KoTc3mrtE/Iendh/H7/bjn341h1SovIvLdmaZJQ/laKna/SsDXQeKAC0noOxvDYg11NBEREZFTwjELifLycuDgm7ZvPv5Gamoqt956a9Anuu+++7jyyiu56KKLWLFiBffeey8vvPDCYWO2bdvGY489xvPPP09iYiJNTU04HLpdtidqa2vjH//4BweKirD7rfSzpuGODjtuGWGaJrmr32fde8toa26g+swMymNGcE3SDaTvWEl7cR7ui+7EGhd8GSYi8q+MQDNFmx+juWYHYdH9SRt2FU53aqhjiYiIiJxSjllI/OEPfwBg7NixXHbZZd/5JDU1NeTm5vLss88CcMEFF3D//fdTW1tLXFzcoXHPPfcc11xzDYmJiQBERkZ+53NKaL3wP4+zv6GWKG8k020TibHbiBl3/Df7BVu/5r2/PoDV7qB5QgoNWTG0rzuNKcMP0JLzKo7R5+IYfmbnX4CI9EqmGaCu5HOcDa/TYrWQMvgHxGbOwDA0Ma6IiIhIVwvqAfxvyojm5mbq6uoOey0zM/O4+5eVlZGcnIzVevA2WKvVSlJSEmVlZYcVEvn5+WRkZLBo0SJaW1s555xzuPHGG4NeiQFg+/btQY892TZs2BCyc3cn+fn5B8uIjnTOcQ0i0majtY+V+shqCjfUHnPfwnWfAFB7wRlUJZVT/tnFXNbfSsMbDxJwJ1CQMAkzhH/O+hqfGvR17p0Mfz321s+x+soJ2NLpCJ9OQVUkBVWbQh1NOom+l08N+jqLiPRcQRUS+fn53H777ezcuRPDMA7N7wCQl5d30sL4/X527drFs88+i8fj4dprryUtLY358+cHfYwRI0bgdDpPWqZgbdiwgfHjx3f5ebubjo4Oli9fjtMXw2RXODE2J9FT0kk4b+Dx921tZtd7TxMeHcfehGr89Wn87rIp9Pn6T/gCHqKu+CPJSf06/yKOQl/jU4O+zr2PGfBTc+Ajqva9jWFxkDLsh+SXOhg/YUKoo0kn0vfyqSFUX+eOjo6Q/hJMRKS3COoe1d/85jdMnjyZtWvXEhERwbp16/jBD37Agw8+GNRJUlNTqaiowO/3AweLh8rKSlJTD7+FPy0tjTlz5uBwOIiIiGDmzJls3br1BC9JQiUQCPDE408CEOlJZ8i4MQc/HpNy3H3LC/J46o5LKMvfQX1iChZHO9nG2Qwoex/fvo2Ez7oeawjLCBHpmdqbiilY90cq975JRPwIBky9j5i00+AE7rwTERERkc4RVCGxc+dO7rjjDqKiojBNk8jISO66665vrbxxNPHx8QwdOpScnBwAcnJyGDp06GGPa8DBuSW++OILTNPE6/Xy9ddfk52dfYKXJKEQCAR4/LHHqawux2JamT7ESaA6gC3WhTM14pj7+n1e/v7Yr/EFDIpHXUjxuQeLq/kZVto+ewF79jQcY+d0xWWISC8R8Hup3LuCfWt/j7ejnoyR15M5+gbszuhQRxMRERGR/yeoQsLpdOLz+QCIjY2ltLSUQCBAfX190Cf6zW9+w9KlS5k9ezZLly7lt7/9LQDXXXcd27ZtA2Du3LnEx8dz/vnnM3/+fAYOHMill156otckIdDe3k5ZeRluTzIzG/yMWngmnqpWXBlRx9yvwxtgyZPP0FRXyb5x2Xim78blT+TG9BuI/+QpLJEJhJ//7yc0j4iInNpa6/PZt+YBqgvfJTplEgOn3kdU8rhQxxIRERGRfxHUHBLjx4/n3Xff5eKLL2b27Nlcd911OBwOpkyZEvSJBgwYwGuvvfat7U8//fShjy0WC7/4xS/4xS9+EfRxpXvwer0AxAVMJt/xU+o/rcL0BrDFhR1zv4df3IJlwys0xfUjMK6ZGXGzWdBnPh1v/Rfepmoir/oTFtex77AQEQEI+NqpzF9BbdGn2F2x9Bl7KxHxw0MdS0RERESOIqhC4p8fzbjtttsYOHAgra2tLFiwoNOCSc9SUVoKgMvhoGpFIf5mD9FT0okcnXzE8Y3V5Xzw3B/x7j6AC5PIS7OwugIs7Hcpnk3v4d31JWFn/Rhbuh7ZEZHja67JpSxvKd72OmIzZpA0cD5WmyvUsURERETkGIIqJP6ZxWJh/vz5eDweXnvtNRYtWtQZuaQH8TU18enzyzAMC8PShuCraSduZj9iz+h71H0Ktq9hf+562u39sU5Oodhexrz4BfgrC2j96Cls/cfhnHJxF16FiPREfm8L5btfp6HsKxzhyfSbcDvhMcdf1UdEREREQu+4hcRXX31FXl4effr0YdasWfh8Pl566SWefvppYmJiVEic4kzTZO8jj1Bm+nH6Y0nyhmONtxI9JeOY+1WXFIBhYfuAy0mf8AYDvVGMfPlvNLW1Ybhjcc+7HcMIaooTETlFNVZuonznMnzeZuL7zSExay4Wqz3UsUREREQkSMcsJJ566imeeOIJBg4cyN69e7niiitYu3YtDoeD+++/nzPPPLOLYkp3VZ2Tw4e1tXS4XGQH0gg0eog5JwuLw3rUfTxtLRTlbSAQlsTAzO20AbPX5RPdbyrWhEwc2adjccd03UWISI/i62igbNfLNFVuwhWRSeaYWwiL6hPqWCIiIiJygo5ZSLzyyissWbKEESNGsHnzZq644gruvvturr766i6KJ91Z7UcfUb18OXVpaSQHUpgWno4j2U3UpLSj7hMI+HnpgRuoKS2kfOjptA/ejc1v0m/hnwhPGdyF6UWkpzFNk4ayrynf/RpmwEPSwPnE9zkHw3L0AlREREREuq9jFhJ1dXWMGDECgDFjxuBwOPjRj37UJcGke2veto2KpUvx9O9Pe3s7gyyRmHYLGTeOP+YSnTWlhdSUFtJ+en/aR1eTWe/h0rDpKiNE5Jg8bdWU5b1ES20uYTEDSRu6GKc7JdSxREREROR7OO4cEqZpHvrP6XQCEAgEDr1useg5/1NRTU4OfncMq2NSMMr2M8yeSsz41GOWEQBfvnFwmdfKvjBoRxRXle0j9tbLuiKyiPRAphmgrugzKvLfwgBShlxObMYZmmNGREREpBc4ZiHR2trKsGHDDn1umuahz03TxDAM8vLyOjehdDv+lhYq9xTyblI/2ssLGWHJwm2EEX3asSeyrCrKJ3/zl3hjHFicSfygeDfOMedgCY/qouQi0pN0tJRTmruEtoZ83PHDSM1ehCMsPtSxREREROQkOWYhsWrVqq7KIT1I844dfBqbRruliSxLChMsA0m5bBj2aNdR9ynbl8urD/0M02LQMn8YEzem4CKXsMnzuzC5iPQEZsBP9f4PqN63EsPqIG3Y1USnTj7uHVgiIiIi0rMcs5BIT0/vqhzSQ3R0dPB/VqzA6zQZZxnMePdAkhZk485OOOo+nyx7lI0fvkrAZaP6koFcOOhy0v/xCBWxo4iNP/ZdFSJyamlrPEBZ7hLam4uITBpH6pDLsTl1F5WIiIhIb3TcOSRE/llVVRVe0yQ5kMRoWz+SFg7DPTDu6OOL89n00eukDB7BpuwGpo+8gMRtNURbWmgdc2EXJheR7izg91BVsJKa/R9is0eQMeonRCWNDXUsEREREelEKiTkhGzZtAmAoUYWRoTjmGUEQMmerZhmgP3TI2l3+RkePQL7jv9ivz+FQeMndkVkEenmWuv3Upq7BE9rBTFpp5E86BKsdneoY4mIiIhIJ1MhIUFrbW1l9ddfYw/YSLZG4EoIP+4+eVs/A6uFUmsNo2Mm0Leqmea2Uj5yX8EYl/76iZzK/L52Kve+SV3xZ9hd8fQZ+1Mi4oeGOpaIiIiIdJET+omwrKyMiooKxowZ01l5pBvbsGEDpmky0JtNhNNGWOaxn+vetWcNpVs20JQdRZw7iR9nXUfrK/9JvRmJf8DpXZRaRLqj5uodlO18EW97HXGZZ5M04EIstqNPjCsiIiIivU9QhURpaSm33XYbO3fuxDAMNm3axHvvvcfnn3/OAw880NkZpZtoqKvDYsJpEQcnO40YkXjUsRs/ep1Plv0PARjD/uUAACAASURBVLuFEZPP45IRNxOoPoC/YAMfeGYxrH9MV8UWkW7E52mmYs/rNJR9jcOdQr8JdxIe0z/UsUREREQkBCzBDLr33ns588wz2bhxIzbbwQ5j2rRprF69ulPDSfdSvHsPbiMMANe0DJypkUcd+9XHy+hIcOL9yXQunXELhmHQvm4FfoudVd7JDO+r58NFTiWmadJYsYH8r39LQ/laErLOp//kX6mMEBERETmFBXWHxLZt23jqqaewWCyH1oGPjIykqampU8NJ99JU30KMEUlTppsB5w444piakgJefOAneNvbsI5MY96AywEItDTg2fYxuWGTiLDGEhdl78roIhJC3o4Gyne+RFPVFlyRfeg79qe4IrXkr4iIiMipLqhCIj4+nv3795OVlXVo2969e0lNTe20YNL9eP0BbBYrKWdlHXVM7lfv421vo35cHBefdzPDokcA0LFxJfi9vN40heFDdHeEyKnANE3qS1dTsed1zICPpIEXE99nJobFGupoIiIiItINBFVIXHPNNdxwww1cf/31+Hw+cnJyePLJJ7nuuus6O590E/X19TQbHlJNJzEpR35UY9/Wr1j7zos4UhKonZpAdr+Dy3qaPg8dG1biyxzHntx4zu+nQkKkt/O0VVOWt5SW2p2ExwwidehinO7kUMcSERERkW4kqELi0ksvJSYmhldeeYXU1FTeeustfvrTnzJr1qzOzifdxNovvgbA7YvDGX