DAI_ASSIGNMENT_1_Q2_800to1300_SVM.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DAI_ASSIGNMENT_1_Q2_SVM.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X8Y6BjULYyWH"
      },
      "source": [
        "Q2. Evaluation Metrics for Bias Detection. ​[50 Marks]\n",
        "Dataset : Fashion MNIST\n",
        "Motive of this question is to use metrics to detect and evaluate the bias in the machine learning model.\n",
        "● AIM: ​Perform a 2-class classification between ​Pullover​ and ​Coat\n",
        "● Training data split:​ Take all the training samples corresponding to Pullover, and only 500\n",
        "samples for Coat.\n",
        "● Testing data split:​ Take all the testing samples corresponding to Pullover and Coat.\n",
        " \n",
        "● Cross-validation:​ Repeat each of the experiments 2 more times by taking different 500 samples of Coat.\n",
        "● Algorithms:​ Perform classification using:\n",
        "○ A linear SVM. (You may use sklearn library)\n",
        "○ A 5 layer neural network with architecture: [ 128 -- 128 -- 128 -- 64 -- 1 ]\n",
        "(These numbers denote the number of nodes in each layer)\n",
        "● Perform a 2 class classification and report the performance as follows:\n",
        "○ Report :\n",
        "■ Testing performance (mean ± std) [5 Marks]\n",
        "■ Comparison of Testing Accuracy and show Confusion matrix for the\n",
        "classification performed: SVM vs Neural Network [5 Marks]\n",
        "○ ROC curve (just 1 ROC for each algorithm with error bars, as shown in below graph) and\n",
        "report EER (Equal Error Rate). [MUST be done from scratch] [15 Marks]\n",
        "○ Draw the Precision-Recall curve. [MUST be done from scratch] [5 Marks]\n",
        "○ Of the two curves stated above, which is more reliable for biased/imbalanced data? Why?\n",
        "Draw inferences from the observed results on why one approach performs worse/better than the other. [20 Marks]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o2QS8RWWY13f"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8_Voj7aDW-Wx"
      },
      "source": [
        "# Pullover = 2, coat = 4"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yNxN-SY-ZMUz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a59f3313-ac76-43e4-ac76-70cff94b7577"
      },
      "source": [
        "# Ref : https://www.tensorflow.org/tutorials/keras/classification\n",
        "# TensorFlow and tf.keras\n",
        "import tensorflow as tf\n",
        "\n",
        "# Helper libraries\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "fashion_mnist = tf.keras.datasets.fashion_mnist\n",
        "\n",
        "(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "8192/5148 [===============================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OYMmd4xXew4B",
        "outputId": "c57f53e9-e346-4ea1-de31-c7eb82c1b4ea"
      },
      "source": [
        "# For pullover\n",
        "idx = (train_labels == 2).reshape(train_images.shape[0])\n",
        "Xtrain_pullover = train_images[idx]\n",
        "Ytrain_pullover = train_labels[idx]\n",
        "print(Xtrain_pullover.shape)\n",
        "print(Ytrain_pullover.shape)\n",
        "idx_test = (test_labels == 2).reshape(test_images.shape[0])\n",
        "Xtest_pullover = test_images[idx_test]\n",
        "Ytest_pullover = test_labels[idx_test]\n",
        "print(Xtest_pullover.shape)\n",
        "print(Ytest_pullover.shape)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(6000, 28, 28)\n",
            "(6000,)\n",
            "(1000, 28, 28)\n",
            "(1000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MBGFHPxdmcrf",
        "outputId": "efe5322c-ce8f-48f4-ca72-ee78d7c5a9b5"
      },
      "source": [
        "# For coat\n",
        "idx = (train_labels == 4).reshape(train_images.shape[0])\n",
        "Xtrain_coat = train_images[idx]\n",
        "Ytrain_coat = train_labels[idx]\n",
        "print(Xtrain_coat.shape)\n",
        "print(Ytrain_coat.shape)\n",
        "idx_test = (test_labels == 4).reshape(test_images.shape[0])\n",
        "Xtest_coat = test_images[idx_test]\n",
        "Ytest_coat = test_labels[idx_test]\n",
        "print(Xtest_coat.shape)\n",
        "print(Ytest_coat.shape)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(6000, 28, 28)\n",
            "(6000,)\n",
            "(1000, 28, 28)\n",
            "(1000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tFRIjVaOsxZI"
      },
      "source": [
        "Training data split:​ Take all the training samples corresponding to Pullover, and only 500 samples for Coat. ● Testing data split:​ Take all the testing samples corresponding to Pullover and Coat."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VL6zMyXepEHj",
        "outputId": "299ac10b-5bf5-4e23-87aa-537cfcac158a"
      },
      "source": [
        "# X_train, Y_train, X_test, Y_test\n",
        "# np.array(Xtrain_coat[0:500]).shape\n",
        "X_train = np.concatenate((Xtrain_pullover, Xtrain_coat[800:1300]))\n",
        "# To take different samples just change teh indices from 500:1000 and then 1500:2000\n",
        "print(X_train.shape) # Expecting 6500\n",