7kxy2Kdm4CwHPRKCJtzYRZD8434dnxGWZrPbuHHPz7ovkjRHov0wxQW/QJlXtXYBgWUrKvJDb9dAwjqCmLREREROQUElQh4ff7mTVrlgqIU5Tf72fHlm04sRPmCDs0j8i/am2sx+Z0sddWzsSoKcDBW7Y71r6FNbEfn9b0we1qJiPR2ZXxRaSLdDSXUpq3lLaGfUTEDyd16CLsrrhQxxIRERGRbiqoQmLatGnMmTOHefPmMX78+M7OJN3Myy+/TFVLHWdYRhM+KP6o42rKCvCH2wgQYEbSmQD4CjfjryqkdtKN7N7QRkK046iFhoj0TGbAR3XhB1QXvIPF5iR9+I+JSpmk73UREREROaagCom//e1v5OTkcPvtt2OxWJg7dy4XXHABQ4YM6ex8EmJVVVXs2LGDdOKJ8MUzau7II45rrK2gomAnzZNTsVscpIUdnEG/+tPX8RmR/GxVKgGLl6tmHb3QEJGep61xP6W5L9DRXEJU8gRShlyGzREV6lgiIiIi0gMEVUgMGzaMYcOGcdddd7F27VpycnL40Y9+RGJiIm+//XZnZ5QQ2rJlCwDTbSNoig4jKiHiiOOqDuQD0JBmZ2HG5dgtdl58dR3nl23mDe8szhiTwLypCQxIC+uy7CLSeQJ+D1X73qZm/0fYHFFkjrqByKQxoY4lIiIiIj1IUIXEP+vfvz8DBgwgLS2NwsLCTogk3YVpmmzZvIXoQDRuw8WAK0YddWxjTTkA/ggbY2LHUdvoxZm7Eq/DzoXXLyYxNbGrYotIJ2up201Z7hI8bVXEpJ9O8sCLsdrDQx1LRERERHqYoAqJxsZG3n//fXJyctiyZQvTpk3j2muvZebMmZ2dT0KotLSUmtoaRhvZ+G0B3GlHXu4ToLBgC6bVIDw6DrfNze79JUy3b6K13xn0Uxkh0iv4fW1U7nmTupJ/YA9LoO+4n+GOyw51LBERERHpoYIqJKZPn87YsWO54IILePTRR4mK0vPBp4KtW7diYDDCnkrU4ISjjitsLmB7yVqssQ6mp5wNgH/LuzgMH9apC7oqroh0oqbqbZTlvYSvo564PrNIGjAPi1Ur5oiIiIjIdxdUIfHhhx+SlJTU2VmkGwkEAmzevJlEMw6X4SDu7P5HHuf38cmyRwg70EzKsNGcnzYP0+cltfgjtvoHc0a/rC5OLiInk8/TTMXuV2koX4vTnUrGqLsIj9b3tYiIiIh8f0ctJNatW8fEiRMByM/PJz8//4jjpk6d2jnJJKRWrVpFU1MTwy0jINrEkeg+4rh3n3mA2jUb8EU7mPejXwNQt+FjwgNNFKT8iBla9k+kRzJNk8aK9ZTvegW/r5WErLkkZM3BYrGHOpqIiIiI9BJHLSR++9vfkpOTA8CvfvWrI44xDINVq1Z1TjIJqe3btwOQZSThzjzyHBA+r4edaz7CG+PA9eNziU5IwesLsP+Tv+PwJ3Dm3BldGVlEThJvex1lO5fRXL0VV1Rf+g77Oa6I9FDHEhEREZFe5qiFxDdlBMDHH3/cJWGkezBNk9qaGiI8aYSF27C6j/yc+OaP3wCgZmoitw/8IQCrPtnB5EAh65MXMCtVS3yK9CSmaVJf+gUVu5djmn6SB11CXObZGBZrqKOJiIiISC9kCWbQjTfeeMTtt9xyy0kNI91DU2MT/kCAcFwYhoEj8cjL+a1/72UAjKxkUsJSAfBt/5gABmdffhGGHtcQ6TE8rVXs3/gwZXkv4orqw4Ap9xDf9xyVESIiIiLSaYKa1HLNmjVH3L527dqTGka6h727CwDIcrgACOsbc9jrZiBA7poPaGmooaWfmxHxowFYm1fPkJa1FLuziY/SUp8iPYFpBqg98DGV+SswDCup2YuISZ+GYQTVV4uIiIiIfGfHLCQeeeQRALxe76GPv1FUVERaWlrnJZOQKS0oAqC/LRFHshtH0uF3SOR+9T7v/fX3mBYwzhzGD/osorLew6svfcyvwurpmHp1CFKLyIlqby6hNHcJ7Y2FRCSMJDX7Suyu2FDHEhEREZFTxDELifLycuDgc8XffPyN1NRUbr311s5LJiFTUXrwa+3EQcoVw7/1+terlxNwWila1J//nvZfWA0rX2yrYrptI6Y9nOTxZ3R1ZBE5AWbAR3Xhe1QVvIvVFkb6iH8jKnmCHrMSERERkS51zELiD3/4AwBjx47lsssu65JAEloej4eCykJSzCSsdgv22MMnptzTtJvKsgK86WHMypqH1Tj4fPnXWyu4zb4D5/CzMOxHngRTREKvraGA0twldLSUEpUykZTBl2FzRIY6loiIiIicgo5aSBQXF5ORkQHA1KlTKSoqOuK4zMzMzkkmIVFXV0eAANm2FCJGfHseiO21W7A3ehg5+mwuSF8AQFW9h9jytTjDPDhHzurqyCIShIDfQ2X+36k9sAqbM5rM0TcRmTgq1LFERERE5BR21EJi3rx5bNq0CYBzzjkHwzAwTfOwMYZhkJeX17kJpUtVVlQC4AjYSZw35Fuvl33wMUYA+vYffWjbF9sbmG7fSCAqFWvG0C7LKiLBaandRWneErxt1cSmTydp0MVYbVqWV0RERERC66iFxDdlBMDOnTu7JIyE3v4dhQc/8DRgsR2+3F91azmta3ZAYiQjTp97aPvWzfn8u60A15ir9Ay6SDfi97VRsecN6ks+xx6WSN9xP8cd9+2iUUREREQkFIJa9vNfFRUVYRjGoUc6pPc4sLcQNy4SY/2Hbfd6Osj5399gBEwyp5x+qHioqveQWrUanOAYeXYoIovIETRVbaVs54v4OhqJ73sOif3nYbE6Qh1LREREROSQoBaav+2229i4cSMAy5cvZ+7cuVxwwQW89tprnRpOupbf56e+tYE4I5KsxXMObQ8E/Kx8+rfU5O6gPSWM6dMuP/TaX98pYbp9E/60kVijk0IRW0T+ic/TRPG2Zyja8r9Y7RFkTbyb5EGXqIwQERERkW4nqDskvvrqKx588EEAnnvuOZ599lmioqK4+eabWbhwYacGlK5h+nx8+YeleAwvdqyE98kkEPDzdc4LrP3wZfwtrXij7LQtHEVKYhYA+0pbqczbSkpYLeHjrw7tBYic4kzTpLFiHeW7XsHvayex/zwS+s3GsHynG+FERERERDpdUO9UvV4vDoeDiooK6uvrGT9+PADV1dWdGk66Tvlna9hU34zH6SW772AA8r76kK/e+hsAtZMS6Hf2Odw56HrMljpaP32ByrwyrnVVYNpdOIZMC2V8kVOat72Wsp3LaK7eRlhUFqnDrsIVkRbqWCIiIiIixxRUITF06FCefPJJSkpKOPPMMwGoqKggIiKiM7NJFynbV8tbKyppcTdgx8bYmZMo2b2N9/76AL4IG86rZ3H/2F9gMSz4a0toevkeAg1VhPuTsLuchE+bi+FwhfoyRE45phmgvuQLKva8gWkGSB68kLjMszCMoJ7GExEREREJqaAKiQceeIBHHnkEm83GnXfeCRxchWPevHmdGk66xub38vC7C2kxWpiVNAVXVjQrHrmDgN2C9cdncfPou7AYFnxFO2h+/XcABC59kLv/5uNH01IYNDU5xFcgcurpaK2gLHcprfV7cMdlk5q9CEd4YqhjiYiIiIgELahCok+fPvz5z38+bNucOXOYM2fOUfaQnsLv9VO7t54KVw1D7X2ZfMHpvLz3eRpK9uNIieWmUXdgM6F93d9pW/VXLDFJRCz8Dcu32YByRmS5Q30JIqcUM+Cn5sAqqva9jWGxkTr0KmLSTtOSuyIiIiLS4wQ929ny5ctZsWIFFRUVJCcnc9FFF3HJJZd0ZjbpAgd2lJPg9OAnwODpI9hDLsWPLsHR4mPq9IuwFe+l8d3HCNSWYOs3BveC/8ASFslXuXtIiXUwvK8KCZGu0t5UTGneEtob9xOZOJqU7CuwO2NCHUtERERE5DsJqpB44okneOutt7jmmmtIS0ujtLSUZ555hsrKSm688cbOziidaHfOBqqsFVgDFoZNHc0zr96OrcXHOdfewYDaAppf/AWWmBTcF92Jfeh0DIuV99ZWs6u4lTkT4vRbWZEuEAh4qS54l+rC97Da3KSPvJaopPH6/hMRERGRHi2oQuK1115jyZIlpKenH9p2+umns3jxYhUSPVjeJzvxNjjIt5cxMLUfH33+DP7Pc7FEu+iz9hU8bU04p15K2OlXYNgPTlr5xfZ6HnmzhMxEJz+anRLiKxDp/Vob9lGWu4SOljKiUyaTPHghNocmFBYRERGRni+oQqKtrY24uLjDtsXExNDe3t4poaRrfPHWblLcLXTgpblxM9WvbsG0GAy2mVjjMgibfQO25P6H7bOtoBnDgEduHkSY0xqi5CK9X8DfQeXeFdQWfYLNGUPmmFuITBgR6lgiIiIiIidNUIXE9OnTueOOO7j99ttJS0ujpKSEhx9+mNNPP72z80kn2bOhlDirgyJjL+FWJ637ttA+MoH5tgCDJv4Q15Qjzw9SWu0hMsyqMkKkEzXX7qQsbynetmpiM2aQNHA+VltYqGOJiIiIiJxUQS1Wf++99+J2u7nwwgsZO3Ys8+fPJywsjHvuuaez80knKVy1m8awvRSb1VjaivHFOpk2eQYZTT7s/ccfcR+f32RHYQvDNJGlSKfwe1spzV3CgY0PYxgW+o6/ndTsK1RGiIiIiEivdNw7JJqamjhw4AD33nsvDz74IHV1dcTGxmKxBNVlSDdUuruajEaTL8xKDMBel499ykAmlLcSiIzHktj3iPtt3NNImyfAWWM0q7/IydZYuZnyncvweZuI7zubxP5zsVgdoY4lIiIiItJpjllIfPrpp/zsZz+jvb0dt9vN448/zpQpU7oqm3QC0zSp/vtuaowavPjok+KgpdzL1DMvI/D3p7APPu2oM/d/trUBh81g8tDoLk4t0nv5Ohop3/UyjZUbcUZkkDnmJsKijlwKioiIiIj0JscsJB555BHuuOMOLrnkEl599VUefvhhXn755a7KJidZS0M7e3N24W+s50P/RgDaavbjibaTZnFitrcc9XENgN1FrfRNduG06+4Yke/LNE0aytdQsfs1Ar4OEgdcRELfczEsmp9FRERERE4Nx/zJsqioiMWLFxMWFsaiRYvYv39/V+WSTrApZzeuXXXk+yoAuOLCC/F1VGDGuYkv3g+GBVvWmG/t19rh579fP0BxdQcThkR1dWyRXsfbXkvR5sco3fEcjvBk+k/5FYlZ56mMEBEREZFTyjHvkAgEAv9/oM2G3+/v9EDSeXyljdgMgwrPLqLioyna/hGWyiYixw3Ct28j1tRBWMIiD9unrKaDv75XxuodDcwcG8MPZiSGKL1Iz2eaAeqK/0Hl3jcxgZTBPyA2cwaGobuOREREROTUc8xCor29nUWLFh36vKWl5bDPAV588cXOSSYn1e51JVDdTofDpMJukBUfRd7ql+mIdzLkzDPxv/0armmXH7aPaZr88q/7qKj3cPmZSfzw3NQQpRfp+TpayinLW0pr/V7ccUNJHboIR1hCqGOJiIiIiITMMQuJBx544LDPL7300k4NI52jdG8tH72wmRkJERS2bsdvNajY+B42m0HlwoGMIx7MAPb+4w7br7rRS3mdhx+dm8LlZyWHKL1Iz2YG/NQc+JCqfTkYFgdpw35IdOrUo04eKyIiIiJyqjhmIbFgwYKuyiGdaN+WMhyGgSsAB8wqrHiwe+pJWDCTG8bejuX9v+J1ubGmDT5sv+p6LwB9k1yhiC3S47U3FVGau4T2pgNEJo4lNftybE6tUiMiIiIiAscpJKR3KMuvZUBmNDT7aKAdw9tC9OhhXD3vNwdn+t+3EVu/sd+aUG91bgMA/VJVSIiciIDfS3XBSqr3f4DVHkHGyOuJSh53/B1FRERERE4hKiR6OU+7j+riRoZmJ9HW1ESDzUJ4Uz2jJiwEIFC1H7O55luPawBs3ddMRoKT1DhnV8cW6bFa6/MpzV2Cp7Wc6NQppAxeiNXuDnUsEREREZFuR4VEL1dRWIdpQpjHzwbvLjDA5m1g7Lg5AHj3bQDA3n/8Yfu1e/wUVXUwdmDkt44pIt8W8LVTmb+C2qJPsbti6TP2ViLih4c6loiIiIhIt6VCopcry6/DZgD1HRQFarCZjcT0S8QRFg6Ad99GLAl9sUQdPtv/J5vraesIcM742BCkFulZmmtyKctbire9jtiMGSQNnI/VpkedRERERESOxRLMII/Hw3//938zc+ZMxo8/+Jv0L774gqVLl3ZqOPn+yvbV0jclEm/AR63Fi729nuHjZwJgetrxFW3/1uMapmmyck010W4rk7OjQhFbpEfwe1so2fE8Bzb9D4bFTr8Jt5OafbnKCBERERGRIARVSPz+979n9+7dPPTQQ4eWqhs0aBDLli3r1HDy/fj9ASoK60mOclFl1mMaBlZPI+MmzQXAd2Ab+H2HCgnTNCmp7uDFVRXkl7az8IwkLU0ochSNFRv/L3t3Hh1Xfdj9/31nn9G+a2RtlhdZ3je2gIGAiSEYHEgIxECTJ0+hvyd5kpa2+YVf8oSl5Dyt29PfOUka+mvyPE2LnWZxUqBxgKQkJGDCKhsvaLO1WLI10mjXaDT73N8fTtT6gYBsS3O1fF7ncI515ztzP9aX8Zn7mXu/l5OvPMpY32sU195I3WX/A1/+cqtjiYiIiIjMG9O6ZOP555/n5z//OT6fD5vtbIdRVlZGf3//rIaTizPYM04yniI7DSfSfWCa+IpcZOUVAWcv18DhxlG9FoCH/qmTN9tCANSUudn1geLf+9oii1UyNkag9fuEgofx5FRRs+lzeHKqrI4lIiIiIjLvTKuQcDqdpFKpc7YNDw+Tn58/K6FkZgQ6hrEBxkiUvuQAdnOSwtLyqccTHYdwVK/FcLh45e1R3mwLsXFZNn/4YT9VxR4cjmmdQCOyKJimyVjgFfrafoSZjlO6/CMUVd/wjtvlioiIiIjI9EzriPPGG2/ki1/8Ij09PQAEg0H+4i/+gptvvnlWw8nFCbQP4y/ykU6lGbDFcMTHuOSKXQCkRvtJD5+eulzjOz/royjXwcP31rKswofLpTJC5HfikUG6D3+D3qYncGdXUHfZ/6C49kaVESIiIiIiF2FaR50PPPAAlZWV3HrrrYyPj7Njxw5KS0v57Gc/O9v55AKZpklfxwhLCnyMECJpGDgTIVZuvRaAZMch4OztPjsDEXoGYlzekIfHrQMskd8xzTTD3S/Q/upjRMbaKa+/i9otf4o7q/z9nywiIiIiIu9pWpdsuFwuvvSlL/GlL32J4eFhCgoKznuxw87OTh588EFGR0fJz89nz5491NbWvuvYjo4ObrvtNnbv3s0Xv/jF89qPnDUaDBOZiJNnGDSn+wDwZdmwO5wAJDoaseWWYCuq5Mkfnz3zZeflRZblFZlrYuE+epv2EhlrJ6toDf5Vu3F59R4REREREZkp0yokfnepxu+Ew+GpP1dVTW8xt4cffpjdu3eza9cunn76aR566CGeeOKJd4xLpVI8/PDDbN++fVqvK+8u0DGMAdjGYvQn+rERo6C0BAAzlSRx6giuhm0YhsHAWILacg+15V5rQ4vMBWaagc5nGez4KTaHm4o1nyKv/DLdcUZEREREZIZNq5C44YYbMAwD0zSntv3uw3lzc/P7Pn9oaIimpia+853vALBz504ee+wxhoeHKSwsPGfst771La699lomJyeZnJyc9l9EztXXPkJRtguSafqNSRyxcfx16wBInmmB2CTOui0kU2mOdUxw1VotUCoSGe/GHXqKgdEhcks3U15/Fw53rtWxREREREQWpGkVEi0tLef8PDAwwN/93d+xdevWae0kEAhQVlaG3X52fQK73U5paSmBQOCcQqKlpYWDBw/yxBNP8Pjjj0/373CO48ePX9DzZkJjY6Nl+/4/dTYNUJftYiIWYdIGvuQ4tpwqGhsbKTjxPPmGjeMj8E/fOEIqbaPQNURj46DVsee8uTTHMoPMJI7oIRzRoxiGh1jWdvoTS+k/fsLqZDKL9H5e+DTHi4PmWURk/ppWIfF/Kikp4ctf/jI7duzglltumZEgiUSCr3zlK/zlX/7lVHFxIdauXYvb7Z6RTOejsbGRLVu2ZHy/72ZyPMproV/gL8/mVPRsmWRfncf2G24DYPzoXliyiqK6LZz4t1Y+fGkh932kUqekv4+5NMcycyZHT9LbtJd4tJ/8ig8QCC9nyyUfsDqWzDK9nxc+zfHiYNU8x2IxS78EExFZKC6okICzC09GIpFpjfX7/fT395NKpbDb7aRSKYLBIH6/f2rMwMAA3d3d3H///QCMj49jmiYTExM89thjFxpzUQq0j1DosOMcinA6fhqMJBvqNgGQDo+RCpzEc/XdnDxz9pKYGy8pUhkhi04qGSV48klGTv8ap6eI6k1/THZRAwF90yYiIiIikhHTKiR27959zgFrJBLh5MmT077tZ1FREQ0NDRw4cIBdu3Zx4MABGhoazrlco6Kigtdee23q52984xtMTk7qLhsXINgU5JLcswtU9jKGMxHimg/cBUCy6zBg4qzbwrM/GSbba6e23GNhWpHMCw0eJ9D8XZKxUQqrrqN0+S5s9syfWSUiIiIisphNq5C44447zvnZ6/WyatWq33vbznfzyCOP8OCDD/L444+Tm5vLnj17ALjvvvv4/Oc/z7p166afWt6TrWOMNNA79iQRr5fy6mKycgoASHQcwvDm8t1jWbzdNchtVxbjdNisDSySIcn4BP1t+xnrew1XVjm1676AL7/O6lgiIiIiIovS+xYSqVSKV199lcceewyXy3XBO1q2bBn79+9/x/Zvf/vb7zr+c5/73AXvazGLR5N4YymiPng90gm+NWz54IcAMM00iY5GkkvW8/1fDVJb5mH39WUWJxaZfaZpEgoeItD6fVKJMMVLP0zx0puw2ZxWRxMRERERWbTet5Cw2+28/PLLWmNgnujvGsVuQCo1StSdA5hsrb8UgFSwEzM8SruzAYA//3gV2d4LXkZEZF5IxMboa/kXQgNH8ORUU7Ppj/HkVFodS0RERERk0ZvWufqf/OQn+cY3vkEikZjtPHKRBo7147HZmBw9RcKZi7PAMXXXkWTHIQD2tZaTl2VnabnXyqgis8o0TUbOvEz7K48wMdRE6fLbWXrJF1VGiIiIiIjMEe/59fiBAwfYuXMn+/btY3BwkO985zsUFhaec7bEr371q9nOKOch+naQPGycGH+ZZHEdS6r/45KMRHsj8bxqTp728sBHK7DZdNaLLEzxyCCB5n2Eh1vw5a/Av/oe3D5dniQiIiIiMpe8ZyHx0EMPsXPnTv7mb/4mU3nkIpw8HMCIpUnaE5x2JcGwU1uzFAAzHiF5upk3XVfjdBhsXZljcVqRmWeaaYZ7XiB48mkMw0b5qt0ULLkKw9DCrSIiIiIic817FhKmaQJw6aWXZiSMXLgzbYO8/MRbXJXrIxHuIeJdAsCGFRsASJw6CukkvxxeyieuK6UwV4v5ycISm+ilt3kvkbFOsovW4m/YjdNT+P5PFBERERERS7xnIZFOp3n11Veniol3c8UVV8x4KDl/Ta/0UOM7exeU01ndJBO51PjzqCw6e718vL2RGE4m8lZw57U6dV0WDjOdZLDrZwx2PovN4WbJmv9CbvmlWohXRERERGSOe89CIh6P8+Uvf/n3FhKGYfCLX/xiVoLJ+ek61s+1WR7SsSFeT/WCI4/Lt14y9Xj8ZCNvJ5ex/coyrR0hC0Zk/BS9TU8QmzhDbtlWyus/jsOVa3UsERERERGZhvcsJLxerwqHeSCVSJGXBhsGp9P9TDjycaeGWH/ZtWcfHwlgjPdxNHkJWwp0qYbMf+lUnIGOnzB06nkc7lyqNvw3cko2WB1LRERERETOw3sWEjI/hEaiFDntDJshfubowUinuOKOGzFsZxfyS3Q0AnA0uYL/UpdtZVSRixYeaSPQtJd4ZID8JVdRtvx27E6f1bFEREREROQ8TWtRS5nbzrQNUu5y8Hz8dbCBWRbkg1tunXo82dHIkFGEs7iC/GydISHzUyoZIXjiSUbOvIjTW0zN5j8hq3CV1bFEREREROQCvWchcfjw4UzlkIvQ9dIpqo0E/bYw7nAv3pv8OG1niwczlSDRdZRDsQ1sXqtbfcr8FBo8RqD5X0jGRims3k7pslux2V1WxxIRERERkYugSzbmubbfdLMykqbdHAQgWTnJJ7Z8PkHlmQAAIABJREFUZurxRE8TJKIcTa3g42vyrIopckGS8Qn6237IWN/ruLMqqFx/P768pVbHEhERERGRGaBCYh4LDUc48XQrqzwujkw2YnPaKLrlUqp81VNjAodfw2faqNlyKevrdIaEzA+maTLe/yZ9rT8glYxQvPRmSpbehGHTP1kiIiIiIguFPt3PY13H+ymz2wmbY4zbDbKyk1xVevU5Y1KdhziRruGuD9VYlFLk/CSiIwRavsfE4FE8uTXUrP4DPNlLrI4lIiIiIiIzTIXEPDbYNkS1086TyTcx7S5WVZSytfDSqcdToWEKoj285vkwH/BpqmVuM02T0d6D9Lf9GNNMUbbioxRWX49h2KyOJiIiIiIis0BHqfOYIzBBykwznE5gmGmuvfSGcx4//MuXWAaUrL/MmoAi0xSfHKC3eR+TI634ClZS0XAPLl+p1bFERERERGQWqZCYp/qOB6lIQi9jYBgUTfZQsHr11OPhaIrBY69Tasvm2u2bLEwq8vuZZprh7l8SbH8aw7Djb7ib/IqrMAzD6mgiIiIiIjLLVEjMU6Ffn8IGHDKaAXAWuc45iGs9NUEDbaQqN+Gw2y1KKfL7RSfO0Nu0l+h4F9nF6/Gv+gROT4HVsUREREREJENUSMxT5liMNxOd9BPCGRuh6gPrz3n8xFvHWWqbhNWX/p5XELGGmU4y2PUcA53PYnd4WbL2v5JbtlVnRYiIiIiILDIqJOahVDhOJBrmCG0AZI23s3TDH0w9nkikSXcdBiCvYaslGUXeTWSsk96mvcTCveSWX0L5yjtxuLKtjiUiIiIiIhZQITHPxAcm6fv+cc6YgwDkDB9nYp2XqrKVAKTTJl9/6jRXxJqZyK+hICvfyrgiAKRTcYLt/8Zw9y9wuPOo2vAZckrWv/8TRURERERkwVIhMc+EjvaTGIzQFZ8EOxi+BCtuvZNCVxEAr7eM8/KhAJ/M6SZr/ccsTisC4eFWepv3kogMUrBkG6Urbsfu8FodS0RERERELKZCYp6Jdo3iLPUx2j4BNpOyVSu5t/ZTU48fbg+xztmOnTSOui3WBZVFL5WM0H/iXxk98xIubwk1mx8gq7De6lgiIiIiIjJHqJCYR9KJFNEzIex1+aQZwjBTLL/8g+eMGRxLssV9ElxeHEtWWZRUFrvQwFECLd8lGRunqOYGSupuwWZ3WR1LRERERETmEBUS80i4aRBSJj3jUWL2MI5knMrateeMiUaTrDVacdZuxLBreiWzkvEQfa0/YLz/TdzZS6ja8Bm8uTVWxxIRERERkTlIR6zzSOitPgB+1fM8KaeXkvAotTnLzhmTGjxFAaM4l19iRURZpEzTZLz/Dfpaf0AqGaWk7haKa3dg2PRPjIiIiIiIvDsdLcwT6XiKSMcoE2aEiDGBJ3yaVU47hmFMjekfiVEdeRvc4Fym231KZiSiwwRavsfE4DG8uUvxr74XT3aF1bFERERERGSOUyExT4y9ehqAttJR6AV3ZABPyblrRBw6EWKjvZVkYS22nCIrYsoiYpppRs8cpP/Ev2KaacpW3kFh1QcxDJvV0UREREREZB5QITFPhFuHMFx2+mOjANiTE2R95MZzxhxp7ucP7d146++wIqIsIrHJfgJN+5gcPUFW4Sr8q+7G5SuxOpaIiIiIiMwjKiTmgXQsRex0iOx1pcROxHAkkjixU7F0zdSYE2cmibUfwu5J416hyzVkdpjpFEM9v2Cg/ScYNgf+hnvJr/jAOZcOiYiIiIiITIcKiXkg3j8BgLPIS+zwBEY6ij3HQ5nv7HX64+EEj/xzJ59wtmF6crBX1FsZVxaoaOg0vc17iY6fIqdkA+WrPoHTnW91LBERERERmadUSMwDqUgCALffw0Q6jMuMMLGmBNtvr9VvPBFiJBTnsuKTuOo2Y9jsVsaVBSadTjDY+SyDXc9hd2ZRue4+cko366wIERERERG5KCok5oHUxNlCouftN8EAzDSFxUumHh8YS1Br68URG9ftPmVGTY51EGjaSywcIK/8MspW3oHDlW11LBERERERWQBUSMwDybEYAO0vPA+5PrzhXjzla6ceHx5PssnRChg46zZblFIWknQqRvDk0wz3vIDDnU/Vxv9OTvHa93+iiIiIiIjINKmQmAdifRMYziTHPeXAOLZUhLwlVVOPj4aT3OBsw76kHpsvz7qgsiBMDLcQaN5HIjJIQeXVlC6/DbvDa3UsERERERFZYFRIzAORzgHaQu2MucZxJhKkcuzkewunHo+Pj1BrO41z2T0WppT5LpWYpP/EjxntfRmXr5SaLX9GVsEKq2OJiIiIiMgCpUJijjvzD/9ILFbNi/YuALJDR4j6nfi9FVNj/KHj2DBxLtftPuXCjAffoq/leyQTIYpqd1Cy9GZsdpfVsUREREREZAFTITFHmaZJ8Ac/IHT4BL0lZxcRrC1ZQqjvJcJra6j0Vk+NrYs1MWHLJb+szqq4Mk8lY+P0tX6f8eAh3NmVVG38DN7cGqtjiYiIiIjIIqBCYo6aOHyY4eeew5bTwBlzCJthp36FjzePwdZlH8Tn8AFgplPUm62cyd5E1W9vAyryfkzTZKzvNfrb9pNOxihZtovimg/plrEiIiIiIpIxKiTmqFggAECy/nJO9b5ARWkVfV1NACwtXz01LtrdRJYRZbhogyU5Zf5JRIcJNH+XiaG38ebVUbH6XtxZfqtjiYiIiIjIIqNCYo5KhUJEDB89p0eJ2OLUr15Oyws/I1rupTz3P+6wMdnyOmnTRsK/zsK0Mh+YZpqR0y8SPPkkJlC+8k4Kqq7B0Jk1IiIiIiJiARUSc9TpfpNDWdspsE0AsGx5NYe/H4Qrqs9Z0DLd2Uhrqoa8wnyroso8EAv3EWjex+ToSbIKG/A33I3LW2x1LBERERERWcRUSMxRRzvSRHO7aCGE3bBxJngEA6it34ztt99op8cHcYyc4q3kTVybq6mUdzLTKYa6/52BjgMYNhcVq/+APP8VGIZhdTQREREREVnkdBQ7B432jTHiHmGQQSpdJdRfsY6OtucxDbh0zU1T4xLtbwLwVmolH811WhVX5qhoqIfepr1EQ93klGzCv+ouHO48q2OJiIiIiIgAKiTmpJ59jUQcw1Qbpdz3f38Wu8/J3z/2HeLFHpYWrJwal2h/g0lXEWdCpRRkayrlrHQqwWDnTxk89XPszmwq191Pbtlmq2OJiIiIiIicQ0exc4yZTBMemyBMlGsv3Ybd5ySVTBI53Uu0IWfqcg0zmSDR+RY9WZfiGrPhcel2jQKTo+30Nu0lPtlHnv9yylfegd2ZZXUsERERERGRd1AhMccMv3qabjMIJmy64TIAgt1tmIkkZkX+VCGR7DkOiSjN9gayvSojFrt0Mkqw/WmGe36F01NA9abPkV20xupYIiIiIiIiv5cKiTlm9HAfXekgJU4fPp8PgN6TxwAorFsxNS5x8g2wO/nVUBX+IrclWWVumBhqItC8j0R0hMKqayldtgubw2N1LBERERERkfekQmIOSU0mGBsYZJQJrl1aP7X9VNthEjlO1ldsnNqWaH+T8YIGgp02PvXhIiviisVSiTB9bT9iLPAKLl8ZtVv/DF/+cqtjiYiIiIiITIsKiTkk1DTAoDkOwJptVwJgmiZnThwj6vdSn9sAQGq4l/TwGTorrwbgsoZcawKLZcb7DxFo/T6pxATFtTdSvPRmbHbdaUVEREREROYPFRJzyOjRIMPpMAAl1dUAjA0EiIfGiW/2U+2rAf7jdp8/6KpiSbELn1trSCwWydgYgdbvEwoexpNTRc2mz+HJqbI6loiIiIiIyHlTITFHmGmT/q4zHDXb8Rk2XC4XAL0njwJQsLQOp+3sN+Dxk28QNEroTxey564ayzJL5pimyVjgFfrafoSZjlO6/DaKqrdj2FRGiYiIiIjI/KRCYo5IjkdpTLUCcF1R4dT2nhNHSbts1NZsAMCMR0l0HeXN2GXc92E/K5b4LMkrmROPDBJo/hfCw01485dT0XAP7qxyq2OJiIiIiIhcFBUSc8RvXjjIaXOQvJSXutLSqe3dbYeJlntZlnv2DhujrYexmUlOedZy5xYtZrmQmWaakZ5f09/+FAZQXn8XBZVXY/z21q8iIiIiIiLzmQqJOeKVt14nn2zqx+zYvF4AouEQ432niV5SzNKsOuDs5Rop08WVN1yOw25YGVlmUSwcoLdpL5GxDrKK1uBftRuXVwWUiIiIiIgsHCok5oDYeIRQIkxNupb8ZDf23xYSve3HwQRnVRkFrkJM08Te3ciR5HJyfW6LU8tsMNMpBk/9nMGOn2JzuKlY8ynyyi/DMFQ+iYiIiIjIwqJCYg4Y6A1iAkbKQ0Gyf+oMid4TxzANqFq2DoD0YDf28CBH0tu4u9RjYWKZDZHxbgJNe4lO9JBbupny+rtwuHVLVxERERERWZhUSMwB3SdOA7BiZRnOgcRUIXGq7TDxYg91BfUAnPzNS5QCZRsvp6JIZ0gsFOlUnIHOnzJ06t9xOLOpXP9H5JZusjqWiIiIiIjIrFIhMQf85sgrAJR6UgB4qqpIJZMEu1qJNmRTl70MgGjbG/SYfu7a2WBZVplZ4ZETBJr3EZ/sJ7/iSspW3I7dmWV1LBERERERkVmnQmIOSMYTuHDgCbSSLivDVVFBX2cz6USChD+bJd4qzGiYJYkO3sy+nvVOu9WR5SKlklGCJ59k5PSvcXqKqN70x2QXqWgSEREREZHFQ4WExdJpk0QyQZWtlHjbWxRcfz2GYdB78hgAxcvqcdgcdLzxawpIk6zaYnFiuVihweMEmr9LMjZKYdV1lC7fhc2uS3BERERERGRxUSFhsZG+CVJmCofNjplMkrN5MwDdrW+RzHFSX74GgN7G3+A0vVyz41Ir48pFSMYn6G/bz1jfa7iyyqld9wV8+XVWxxIREREREbGECgmLJWJJUkYKeyKOPSsL7/LlAJxpP0a03MPS7OWYZprK8HFOexuoytI36fONaZqEgocItH6fVCJM8dIPU7z0Jmw2p9XRRERERERELKNCwmLxaIIUaeypJN76egy7nXhkktj4GPHVJdRl1dHf2kKeMUF35War48p5SsRG6Wv5HqGBI3hyqqnZ9Md4ciqtjiUiIiIiImI5FRIWi4QjADjSKUpuvxOAeHQSAG9WHjnOXI785iUaTIOaS66wLKecH9M0Ge39Df0nfoSZTlK6/HaKqq/HsGlBUhEREREREVAhYbno4TMAuBw2PJVnvzlPxKMAlGSVE42n8ATeIuCsZu3ScstyyvTFJwcItHyX8HALvvwV+Fffg9tXZnUsERERERGROUWFhMUSobNnQ7hy/mMqBkK9AJRmL+HRfzjCn9NDsO52S/LJ9JlmmuGeFwiefBrDsFG+ajcFS67CMGxWRxMREREREZlzVEhYLBIJA1CwvHpq24lTh87+wb6SrOBRbF6TFVdusyKeTFNsopfe5r1ExjrJLlqLv2E3Tk+h1bFERERERETmLBUSFkpFkkxEJwDwr2mY2n6muxnTgHCyno2OF8CXj718mVUx5T2Y6SSDXT9jsPNZbA43S9Z+mtyySzAMw+poIiIiIiIic5oKCQuN/LqLkBkBE8qW101tDwV6oSCLcMTgGscJnMuu1Gn/c1Bk/BS9TU8QmzhDbtlWyuvvxOHKsTqWiIiIiIjIvJCxQqKzs5MHH3yQ0dFR8vPz2bNnD7W1teeM+eY3v8kzzzyDzWbD6XTywAMPsG3bwr1UIdI1whlzCI+ZhcfjBiCdSmKeHsSoK8M7fJJsI4Jr+VaLk8p/lk7FGej4CUOnnsfhzqVqw38jp2SD1bFERERERETmlYwVEg8//DC7d+9m165dPP300zz00EM88cQT54xZv349n/70p/F6vbS0tHDPPfdw8OBBPB5PpmJmTGIkwljvCAPpMRoq6qe2nzl5DCOaxFNXSWnPMVLYcC7dbGFS+c/CI20EmvYSjwyQv+QqylZ8FLvDa3UsERERERGReScj1wEMDQ3R1NTEzp07Adi5cydNTU0MDw+fM27btm14vWcP7urr6zFNk9HR0UxEzLjQsQFOm0NgwMbLL5nafvDkzwBYUr2a6ujbnLItxfBkWRVTfiuVjOAMH+RU4/+LiUnN5j+houEelREiIiIiIiIXKCNnSAQCAcrKyrDb7QDY7XZKS0sJBAIUFr77nQieeuopqqurKS8vP699HT9+/KLzXqjGxsZpjTOiafJ+MUlPegBb2kmEMI2NjaRI8fbgWxQABcPZ+FNneNa+A6b5ujI7bIlunOGD2M1JEu51RNxbGe8MQ6fmZaGa7ntZ5jfN88KnOV4cNM8iIvPXnFzU8vXXX+drX/sa//iP/3jez127di1ut3sWUr23xsZGtmzZMq2xE8eDBMy36UkPkEsBl1569gyJlrEmjJcjGHY7q/uPAZCuvXzaryszKxmfoL/th4yNvI47q4Ix23Y2X3aT1bFklp3Pe1nmL83zwqc5XhysmudYLGbpl2AiIgtFRgoJv99Pf38/qVQKu91OKpUiGAzi9/vfMfbw4cN84Qtf4PHHH6euru5dXm3+i/aMM8AYSSOFP6toavsr3T/FORIjxwmOnkP0pou55IrVFiZdnEzTZLz/Tfpaf0AqGaGkbifFtTdy6PARq6OJiIiIiIgsGBkpJIqKimhoaODAgQPs2rWLAwcO0NDQ8I7LNY4ePcoDDzzA17/+ddasWZOJaJZIjsc4wzCYsLKuCoAzk6dpnGxh6akwpesu5f/zfIwjAQdP+H0Wp11cEtERAi3fY2LwKJ7cWmpW34sne4nVsURERERERBacjF2y8cgjj/Dggw/y+OOPk5uby549ewC47777+PznP8+6det49NFHiUajPPTQQ1PP++u//mvq6+t/38vOS7HJKIeTJ3Gms1hxyXIAXjn9HGXPngHAV7OVF9/w8ZEri3E5M7Lu6KJnmiajvQfpb/sxppmibMXHKKy+DsPQ719ERERERGQ2ZKyQWLZsGfv373/H9m9/+9tTf/7xj3+cqTiWOjMcBCA/YadgaQUALwcOUtkdJie/iED2NcAAH91WYmHKxSM+OUBv8z4mR1rxFdRT0XAPLp9+9yIiIiIiIrNpTi5qudCNhUMANKQTGA4HY5NB3N1hAHZ+9qv8z5+GqCx2U5znsjLmgmeaaYa7f0mw/WkMw46/4W7yK67CMAyro4mIiIiIiCx4KiQyzDRNAtFBAGqzYwA0tz1D8a/6sDkceEtX0BFo4ebLi97rZeQiRSfO0Nu0l+h4F9nF6/Gv+gROT4HVsURERERERBYNFRIZ1vRaFyc5DUBWWRkAAx1vYphQs/FyDp+cxAS2b9bB8Www00kGOp9lsOs57A4fS9b+IbllW3RWhIiIiIiISIapkMiwvo6zZ0esTvhwFhSQnhihv//sYpart97AUy3jFGQ7WLFEd9eYaZGxTnqb9hIL95JXfillKz+Ow5VtdSwREREREZFFSYVEhoVDowAUJVI4CgqIN/2a4GgMN1C5chONz57hijW52Gz6xn6mpFNxgu3/xnD3L3C486na+FlyitdZHUtERERERGRRUyGRYcHRs2dDFMXjeGpqiL32D9gHYphOG+1DNiaiKS6pz7U45cIRHm6lt3kvicggBUuupnTFbdgdXqtjiYiIiIiILHoqJDJsJDxEiZFHbqIfuxHnxKk2HKNx8nZu4/WWCew22Lwix+qY814qGaH/xI8ZPXMQl7eEmi1/SlbBSqtjiYiIiIiIyG+pkMig0z2nGUkOUWeUY6YipHre4EgsTdplo+6yazjwb5P4C91keexWR53XQgNHCbR8l2RsnKKaGyipuwWbXbdQFRERERERmUtUSGTQsTdaAai2lWLPaiXwmycZGoiSLPOyvmgLfz/Qx/qlWRannL+S8RB9rT9gvP9N3NlLqNrwGby5NVbHEhERERERkXehQiKDxkcmAChLZ+FxDdNlzwHCZH1wK24zn4nIGeoqtL7B+TJNk/G+1+lr+yHpZIySulsprv0Qhk3/e4uIiIiIiMxVOmLLoEDX2Vt+uuLDeLIm6fDkk3YYXLHpVo52nC0rKks8VkacdxLRYQIt32Ni8BjevKVUNNyLO7vC6lgiIiIiIiLyPlRIZIiZNgnFB3DY7bic4zg9SYZPnWZyeS7r8rfwB3/fSmGOg83Ls62OOi+YZpqRMwcJnvhXTDNN2co7KKz6IIZhszqaiIiIiIiITIMKiQyJTMTBSGEATvcgKVcWZnQce0E5Pf0JQpEUf/LRSnJ8mpL3E5vsJ9C0j8nRE2QVrsK/6m5cvhKrY4mIiIiIiMh50NFvhgx0j5I2ktQapWQVtzFuLwcC5OQX8VZ7CICNdTo74r2Y6RRD3b9goOMnGDYH/oZ7ya/4AIZhWB1NREREREREzpMKiQyJRZMkjQTOVJKTPSc4GAwD4Cvz83pziNoyD2WFbotTzl3R0Gl6m/cSHT9FTskGyld9Aqc73+pYIiIiIiIicoFUSGTIWHc/KdJ4knFawhFycvM5tcZFWUUNbz8X5q5rS62OOCel0wkGO59lsOs57M4sKtfdR07pZp0VISIiIiIiMs+pkMiQlra3AKjwpXllOE3N6uUc2TzB+FA+pgmXNeRZnHDumRztoLf5CeLhPvL8l1O24mM4XLqsRUREREREZCFQIZEhoXCIHHyk