        "Y_train = np.concatenate((Ytrain_pullover, Ytrain_coat[800:1300]))\n",
        "# To take different samples just change teh indices from 500:1000 and then 1500:2000\n",
        "print(Y_train.shape) # Expecting 6500\n",
        "X_test = np.concatenate((Xtest_pullover, Xtest_coat))\n",
        "print(X_test.shape)\n",
        "Y_test = np.concatenate((Ytest_pullover, Ytest_coat))\n",
        "print(Y_test.shape)\n"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(6500, 28, 28)\n",
            "(6500,)\n",
            "(2000, 28, 28)\n",
            "(2000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vI4qmV_fuePb"
      },
      "source": [
        "Algorithms:​ Perform classification using: ○ A linear SVM. (You may use sklearn library) ○ A 5 layer neural network with architecture: [ 128 -- 128 -- 128 -- 64 -- 1 ] (These numbers denote the number of nodes in each layer) "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iJBogGf6vqB2"
      },
      "source": [
        "# Linear SVM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1cuP04v5s8Kr",
        "outputId": "0d0ddd7b-6b31-4501-f9f4-e231de12c821"
      },
      "source": [
        "# xtrain = X_train.reshape(-1, 784)\n",
        "# xtrain.shape\n",
        "# from sklearn import svm\n",
        "# # Create a classifier: a support vector classifier\n",
        "# classifier = svm.SVC(gamma=0.001)\n",
        "# #fit to the trainin data\n",
        "# classifier.fit(X_train,Y_train)\n",
        "\n",
        "\n",
        "X_train = X_train.reshape(-1, 784)\n",
        "Y_train = Y_train\n",
        "X_test = X_test.reshape(-1, 784)\n",
        "Y_test = Y_test\n",
        "print(X_train.shape)\n",
        "print(Y_train.shape)\n",
        "print(X_test.shape)\n",
        "print(Y_test.shape)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(6500, 784)\n",
            "(6500,)\n",
            "(2000, 784)\n",
            "(2000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Gh9oMEaOL4Zn"
      },
      "source": [
        "# RBF KERNEL\n",
        "\n",
        "# # C=1 Default kernel = default (rbf)\n",
        "# from sklearn import svm\n",
        "# # Create a classifier: a support vector classifier\n",
        "# classifier = svm.SVC(gamma=0.001)\n",
        "# #fit to the trainin data\n",
        "# classifier.fit(X_train,Y_train)\n",
        "# Y_pred = classifier.predict(X_test)\n",
        "# accu = classifier.score(X_test, Y_test)\n",
        "# print(\"Testing accuracy:\",accu)\n",
        "# from sklearn.metrics import classification_report, confusion_matrix\n",
        "# print(\"Classification Report: \\n\",classification_report(Y_test, Y_pred))\n",
        "# print(\"Confusion Matrix: \\n\", confusion_matrix(Y_test, Y_pred))\n",
        "\n",
        "# import seaborn as sns\n",
        "\n",
        "# matrix = confusion_matrix(Y_test, Y_pred)\n",
        "# sns.heatmap(matrix, square=True, annot=True, fmt='d', cbar=False,\n",
        "#             xticklabels=classes,\n",
        "#             yticklabels=classes)\n",
        "# plt.xlabel('true label')\n",
        "# plt.ylabel('predicted label');"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 568
        },
        "id": "Rq6vpp9A6mNW",
        "outputId": "14bd3781-9d13-4ed3-ec5d-94cab09fab33"
      },
      "source": [
        "# C=1 kernel = linear\n",
        "from sklearn import svm\n",
        "# Create a classifier: a support vector classifier\n",
        "classifier = svm.SVC(kernel='linear', probability=True)\n",
        "#fit to the trainin data\n",
        "classifier.fit(X_train/255,Y_train)\n",
        "# p = np.array(classifier.decision_function(X_train))\n",
        "# prob = np.exp(p)/np.sum(np.exp(p)) # softmax \n",
        "# Y_score = prob\n",
        "Y_pred = classifier.predict(X_test/255)\n",
        "accu = classifier.score(X_test/255, Y_test)\n",
        "print(\"Testing accuracy:\",accu)\n",
        "print(\"----------------------------------------\")\n",
        "from sklearn.metrics import classification_report, confusion_matrix, roc_curve\n",
        "print(\"Classification Report: \\n\",classification_report(Y_test, Y_pred))\n",
        "print(\"Confusion Matrix: \\n\", confusion_matrix(Y_test, Y_pred))\n",
        "print(\"----------------------------------------\")\n",
        "import seaborn as sns\n",
        "\n",
        "matrix = confusion_matrix(Y_test, Y_pred)\n",
        "sns.heatmap(matrix, square=True, annot=True, fmt='d', cbar=False,)\n",
        "plt.xlabel('true label')\n",
        "plt.ylabel('predicted label');\n",
        "print(\"----------------------------------------\")\n"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Testing accuracy: 0.732\n",
            "----------------------------------------\n",
            "Classification Report: \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           2       0.66      0.98      0.79      1000\n",
            "           4       0.96      0.48      0.64      1000\n",
            "\n",
            "    accuracy                           0.73      2000\n",
            "   macro avg       0.81      0.73      0.71      2000\n",
            "weighted avg       0.81      0.73      0.71      2000\n",
            "\n",
            "Confusion Matrix: \n",
            " [[979  21]\n",
            " [515 485]]\n",
            "----------------------------------------\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": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YCNK8yik0aGz",
        "outputId": "5e67443e-d3cc-4357-86d8-b1799efacfc5"
      },
      "source": [
        "print(\"Mean / STD:\")\n",
        "a_mean=(matrix[0,0]+matrix[1,1])/(matrix[0,0]+matrix[0,1]+matrix[1,0]+matrix[1,1])\n",
        "result=a_mean*100\n",
        "print(result)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mean / STD:\n",
            "73.2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1_e75za46LMu"
      },
      "source": [
        ""
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "MTsjovHOh3Et",
        "outputId": "16c13bd6-f3d2-459c-e9ae-2ee1e577c9ba"
      },
      "source": [
        "fpr, tpr, thresholds = roc_curve(Y_test, Y_pred, pos_label=2)\n",
        "plt.plot(tpr, fpr)\n",
        "print(thresholds)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[5 4 2]\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": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rAm3FsTlUOkN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "932e20de-af9c-464d-e86c-b23d80925557"
      },
      "source": [
        "# tn = 817\n",
        "# fp = 233\n",
        "# fn = 183\n",
        "# tp = 767\n",
        "# print(Y_test)\n",
        "# print(Y_pred)\n",
        "Y_test = np.where(Y_test==2, 0, Y_test)\n",
        "Y_test = np.where(Y_test==4, 1, Y_test) \n",
        "Y_pred = np.where(Y_pred==2, 0, Y_pred)\n",
        "Y_pred = np.where(Y_pred==4, 1, Y_pred)\n",
        "print(Y_test) # IMP\n",
        "print(Y_pred)\n"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0 0 0 ... 1 1 1]\n",
            "[0 0 0 ... 0 1 0]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oZPhrrCSTyBc"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hgony12ZUS9F"
      },
      "source": [
        "# fpr = fp/(fp+tn)\n",
        "# tpr = tp/(tp+fn)\n",
        "# plt.plot(fpr,tpr,linestyle='--')"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mex_VUHkUuaQ"
      },
      "source": [
        "# Didnt work!!!\n",
        "# tns=[]\n",
        "# fps=[]\n",
        "# fns=[]\n",
        "# tps=[]\n",
        "\n",
        "# tn=0\n",
        "# fp=0\n",
        "# fn=0\n",
        "# tp=0\n",
        "# for i in range(len(Y_test)):\n",
        "#   if(Y_test[i]):\n",
        "#     if(Y_test[i]==Y_pred[i]):\n",
        "#       tp=tp+1\n",
        "#       tps.append(tp)\n",
        "#     else:\n",
        "#       fn=fn+1\n",
        "#       fns.append(fn)\n",
        "#   else:\n",
        "#     if(Y_test[i]==Y_pred[i]):\n",
        "#       tn=tn+1\n",
        "#       tns.append(tn)\n",
        "#     else:\n",
        "#       fp=fp+1\n",
        "#       fps.append(fp)\n",
        "\n",
        "# # print(tn)\n",
        "# # print(fp)\n",
        "# # print(fn)\n",
        "# # print(tp)"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p_n8UY4RIQIC"
      },
      "source": [
        "to code beyond this,I referred. specially the threshold part. hence referred GitHub Repo: The notebook link: https://github.com/tejas2454/Bias-Detection-and-Mitigation-Trusted-AI/blob/master/MNIST%20Bias%20detecton%20ROC%20scratch%20code.ipynb"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cgJDWYqwhCWw",
        "outputId": "99223ffb-074d-4d22-e12d-d4091119e255"
      },
      "source": [
        "prb = classifier.predict_proba(X_test/255)\n",
        "print(prb.shape)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(2000, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "vXgfbSpcig-R",
        "outputId": "3d8781aa-f06a-45fd-edb8-8768709c8c71"
      },
      "source": [
        "prb\n",
        "plt.plot(prb)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7fe29096fb00>,\n",
              " <matplotlib.lines.Line2D at 0x7fe29096fc18>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2ziQBc9k9N0i"
      },
      "source": [
        "prb=prb[:,0]"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3-h_YbkoKVXj",
        "outputId": "e91027cd-9663-40ee-da9d-f4cb2048d074"
      },
      "source": [
        ""
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, ..., 1, 1, 1], dtype=uint8)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tDJWfCK49vi-"
      },
      "source": [
        "roc=np.concatenate((Y_test.astype(int).reshape((2000,1)),Y_pred.reshape((2000,1)),prb.reshape((2000,1))),axis=1)"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C3PCvfyM_lJW",
        "outputId": "f89a0b5e-78f2-4ab4-9e6c-85d634c4a424"
      },
      "source": [
        "roc"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.        , 0.        , 0.95353829],\n",
              "       [0.        , 0.        , 0.99693415],\n",
              "       [0.        , 0.        , 0.99991678],\n",
              "       ...,\n",
              "       [1.        , 0.        , 0.92680459],\n",
              "       [1.        , 1.        , 0.41224875],\n",
              "       [1.        , 0.        , 0.8637605 ]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9kZD5LHKAuAQ",