3KcACBakATjV6SPHa2fFEq+V8eaUdCpG8OTTDPe8gMOdT9XG/05O8VqrY4mIiIiIiMgMUiGRAaPBCaKpOC4cHO1uBaBpxQRLvNWc6ithSbETm02XIABMDDUTaN5HIjpEQeU1lC6/DbvDY3UsERERERERmWE2qwMsBk2vnmKMELlpg2giSXC7nzWFG3iw4SuEJyEvS71QKjFJb9MTdB/+GobNTs2WP8O/6hMqI0RERERERBYoHQlnQHvXCRIkqTEdHANiJR6uKrkGA4OJaIrcRV5IjAffoq/leyQTIYpqd1Cy9GZsdpfVsURERERERGQWLe4j4QyJDg5iAK7kEADunBxqvMt58H+1E0uYVJcuzrMAkrFx+lq/z3jwEJ7sKqo2fhZvbrXVsURERERERCQDVEjMMjNtEklOkI2XoeRh0g6Dayo/REdPmqMdYT66rZjbriy2OmZGmabJWOA1+tp+iJmKU7JsF8U1H8Kw2a2OJiIiIiIiIhmiQmKWBTpGCJuT5BoexuwxUlkOLiu5ikPHEwBs31yI3b54lvJIRIfpbf4u4aG38ebVUbH6D3BnlVsdS0RERERERDJMhcQsMk2TXzzza0LGBFVmKYPpGLYsF0XuIo51dGMAhTlOq2NmhGmmGTn9IsGTT2IC5SvvpKDqGgxj8ZQxIiIiIiIi8h9USMyivr4+mvveAMCf9tAXilFYV0lwNM4LR0bZWp+zKBa0jIX76G3eR2T0JFmFDfgb7sblXVyXqYiIiIiIiMi5Fv7RsIV6u/sB2G5soIgQhgnr1m7jZ28MkUyZ3PfhCosTzi4znWKo+98Z6DiAYXNRsfqT5PkvxzAMq6OJiIiIiIiIxVRIzKLTJ88WErn2bNochzBtsPHSj/Lcv47gc9uoWsB314iGeuht2ks01E1O6Sb89XfhcOdZHUtERERERETmCBUSs2giFAbAhYOBSA9FOT7c2fn0j/RTUeS2ON3sSKcSDHb+lMFTP8fuzKZy3f3klm22OpaIiIiIiIjMMSokZlEoNAaA0zSZHDrNipXLAYgl0hTnLbzFLCdHT9LbtJf4ZD95/isoX/kx7M4sq2OJiIiIiIjIHKRCYhb1jfbgMh2knEkM02Tpyo3E4ilGJpKsWOKzOt6MSSejBNufZrjnVzg9BVRv+jzZRautjiUiIiIiIiJzmAqJWRIMBombUVayhJg5iWkzWLJ+G99+JkAklubq9QtjPYWJoSYCzftIREcorLqW0mW7sDkW7toYIiIiIiIiMjNUSMyS9vZ2ANbaaxlPn8LIcZAsXMbzh7q4pD6HD6zJtzjhxUklwvS1/YixwCu4fGXUbv0zfPnLrY4lIiIiIiIi84QKiVkSCoXAhHxbFqcT/dh9Lp49EiGWMPnY1aVWx7so48HDBFq+RyoxQXHtTRQv/TA2+8JbE0NERERERERmjwqJWTIyMITDdGEYBtGJIVyVPp57Y5iKIhfrls7PhR6TsTECrd8nFDyMJ6eKmk2fw5NTZXUsERERERERmYdUSMyS3uMtODl7a89oYoy4t4RAb5z7b67AMAyL050f0zQZC7xCX9uPMNNxSpffRlH1dgyb3epoIiIiIiIiMk+pkJgF8f5+xtJpimw+0qQJJQboH76e6zbms+sDRVbHOy/xyCCB5u8SHm7Gm7+cioZ7cGeVWx1LRERERERE5jkVErOg+3//MwmbSa7Nw7Cth4gzQkdyKx/dWIDNZrM63rSYZpqRnl/T3/4UBlBefxcFlVdjGPMjv4iIiIiIiMxtKiRmQXPQjulLk2X3EI4OkSpwM54sZKl/ftwOMxYO0Nu0l8hYB1lFa6houBunp9DqWCIiIiIiIrKAqJCYYWY6TQ9hMKGWEoKRLhKr8ykfdVGU67I63nsy0ykGT/2MwY5nsDncVKz5FHnll827NS9ERERERERk7lMhMcOiPT2MOk0KjVxybW7ejL5FpKySKypyrY72niLj3fQ2PUFs4jS5ZVsoX3knDvfcziwiIiIiIiLzlwqJGTbZ20vYEafWVk6v8wij7jEi6XVsWJltdbR3lU7FGej4KUPd/47DmU3l+v+L3NKNVscSERERERGRBU6FxAzb/5NfYxomZUYBvbFDpF12xoJLKbzEaXW0dwiPnCDQvI/4ZD/5FVdStuJ27M4sq2OJiIiIiIjIIqBCYoaNpJJghypfOW8F27Fl2+nr2kSOz251tCmpZITgyacYOf1rnN5iqjf/CdmFq6yOJSIiIiIiIouICokZZJomCTPFEqMIp9cBYzF8VeUwYMN38LD6AAAW3ElEQVTnnhuFRGjwOIHm75KMjVJYdR2ly3dhs7utjiUiIiIiIiKLjAqJGRQfHiZmj5NnZBOyjwKQ7anAZoMsr7WFRDI+QX/bfsb6XsOd5ady/Rfw5dVZmklEREREREQWLxUSM6jv5ElMw6TIyKY/0QWAaW6gINuB3WbNrTNN02Q82Ehfyw9IJcMUL/0wxUtvwmabe2taiIiIiIiIyOKhQmIGtbeeAKDQyOX0eAum28aJyEpK8lyW5EnERulr+R6hgSN4cmuoafhjPDmVlmQRERERERER+c9USMyg7kAfdtNBETl0Dg6TU5XNS2PZXFOb2TUaTNNktPc39J/4EWY6SemK2ymquh7DNjfWsRARERERERFRITGDJiMxHKYLwzCIpyZZVlpBvN1gabknYxnikwP0Nu9jcqQVX/4KKlbfi8tXmrH9i4iIiIiIiEyHCokZNBaO4bVnAxAxxynPvhKAunLvrO/bNNMM97xA8OTTGIYN/6rd5C+5CsOwzfq+RURERERERM6XComZkkoRs5kUGWfXi5j0jjBILQA1s1xIRCd6CTTtJTLeSXbxOvyrduP0FMzqPkVEREREREQuhgqJGRIdGiJuT1Fmy6Z3solsl42WWDk+t428rNlZu8FMJxns+hkDnc9gd3hZsva/klu2FcOw5o4eIiIiIiIiItOlQmKGnOnsAqDGKGUkeYxLSnz8w1AeZYWOWSkIImNd9DbvJTZxhtyySyiv/zgOV86M70dERERERERkNqiQmCGRiTAAhUYOAcckFUVL6epKcvf1RTO6n3QqzkDHTxg69TwOdx5VGz5DTsn6Gd2HiIiIiIiIyGxTITFDUpE4AAYGLleUEV8NAFeuyZuxfYSHWwk07yMeGSB/yTbKVtyO3TH7C2aKiIiIiIiIzDQVEjMkGTdxYMcwDDZ5hul0Xw5AXvbF/4pTyQjBE//KyJmXcHpLqNn8AFmF9Rf9uiIiIiIiIiJWUSExQ+IpB16bi670UTbZgvQ5KgHI9lzcgpahgaMEWv6FZGyMopobKKm7BZvdNRORRURERERERCyjQmKGxEjhMVwM5hzGSNv5eUcOBdkGLqftgl4vGQ/R1/pDxvvfwJ1VQdX6P8Kbt3SGU4uIiIiIiIhYQ4XEDImTJB8vlU4bMXclp3pS/NHOivN+HdM0Ge9/k77WH5BKRiipu4Xi2h0YNk2ViIiIiIiILBw6yp0BpmkSN5K4cFCRGOO0eyMOu8GHthSc1+skoiMEWv6FicFjeHOX4l99L57s8y81REREREREROY6FRIzYHJggBhJsnCSlwjzm5ifiiIXPs/0fr2mmWb0zMv0n/gxppmibMXHKKy+DsO4sMs9REREREREROY6FRIz4HTj22BAIj2OYRgcGi1h5VrftJ4bnwzS27yPyZE2fAX1VDTcg8tXMsuJRURERERERKylQmIGBJuHAAgbA5gYdCbK+NTWwvd8jplOMdTzCwbaf4Jh2PE33EN+xZUYhpGJyCIiIiIiIiKWUiExAybHkgCM28/Qmypm85pS1i3N/r3joxNn6G16guj4KbKL1+Nf9QmcnvNbb0JERERERERkPlMhMQPiybOFhNMeYyxrOX9+R9W7jkunEwx2Psdg17PYHVksWfeH5JZu0VkRIiIiIiIisuiokJgBCc4WEl5HgopVq/C47O8YMznWSaBpL7FwL3nll1G28g4crt9/FoWIiIiIiIjIQqZCYgbEiAGQbU9RuXrNOY+lUzGC7f/GcPcvcbjzqdr4WXKK11kRU0RERERERGTOUCExA+JmArtpUOWyYy9fNrV9YriFQPM+EpFBCiqvpnT5bdgdXguTioiIiIiIiMwNKiRmQIwkTgxWL9mKzZNNKjFJ/4kfM9r7Mi5fKTVb/oysghVWxxQRERERERGZMzJWSHR2dvLggw8yOjpKfn4+e/bsoba29pwxqVSKr371q7z00ksYhsH999/PHXfckamIFyxOEqdpJ3fXnxMKvkWg5XskEyGKanZQUnczNrvL6ogiIiIiIiIic0rGComHH36Y3bt3s2vXLp5++mkeeughnnjiiXPG/OQnP6G7u5uf//znjI6O8pGPfIQrrriCysrKTMU8b/HhQeIkyXKZBE98l1CwEXd2JVUbP4M3t8bqeCIiIiIiIiJzki0TOxkaGqKpqYmdO3cCsHPnTpqamhgeHj5n3DPPPMMdd9yBzWajsLCQ7du389xzz2Ui4gUbOtpKUdkoV18+wMTAEUqW3Urdpf+PyggRERERERGR95CRMyQCgQBlZWXY7Wdvh2m32yktLSUQCFBYWHjOuIqKiqmf/X4/fX19mYh4wZo7W7hkzShjY04arvkS7iy/1ZFERERERERE5rwFt6jl8ePHM7q/vPUbaf3NEAUNGzje0gv0ZnT/klmNjY1WR5AM0DwvDprnhU9zvDhonkVE5q+MFBJ+v5/+/n5SqRR2u51UKkUwGMTv979jXG9vL+vXrwfeecbEdKxduxa32z1j2adjy5YtNDY2smXLlozuVzJLc7w4aJ4XB83zwqc5XhysmudYLJbxL8FERBaijKwhUVRURENDAwcOHADgwIEDNDQ0nHO5BsCNN97I/v37SafTDA8P8/zzz7Njx45MRBQRERERERGRDMpIIQHwyCOPsG/fPnbs2MG+fft49NFHAbjvvvs4duwYALt27aKyspIPfehDfPzjH+ezn/0sVVVVmYooIiIiIiIiIhmSsTUkli1bxv79+9+x/dvf/vbUn+12+1RRISIiIiIiIiILV8bOkBARERERERER+R0VEiIiIiIiIiKScSokRERERERERCTjVEiIiIiIiIiISMapkBARERERERGRjFMhISIiIiIiIiIZp0JCRERERERERDJOhYSIiIiIiIiIZJwKCRERERERERHJOBUSIiIiIiIiIpJxKiREREREREREJONUSIiIiIiIiIhIxqmQEBEREREREZGMc1gdYKaYpglAPB63LEMsFrNs35IZmuPFQfO8OGieFz7N8eJgxTz/7vPm7z5/iojIhTHMBfIvaSgUoq2tzeoYIiIiIrJIrFy5kpycHKtjiIjMWwumkEin04TDYZxOJ4ZhWB1HRERERBYo0zRJJBJkZWVhs+kKaBGRC7VgCgkRERERERERmT9U6YqIiIiIiIhIxqmQEBEREREREZGMUyEhIiIiIiIiIhmnQkJEREREREREMk6FhIiIiIiIiIhknAoJEREREREREck4FRIiIiIiIiIiknEqJEREREREREQk41RInIfOzk7uvPNOduzYwZ133klXV9c7xqRSKR599FG2b9/ODTfcwP79+zMfVC7YdOb4m9/8JjfffDO33HILt99+Oy+99FLmg8pFmc48/05HRwcbNmxgz549mQsoM2K68/zMM89wyy23sHPnTm655RYGBwczG1QuynTmeWhoiPvvv59bbrmFm266iUceeYRkMpn5sHJB9uzZw3XXXUd9fT1tbW3vOkafv0RE5ilTpu3ee+81n3rqKdM0TfOpp54y77333neMefLJJ81Pf/rTZiqVMoeGhsxt27aZPT09mY4qF2g6c/ziiy+ak5OTpmmaZnNzs7llyxYzEolkNKdcnOnMs2maZjKZNO+55x7zT//0T82/+qu/ymREmQHTmeejR4+aN910kxkMBk3TNM3x8XEzGo1mNKdcnOnM81e/+tWp93A8Hjc/9rGPmT/96U8zmlMu3BtvvGH29vaaH/zgB83W1tZ3HaPPXyIi85POkJimoaEhmpqa2LlzJwA7d+6kqamJ4eHhc8Y988wz3HHHHdhsNgoLC9m+fTvPPfecFZHlPE13jrdt24bX6wWgvr4e0zQZHR3NeF65MNOdZ4BvfetbXHvttdTW1mY4pVys6c7zP/3TP/HpT3+akpISAHJycnC73RnPKxdmuvNsGAbhcJh0Ok08HieRSFBWVmZFZLkAW7duxe/3v+cYff4SEZmfVEhMUyAQoKysDLvdDoDdbqe0tJRAIPCOcRUVFVM/+/1++vr6MppVLsx05/g/e+qpp6iurqa8vDxTMeUiTXeeW1paOHjwIJ/61KcsSCkXa7rz3N7eTk9PD3fffTe33XYbjz/+OKZpWhFZLsB05/kzn/kMnZ2dXHXVVVP/bdmyxYrIMkv0+UtEZH5SISFygV5//XW+9rWv8bd/+7dWR5EZlkgk+MpXvsKjjz46daAjC1MqlaK1tZXvfOc77N27lxdffJGnn37a6lgyw5577jnq6+s5ePAgL774Im+++aa+PRcREZkDVEhMk9/vp7+/n1QqBZz9EBsMBt9xCqHf76e3t3fq50AgoG/P54npzjHA4cOH+cIXvsA3v/lN6urqMh1VLsJ05nlgYIDu7m7uv/9+rrvuOv75n/+ZH/7wh3zlK1+xKracp+m+nysqKrjxxhtxuVxkZ2dz/fXXc/ToUSsiywWY7jzv27ePW2+9FZvNRk5ODtdddx2vvfaaFZFllujzl4jI/KRCYpqKiopoaGjgwIEDABw4cICGhgYKCwvPGXfjjTeyf/9+0uk0w8PDPP/88+zYscOKyHKepjvHR48e5YEHHuDrX/86a9assSKqXITpzHNFRQWvvfYav/zlL/nlL3/JJz/5ST7+8Y/z2GOPWRVbztN03887d+7k4MGDmKZJIpHg1VdfZdWqVVZElgsw3XmurKzkxRdfBCAej/PKK6+wYsWKjOeV2aPPXyIi85Nh6mLZaWtvb+fBBx9kfHyc3Nxc9uzZQ11dHffddx+f//znWbfu/2/v/mOqqv84jj+9nsSQKWBQqK3N1pwo6tVzuS7vLWdbN+0m1JVdnbLC1ghz1D8NgfzR7W7+mI3M2lgp+kc/UDdHhgTmBupUCOct+rkp80eThIi09GrivXz/aN19KagLrkt9v6/Hdv84n/M5n/M652wM3nzO52YQCoXw+XwcPXoUgGeeeQav1zvEySVa0Txjj8fDhQsXei2ItmnTJiZNmjSEyWUgonnO/23r1q0Eg0GKioqGKLEMRjTPORwOs3HjRg4fPozFYsHhcFBUVITFonr9v0U0z/n8+fOsXbuWzs5OQqEQdrud0tJSDMMY6vgSBb/fz4EDB+js7CQpKYnExET279+v379ERP4HqCAhIiIiIiIiIjGnfwGJiIiIiIiISMypICEiIiIiIiIiMaeChIiIiIiIiIjEnAoSIiIiIiIiIhJzKkiIiIiIiIiISMypICEi8jfJzc1lz549Qx3jT+3bt4/ly5f3u//EiRO4XK4YJhIRERGR/xcqSIiIRGHevHlMmzYNq9Ua+bS3t8c8R25uLhkZGVitVux2OytXrqSjo2PQ4y1cuJCKiorI9qRJkzh37lxk2zRN6urqbilzX7Zu3cqUKVOwWq2YpsnixYsJBAJRH//7nCIiIiLy76OChIhIlMrLywkEApHPnXfeOSQ51qxZQyAQoK6ujp9++on169cPSY5bNX/+fAKBAI2Njdjtdp5//vmhjiQiIiIiMaSChIjIIF2+fJn8/Hxmz56NzWYjPz+fixcv9tn33LlzLFu2jFmzZmG323nhhRci+1pbW8nLyyMzMxOXy0VNTU1U509MTMTlcnHq1CkATp48icfjYdasWXg8Hk6ePBnpu3fvXh566CGsVivz5s1j3759kfYlS5YAsHTpUgCysrKwWq3U1NTQ1NTEAw88AMBbb71FYWFhrwx+vx+/3w/Azz//TElJCQ6HA6fTSVlZGaFQ6C+vwzAMHnvsMdrb2+nq6gKgpaUFr9eLaZo4HA58Ph83btzoNydAfX09WVlZkRkX33zzTVT3UURERESGhgoSIiKDFA6HeeKJJ6ivr6e+vp64uDh8Pl+ffbds2cKcOXNobm7m8OHDLFu2DIBgMMjy5ctxu90cO3aMsrIyXn75ZU6fPv2X5+/q6qKuro7Jkydz6dIl8vPzyc3Npampiby8PPLz8/nxxx8JBoP4/X7efvttAoEAlZWVTJ48+Q/jvfvuuwB88MEHBAIBFixY0Gv/o48+yqFDh7hy5QoAoVCI2tpa3G43AKtWrcIwDA4cOEBVVRVHjx6Nag2NGzduUFVVRWJiIqNHjwbAYrFQXFxMY2MjlZWVHD9+nPfee6/fnF999RUlJSX4fD6amprwer2sWLEiUsQQERERkX8eFSRERKL03HPPYZompmmyYsUKkpKScLlc3H777SQkJFBQUEBzc3OfxxqGQVtbGx0dHcTFxWGaJgANDQ2MHz8ej8eDYRikp6fjcrmora3tN4ff78c0TbKyskhJSaG4uJiGhgbuuecesrOzMQwDt9vNxIkTqa+vB379A//UqVNcv36d1NRU7rvvvgFf//jx40lPT+fgwYMANDY2MnLkSGbMmEFnZyeHDh2ipKSE+Ph4xo4dy1NPPcX+/fv7Ha+2thbTNJk+fTp79uzh9ddfxzAMAKZOncqMGTMwDIMJEybg9Xr7vbcAu3btwuv1Mn36dIYPH87jjz/Obbfdxqeffjrg6xQRERGR2DCGOoCIyL/Fm2++yf333x/ZvnbtGuvXr+fIkSNcvnwZgKtXrxIKhRg+fHivY1988UW2bNnCokWLGDNmDHl5eSxatIgLFy7Q0tISKVDArzMPFi5c2G+Ol156iZycnF5tHR0djBs3rlfbuHHjaG9vJz4+nrKyMioqKigtLWXmzJkUFRVx7733DvgeuN1uqquryc7Oprq6OjI7oq2tjZs3b+JwOCJ9w+EwaWlp/Y71yCOPsHnzZrq6uigsLOTLL7/EbrcDcObMGTZs2MAXX3zBtWvXCIVCTJkypd+x2traqKqq4p133om0dXd339KCnyIiIiLy91JBQkRkkCoqKjhz5gy7d+8mJSWFr7/+muzsbHp6ev7QNyUlJbLWwokTJ8jLy8Nms5GWlobNZmPHjh23lCU1NZW2trZebd999x1OpxMAp9OJ0+nk+vXrvPbaa6xevTryCsRAzJ8/n40bN3Lx4kU+/vhjdu3aBcBdd93FiBEjaGxsjMxyiFZycjI+nw+Px4Pb7SY1NZV169aRnp7Oq6++SkJCAjt37vzTb/tIS0vj2WefpaCgYMDXJCIiIiJDQ69siIgM0tWrV4mLi2P06NFcunSJN954o9++H330UWTByzFjxjBs2DAsFgtz587l7NmzVFVV0d3dTXd3Ny0tLbS2tg4oy4MPPsjZs2f58MMPuXnzJjU1NZw+fZq5c+fS2dnJwYMHCQaDjBgxgvj4eCyWvn/833HHHXz77bf9nic5OZnMzEyKi4uZMGFCZJZFamoqc+bMYcOGDVy5coVwOMz58+f55JNPoso/ceJEnE4n27ZtA369t6NGjWLUqFG0trby/vvv/2nOnJwcKisr+eyzz+jp6SEYDNLQ0BBZ70JERERE/nlUkBARGaQnn3ySX375hdmzZ+P1eiOzEfry+eefk5OTg9VqpaCggNLSUu6++24SEhLYvn07NTU1OJ1OHA4HmzdvHvBijElJSZSXl7Njxw7sdjvbtm2jvLyc5ORkwuEwO3fuxOl0kpmZSXNzM+vWretznJUrV7Jq1SpM0+z32z5+W4Dzt9c1frNp0ya6u7tZsGABNpuNwsJCvv/++6iv4emnn2b37t388MMPFBUVUV1dzcyZM1m9evUfFtj8fc6MjAxeeeUVfD4fNpuNhx9+mL1790Z9bhERERGJvWE9fc0tFhERERERERH5G2mGhIiIiIiIiIjEnAoSIiIiIiIiIhJzKkiIiIiIiIiISMypICEiIiIiIiIiMaeChIiIiIiIiIjEnAoSIiIiIiIiIhJzKkiIiIiIiIiISMypICEiIiIiIiIiMfcfo0z02nDlOOIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RscUpt2p4y59",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        },
        "outputId": "66673a45-c6f6-417a-edf7-fa796aecddc9"
      },
      "source": [
        "result_df.sort_values(by='auc',ascending=False)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Dataset</th>\n",
              "      <th>log_loss</th>\n",
              "      <th>auc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>XGBoost Classifer(top 250)</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.132907</td>\n",
              "      <td>0.846664</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>XGBoost Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.133102</td>\n",
              "      <td>0.845507</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>LightGBM Classifer</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.132863</td>\n",
              "      <td>0.844955</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>LightGBM Classifer(top 250)</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.133097</td>\n",
              "      <td>0.844849</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Random Forest Classifier (top 250)</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137054</td>\n",
              "      <td>0.834184</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.137490</td>\n",
              "      <td>0.832512</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logisic Regression</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.542129</td>\n",
              "      <td>0.812101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>Normal</td>\n",
              "      <td>0.148082</td>\n",
              "      <td>0.779135</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                Model Dataset  log_loss       auc\n",
              "6          XGBoost Classifer(top 250)  Normal  0.132907  0.846664\n",
              "3                   XGBoost Classifer  Normal  0.133102  0.845507\n",
              "4                  LightGBM Classifer  Normal  0.132863  0.844955\n",
              "7         LightGBM Classifer(top 250)  Normal  0.133097  0.844849\n",
              "5  Random Forest Classifier (top 250)  Normal  0.137054  0.834184\n",
              "2            Random Forest Classifier  Normal  0.137490  0.832512\n",
              "0                  Logisic Regression  Normal  0.542129  0.812101\n",
              "1            Decision Tree Classifier  Normal  0.148082  0.779135"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "041Z6c2qmN9a",
        "colab_type": "text"
      },
      "source": [
        "# Saving all the datasets created"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EH0NANEcmKmJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import joblib\n",
        "filename = 'normal_train.pkl'\n",
        "X_train.to_pickle(os.path.join(file_path,'best_models',filename))\n",
        "\n",
        "filename = 'normal_val.pkl'\n",
        "X_val.to_pickle(os.path.join(file_path,'best_models',filename))\n",
        "\n",
        "filename = 'normal_test.pkl'\n",
        "X_test.to_pickle(os.path.join(file_path,'best_models',filename))\n",
        "\n",
        "filename = 'normal_train_250.pkl'\n",
        "X_train_200.to_pickle(os.path.join(file_path,'best_models',filename))\n",
        "\n",
        "filename = 'normal_val_250.pkl'\n",
        "X_val_200.to_pickle(os.path.join(file_path,'best_models',filename))\n",
        "\n",
        "filename = 'normal_test_250.pkl'\n",
        "X_test_200.to_pickle(os.path.join(file_path,'best_models',filename))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AW7MYD5OquH2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "filepath = '/content/drive/My Drive/Santander_Customer_Satisfaction/best_models'\n",
        "y_train = joblib.load(os.path.join(filepath,'y_train.pkl'))\n",
        "y_val = joblib.load(os.path.join(filepath,'y_val.pkl'))"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}