        "outputId": "65b4ae38-a0ce-44c6-8d01-09fd2c62479f"
      },
      "source": [
        "pullover_2 = []\n",
        "coat_4 = []\n",
        "for i in range(len(roc)):\n",
        "  if (roc[i,0]==1):\n",
        "    coat_4.append(roc[i,:])\n",
        "  else:\n",
        "    pullover_2.append(roc[i,:])\n",
        "\n",
        "print(\"coat_4\")\n",
        "print(coat_4)\n",
        "print('\\n')\n",
        "print(\"pullover_2\")\n",
        "print(pullover_2)\n",
        "\n",
        "thresh=np.unique(np.concatenate((np.asarray(coat_4)[:,-1],np.asarray(pullover_2)[:,-1]),axis=0))\n",
        "thresh_matrix=np.sort(thresh[::-1].T).reshape((len(thresh),1))[::-1]\n",
        "sequence_arr=np.concatenate((np.asarray(coat_4)[:,:],np.asarray(pullover_2)[:,:]),axis=0)\n",
        "\n"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "coat_4\n",
            "[array([1.        , 1.        , 0.32452456]), array([1.        , 0.        , 0.71007676]), array([1.        , 0.        , 0.88933931]), array([1.        , 0.        , 0.99702148]), array([1.        , 0.        , 0.93746265]), array([1.        , 0.        , 0.68618004]), array([1.        , 0.        , 0.91174152]), array([1.        , 0.        , 0.79953144]), array([1.        , 1.        , 0.66495649]), array([1.        , 0.        , 0.90674529]), array([1.        , 1.        , 0.42633481]), array([1.        , 0.        , 0.96632826]), array([1.        , 0.        , 0.95599638]), array([1.        , 0.        , 0.92838175]), array([1.        , 0.        , 0.99097753]), array([1.        , 0.        , 0.95191344]), array([1.        , 1.        , 0.47313363]), array([1.        , 0.        , 0.77497125]), array([1.        , 1.        , 0.45284579]), array([1.        , 0.        , 0.81797683]), array([1.       , 0.       , 0.7499007]), array([1.        , 1.        , 0.64411644]), array([1.        , 0.        , 0.86599919]), array([1.        , 0.        , 0.72747896]), array([1.        , 1.        , 0.11663569]), array([1.        , 0.        , 0.99892599]), array([1.        , 0.        , 0.96755082]), array([1.       , 1.       , 0.5822601]), array([1.        , 0.        , 0.79947086]), array([1.        , 1.        , 0.22331259]), array([1.        , 1.        , 0.51139934]), array([1.        , 0.        , 0.79396655]), array([1.        , 0.        , 0.88319953]), array([1.        , 0.        , 0.86033778]), array([1.        , 1.        , 0.12287031]), array([1.       , 1.       , 0.1046691]), array([1.        , 0.        , 0.73452773]), array([1.        , 0.        , 0.74160777]), array([1.        , 0.        , 0.77738009]), array([1.        , 0.        , 0.92810484]), array([1.        , 1.        , 0.29564426]), array([1.        , 1.        , 0.13098762]), array([1.        , 1.        , 0.20720371]), array([1.        , 0.        , 0.69206014]), array([1.        , 1.        , 0.52091216]), array([1.        , 1.        , 0.20335364]), array([1.        , 0.        , 0.93665476]), array([1.        , 0.        , 0.80692542]), array([1.        , 0.        , 0.73539367]), array([1. , 1. , 0.5]), array([1.        , 1.        , 0.11242833]), array([1.        , 0.        , 0.83406983]), array([1.        , 1.        , 0.60325504]), array([1.        , 0.        , 0.99517184]), array([1.        , 0.        , 0.89416106]), array([1.        , 0.        , 0.91220196]), array([1.        , 1.        , 0.37579076]), array([1.        , 0.        , 0.91974456]), array([1.        , 0.        , 0.96794417]), array([1.        , 0.        , 0.68298125]), array([1.        , 1.        , 0.56427439]), array([1.        , 0.        , 0.98915234]), array([1.        , 1.        , 0.09301517]), array([1.        , 1.        , 0.64123965]), array([1.        , 0.        , 0.76758134]), array([1.        , 0.        , 0.68961868]), array([1.        , 1.        , 0.34296034]), array([1.        , 1.        , 0.59022641]), array([1.        , 1.        , 0.03775493]), array([1.        , 1.        , 0.40541568]), array([1.        , 0.        , 0.99618868]), array([1.        , 1.        , 0.19059796]), array([1.        , 1.        , 0.52511521]), array([1.        , 1.        , 0.18249712]), array([1.        , 1.        , 0.16474872]), array([1.        , 1.        , 0.55103405]), array([1.        , 0.        , 0.91545387]), array([1.        , 0.        , 0.93044702]), array([1.        , 0.        , 0.78586665]), array([1.        , 1.        , 0.62069901]), array([1.        , 0.        , 0.91036891]), array([1.       , 0.       , 0.9144568]), array([1.        , 0.        , 0.89009507]), array([1.       , 0.       , 0.6912221]), array([1.        , 0.        , 0.85421355]), array([1.        , 0.        , 0.85890363]), array([1.        , 0.        , 0.87273274]), array([1.        , 1.        , 0.47743675]), array([1.        , 1.        , 0.08887236]), array([1.        , 0.        , 0.89568315]), array([1.        , 0.        , 0.72190534]), array([1.       , 0.       , 0.9995523]), array([1.        , 0.        , 0.95772858]), array([1.        , 0.        , 0.66962436]), array([1.        , 1.        , 0.37910913]), array([1.        , 0.        , 0.67059806]), array([1.        , 1.        , 0.36727144]), array([1.        , 0.        , 0.79398578]), array([1.        , 0.        , 0.72197172]), array([1.        , 0.        , 0.95830354]), array([1.        , 0.        , 0.91020802]), array([1.        , 0.        , 0.98982792]), array([1.        , 0.        , 0.84207151]), array([1.        , 1.        , 0.64230749]), array([1.        , 1.        , 0.07770754]), array([1.        , 1.        , 0.15419017]), array([1.        , 1.        , 0.56322168]), array([1.        , 1.        , 0.41488749]), array([1.        , 0.        , 0.72555557]), array([1.        , 0.        , 0.78335519]), array([1.        , 1.        , 0.12477319]), array([1.        , 1.        , 0.19630956]), array([1.        , 0.        , 0.93322582]), array([1.        , 0.        , 0.80989636]), array([1.        , 0.        , 0.77894755]), array([1.        , 1.        , 0.49147734]), array([1.        , 0.        , 0.74900335]), array([1.        , 1.        , 0.46980879]), array([1.        , 0.        , 0.74591593]), array([1.        , 0.        , 0.96473591]), array([1.        , 0.        , 0.89802965]), array([1.       , 1.       , 0.2361766]), array([1.        , 0.        , 0.97550947]), array([1.        , 0.        , 0.98036666]), array([1.        , 0.        , 0.95762042]), array([1.       , 1.       , 0.1745326]), array([1.        , 1.        , 0.50869821]), array([1.        , 0.        , 0.88179709]), array([1.        , 0.        , 0.82869085]), array([1.        , 1.        , 0.61019825]), array([1.        , 1.        , 0.29340114]), array([1.        , 1.        , 0.18594181]), array([1.        , 0.        , 0.77827737]), array([1.        , 1.        , 0.52890039]), array([1.       , 1.       , 0.6285926]), array([1.        , 1.        , 0.59712669]), array([1.       , 1.       , 0.6529871]), array([1.        , 0.        , 0.74485116]), array([1.        , 1.        , 0.54932349]), array([1.        , 1.        , 0.17631456]), array([1.        , 0.        , 0.94353248]), array([1.      , 0.      , 0.835302]), array([1.        , 0.        , 0.74485959]), array([1.        , 0.        , 0.78102675]), array([1.        , 1.        , 0.56567716]), array([1.        , 0.        , 0.73650971]), array([1.        , 1.        , 0.39435987]), array([1.       , 0.       , 0.8108845]), array([1.        , 1.        , 0.42433083]), array([1.        , 1.        , 0.53408959]), array([1.        , 1.        , 0.27968637]), array([1.     , 1.     , 0.40977]), array([1.        , 0.        , 0.94911972]), array([1.        , 1.        , 0.55423681]), array([1.        , 1.        , 0.43108135]), array([1.        , 0.        , 0.85282383]), array([1.        , 0.        , 0.79194825]), array([1.        , 1.        , 0.60667315]), array([1.        , 0.        , 0.96980722]), array([1.        , 1.        , 0.61381008]), array([1.        , 0.        , 0.90044648]), array([1.        , 0.        , 0.94463765]), array([1.        , 0.        , 0.99323495]), array([1.        , 1.        , 0.10866174]), array([1.        , 1.        , 0.06875886]), array([1.        , 1.        , 0.20998172]), array([1. , 1. , 0.5]), array([1.        , 0.        , 0.94380103]), array([1.       , 1.       , 0.5064333]), array([1.        , 0.        , 0.96171524]), array([1.        , 1.        , 0.14556505]), array([1.        , 1.        , 0.42065383]), array([1.        , 0.        , 0.95086224]), array([1.       , 1.       , 0.6190364]), array([1.       , 0.       , 0.8686604]), array([1.       , 1.       , 0.4502214]), array([1.       , 0.       , 0.9386081]), array([1.        , 0.        , 0.90720989]), array([1.      , 1.      , 0.482807]), array([1.        , 1.        , 0.08619513]), array([1.        , 1.        , 0.64887593]), array([1.        , 0.        , 0.80115502]), array([1.       , 0.       , 0.7216696]), array([1.        , 0.        , 0.99571281]), array([1.        , 1.        , 0.19822803]), array([1.        , 0.        , 0.79437421]), array([1.        , 0.        , 0.79122578]), array([1.        , 0.        , 0.98970122]), array([1.        , 0.        , 0.87538574]), array([1.        , 1.        , 0.28045772]), array([1.        , 1.        , 0.17909902]), array([1.        , 0.        , 0.95832575]), array([1.        , 1.        , 0.24080577]), array([1.       , 1.       , 0.0938061]), array([1.        , 1.        , 0.35389992]), array([1.        , 0.        , 0.79612816]), array([1.        , 1.        , 0.60312579]), array([1.        , 1.        , 0.43051237]), array([1.        , 1.        , 0.44499764]), array([1.       , 1.       , 0.3843568]), array([1.        , 1.        , 0.60933403]), array([1.        , 1.        , 0.57831974]), array([1.        , 1.        , 0.39706031]), array([1.        , 0.        , 0.79405353]), array([1.        , 0.        , 0.79200582]), array([1.        , 0.        , 0.96441808]), array([1.        , 0.        , 0.96396851]), array([1.        , 1.        , 0.02049845]), array([1.        , 0.        , 0.69300386]), array([1.        , 0.        , 0.67224566]), array([1.        , 0.        , 0.98914442]), array([1.        , 0.        , 0.87132399]), array([1.        , 0.        , 0.71969908]), array([1.        , 1.        , 0.51606174]), array([1.        , 1.        , 0.11129131]), array([1.        , 0.        , 0.72852478]), array([1.        , 0.        , 0.86307277]), array([1.        , 0.        , 0.99621196]), array([1.        , 1.        , 0.60570063]), array([1.        , 0.        , 0.80207757]), array([1.        , 1.        , 0.64930885]), array([1.        , 0.        , 0.99962653]), array([1.        , 1.        , 0.29427187]), array([1.        , 1.        , 0.64807322]), array([1.        , 0.        , 0.90553934]), array([1.        , 0.        , 0.95043342]), array([1.       , 1.       , 0.2324429]), array([1.        , 0.        , 0.93558282]), array([1.        , 1.        , 0.60993747]), array([1.        , 1.        , 0.22595271]), array([1.        , 1.        , 0.65459159]), array([1.        , 1.        , 0.65011994]), array([1.        , 1.        , 0.63924222]), array([1.       , 0.       , 0.7377184]), array([1.        , 0.        , 0.98330755]), array([1.        , 1.        , 0.43984778]), array([1.        , 1.        , 0.40119367]), array([1.        , 0.        , 0.99696812]), array([1.        , 1.        , 0.65627208]), array([1.        , 0.        , 0.67818573]), array([1.        , 1.        , 0.61706113]), array([1.        , 1.        , 0.61988364]), array([1.        , 1.        , 0.60262055]), array([1.        , 0.        , 0.76359599]), array([1.        , 1.        , 0.55139115]), array([1.        , 0.        , 0.76687696]), array([1.        , 1.        , 0.45623788]), array([1.        , 0.        , 0.67645818]), array([1.        , 1.        , 0.60072937]), array([1.        , 0.        , 0.88541538]), array([1.        , 1.        , 0.53483841]), array([1.        , 1.        , 0.42003407]), array([1.        , 1.        , 0.48646531]), array([1.        , 1.        , 0.34428184]), array([1.        , 1.        , 0.37922509]), array([1.       , 1.       , 0.5410167]), array([1.        , 0.        , 0.98567551]), array([1.        , 0.        , 0.78550775]), array([1.        , 1.        , 0.46869093]), array([1.        , 0.        , 0.93124564]), array([1.       , 0.       , 0.7821432]), array([1.        , 0.        , 0.82166985]), array([1.        , 0.        , 0.87963648]), array([1.        , 0.        , 0.67957879]), array([1.        , 1.        , 0.13778466]), array([1.        , 0.        , 0.97303388]), array([1.        , 0.        , 0.93950722]), array([1.        , 0.        , 0.74472029]), array([1.        , 1.        , 0.48569848]), array([1.        , 1.        , 0.17007855]), array([1.        , 0.        , 0.98426404]), array([1.        , 0.        , 0.99433163]), array([1.        , 1.        , 0.58917103]), array([1.        , 1.        , 0.45365467]), array([1.        , 1.        , 0.30816737]), array([1.        , 1.        , 0.05607881]), array([1.       , 0.       , 0.9318562]), array([1.        , 1.        , 0.42785202]), array([1.        , 0.        , 0.69017863]), array([1.        , 1.        , 0.29056661]), array([1.        , 0.        , 0.77851364]), array([1.        , 1.        , 0.58245379]), array([1.        , 0.        , 0.99366315]), array([1.        , 1.        , 0.33165925]), array([1.        , 1.        , 0.49394304]), array([1.        , 0.        , 0.73956164]), array([1.        , 1.        , 0.14248067]), array([1.        , 1.        , 0.28315346]), array([1.        , 0.        , 0.99066834]), array([1.        , 1.        , 0.62082777]), array([1.        , 1.        , 0.54493824]), array([1.        , 0.        , 0.73646491]), array([1.        , 1.        , 0.40837673]), array([1.        , 0.        , 0.81184855]), array([1.        , 1.        , 0.48041183]), array([1.        , 0.        , 0.92700951]), array([1.       , 0.       , 0.8375727]), array([1.        , 1.        , 0.62923915]), array([1.        , 1.        , 0.20191965]), array([1.        , 0.        , 0.99232338]), array([1.        , 0.        , 0.87495068]), array([1.        , 1.        , 0.22293889]), array([1.        , 1.        , 0.66621661]), array([1.        , 1.        , 0.56940236]), array([1.        , 0.        , 0.71587646]), array([1.        , 0.        , 0.95931527]), array([1.        , 1.        , 0.37571495]), array([1.        , 1.        , 0.35457913]), array([1.        , 1.        , 0.26874957]), array([1.        , 0.        , 0.77245472]), array([1.        , 1.        , 0.24353731]), array([1.        , 1.        , 0.61047156]), array([1.        , 1.        , 0.44675521]), array([1.        , 0.        , 0.90284908]), array([1.        , 0.        , 0.83471143]), array([1.       , 0.       , 0.6787325]), array([1.        , 1.        , 0.52371099]), array([1.        , 0.        , 0.90953141]), array([1.        , 0.        , 0.99505356]), array([1.       , 1.       , 0.2470092]), array([1.        , 0.        , 0.86629566]), array([1.        , 0.        , 0.83398025]), array([1.        , 0.        , 0.70677959]), array([1.       , 1.       , 0.0852353]), array([1.        , 0.        , 0.91783563]), array([1.        , 0.        , 0.88449282]), array([1.       , 0.       , 0.9397142]), array([1.       , 0.       , 0.9860384]), array([1.        , 0.        , 0.94275988]), array([1.        , 0.        , 0.96496176]), array([1.        , 1.        , 0.39093984]), array([1.        , 0.        , 0.94738326]), array([1.        , 0.        , 0.83595416]), array([1.        , 1.        , 0.12131688]), array([1.        , 1.        , 0.22699266]), array([1.        , 0.        , 0.77668471]), array([1.        , 0.        , 0.95043523]), array([1.        , 1.        , 0.01401478]), array([1.        , 1.        , 0.57672466]), array([1.        , 0.        , 0.88733201]), array([1.        , 0.        , 0.84047358]), array([1.        , 1.        , 0.50892046]), array([1.        , 0.        , 0.96970576]), array([1.        , 0.        , 0.67172631]), array([1.        , 1.        , 0.38461885]), array([1.        , 1.        , 0.66516159]), array([1.        , 0.        , 0.98826636]), array([1.        , 1.        , 0.55360207]), array([1.        , 1.        , 0.12841886]), array([1.        , 0.        , 0.72244001]), array([1.        , 1.        , 0.28487136]), array([1.        , 1.        , 0.34402053]), array([1.        , 1.        , 0.46930273]), array([1.        , 1.        , 0.42789154]), array([1.        , 0.        , 0.79657308]), array([1.        , 0.        , 0.74797555]), array([1.        , 1.        , 0.65672417]), array([1.        , 1.        , 0.53224816]), array([1.        , 0.        , 0.82828223]), array([1. , 1. , 0.5]), array([1.        , 1.        , 0.37510673]), array([1.        , 0.        , 0.78465583]), array([1.        , 1.        , 0.34843098]), array([1.        , 0.        , 0.80677457]), array([1.      , 0.      , 0.998126]), array([1.        , 1.        , 0.31182839]), array([1.        , 0.        , 0.98209685]), array([1.        , 0.        , 0.78833894]), array([1.        , 1.        , 0.45617254]), array([1.        , 0.        , 0.99836762]), array([1.        , 1.        , 0.31238951]), array([1.        , 1.        , 0.42859907]), array([1.        , 1.        , 0.65903173]), array([1.        , 1.        , 0.25693532]), array([1.        , 1.        , 0.06500735]), array([1.        , 1.        , 0.23408016]), array([1.        , 0.        , 0.81814774]), array([1.        , 0.        , 0.90093681]), array([1.        , 0.        , 0.76476944]), array([1.        , 1.        , 0.54564211]), array([1.        , 1.        , 0.50722886]), array([1.        , 1.        , 0.37597164]), array([1.        , 0.        , 0.91084426]), array([1.        , 0.        , 0.85362655]), array([1.       , 1.       , 0.5546193]), array([1.        , 1.        , 0.26361089]), array([1.       , 0.       , 0.7808596]), array([1.        , 0.        , 0.95357521]), array([1.        , 1.        , 0.30343961]), array([1.        , 0.        , 0.69727416]), array([1.        , 1.        , 0.48056708]), array([1.        , 0.        , 0.86725714]), array([1.        , 1.        , 0.65088745]), array([1.        , 1.        , 0.26542632]), array([1.        , 0.        , 0.91898756]), array([1.        , 1.        , 0.15133286]), array([1.        , 0.        , 0.95715719]), array([1.        , 1.        , 0.45622497]), array([1.        , 1.        , 0.48694988]), array([1.        , 1.        , 0.17935214]), array([1.        , 0.        , 0.90359945]), array([1.        , 0.        , 0.72575863]), array([1.        , 1.        , 0.15892343]), array([1.        , 0.        , 0.71208577]), array([1.        , 0.        , 0.98533448]), array([1.        , 1.        , 0.36087269]), array([1.        , 1.        , 0.61595163]), array([1.        , 0.        , 0.76778882]), array([1.        , 0.        , 0.77456357]), array([1.        , 1.        , 0.11448898]), array([1.        , 0.        , 0.91755043]), array([1.        , 1.        , 0.22251967]), array([1.        , 1.        , 0.64250423]), array([1.        , 0.        , 0.82604336]), array([1.        , 0.        , 0.91155976]), array([1.        , 0.        , 0.94259388]), array([1.        , 1.        , 0.54722531]), array([1.        , 1.        , 0.61986485]), array([1.        , 1.        , 0.03582163]), array([1.        , 0.        , 0.83173839]), array([1.        , 1.        , 0.52721854]), array([1.        , 1.        , 0.52425868]), array([1.        , 1.        , 0.56641842]), array([1.        , 1.        , 0.66048548]), array([1.        , 0.        , 0.77196546]), array([1.        , 0.        , 0.88321548]), array([1.        , 0.        , 0.77780121]), array([1.        , 1.        , 0.62319765]), array([1.        , 0.        , 0.78460741]), array([1.        , 0.        , 0.96038152]), array([1.        , 1.        , 0.56395457]), array([1.        , 0.        , 0.88915822]), array([1.        , 0.        , 0.91717502]), array([1.        , 0.        , 0.98440207]), array([1.        , 0.        , 0.88113459]), array([1.        , 1.        , 0.03903721]), array([1.        , 0.        , 0.94519307]), array([1.        , 1.        , 0.59477759]), array([1.        , 1.        , 0.60386879]), array([1.        , 0.        , 0.84922221]), array([1.        , 1.        , 0.56070456]), array([1.        , 0.        , 0.69279046]), array([1.      , 0.      , 0.905765]), array([1.       , 0.       , 0.8814224]), array([1.        , 0.        , 0.74139183]), array([1.        , 1.        , 0.35098658]), array([1.        , 1.        , 0.15553688]), array([1.        , 0.        , 0.87104479]), array([1.        , 0.        , 0.88923859]), array([1.        , 0.        , 0.93701542]), array([1.        , 1.        , 0.23161168]), array([1.        , 0.        , 0.86446358]), array([1.        , 1.        , 0.28319648]), array([1.        , 0.        , 0.66930014]), array([1.        , 0.        , 0.89070915]), array([1.       , 1.       , 0.3447024]), array([1.        , 0.        , 0.93347805]), array([1.       , 0.       , 0.7240028]), array([1.        , 1.        , 0.46545261]), array([1.        , 1.        , 0.26243023]), array([1.        , 0.        , 0.95943149]), array([1.        , 0.        , 0.71833479]), array([1.        , 1.        , 0.15900112]), array([1.        , 1.        , 0.57535605]), array([1.        , 1.        , 0.32077296]), array([1.        , 0.        , 0.95699005]), array([1.        , 1.        , 0.59489903]), array([1.        , 1.        , 0.62198073]), array([1.        , 1.        , 0.63297463]), array([1.        , 1.        , 0.20352705]), array([1.        , 0.        , 0.72126903]), array([1.        , 1.        , 0.27623104]), array([1.        , 1.        , 0.16026703]), array([1.        , 0.        , 0.99349721]), array([1.        , 0.        , 0.74064785]), array([1.        , 1.        , 0.48399304]), array([1.        , 0.        , 0.92936492]), array([1.        , 0.        , 0.93011172]), array([1.        , 1.        , 0.40134602]), array([1.        , 0.        , 0.73754374]), array([1.       , 1.       , 0.2529357]), array([1.        , 1.        , 0.47373231]), array([1.        , 0.        , 0.86886663]), array([1.        , 1.        , 0.58968213]), array([1.        , 0.        , 0.98372439]), array([1.        , 0.        , 0.96063115]), array([1.        , 1.        , 0.20500316]), array([1.        , 0.        , 0.99961634]), array([1.        , 0.        , 0.78777818]), array([1.        , 0.        , 0.89363929]), array([1.        , 1.        , 0.56655018]), array([1.        , 0.        , 0.97320139]), array([1.        , 0.        , 0.95004234]), array([1.        , 1.        , 0.15393089]), array([1.        , 1.        , 0.46275393]), array([1.        , 0.        , 0.67952869]), array([1.        , 0.        , 0.97042824]), array([1.        , 1.        , 0.06806941]), array([1.        , 0.        , 0.72370964]), array([1.        , 0.        , 0.89996995]), array([1.        , 1.        , 0.62624655]), array([1.        , 0.        , 0.92720334]), array([1.        , 1.        , 0.30215344]), array([1.        , 1.        , 0.58358007]), array([1.        , 0.        , 0.81223949]), array([1.        , 0.        , 0.85747723]), array([1.        , 0.        , 0.87116642]), array([1.        , 1.        , 0.41921349]), array([1.        , 0.        , 0.94435687]), array([1.        , 0.        , 0.74425804]), array([1.        , 1.        , 0.65856283]), array([1.        , 1.        , 0.30302137]), array([1.        , 1.        , 0.43701951]), array([1.       , 1.       , 0.6271652]), array([1.        , 0.        , 0.82550863]), array([1. , 1. , 0.5]), array([1.        , 0.        , 0.78343888]), array([1.        , 0.        , 0.90663703]), array([1.        , 1.        , 0.27229629]), array([1.        , 1.        , 0.03018779]), array([1.        , 0.        , 0.94830704]), array([1.        , 0.        , 0.79954736]), array([1.        , 0.        , 0.87915946]), array([1.        , 0.        , 0.97562026]), array([1.        , 1.        , 0.58741947]), array([1.        , 1.        , 0.21559409]), array([1.        , 0.        , 0.98880007]), array([1.        , 1.        , 0.35504159]), array([1.        , 0.        , 0.99137155]), array([1.        , 1.        , 0.17717499]), array([1.        , 0.        , 0.87877151]), array([1.        , 0.        , 0.90742822]), array([1.        , 0.        , 0.67248665]), array([1.       , 1.       , 0.2783561]), array([1.        , 1.        , 0.25587936]), array([1.        , 0.        , 0.86225008]), array([1.        , 0.        , 0.68439758]), array([1.        , 0.        , 0.91618198]), array([1.       , 1.       , 0.5882328]), array([1.       , 1.       , 0.3476649]), array([1.        , 0.        , 0.98866491]), array([1.        , 0.        , 0.79942029]), array([1.        , 1.        , 0.66636409]), array([1.        , 1.        , 0.40027958]), array([1.       , 0.       , 0.9178165]), array([1.        , 0.        , 0.90152602]), array([1.        , 1.        , 0.26494802]), array([1.        , 1.        , 0.29633663]), array([1.        , 1.        , 0.34257815]), array([1.        , 0.        , 0.96062647]), array([1.        , 1.        , 0.31293497]), array([1.        , 0.        , 0.81173768]), array([1.        , 0.        , 0.92849357]), array([1.        , 1.        , 0.49295323]), array([1.        , 1.        , 0.18402532]), array([1.        , 0.        , 0.78743474]), array([1.        , 1.        , 0.16904672]), array([1.        , 1.        , 0.34442993]), array([1.        , 0.        , 0.72334474]), array([1.        , 1.        , 0.20622007]), array([1.        , 0.        , 0.74039137]), array([1.        , 0.        , 0.75422056]), array([1.        , 0.        , 0.69334254]), array([1.        , 0.        , 0.84405317]), array([1.        , 1.        , 0.36426967]), array([1.        , 1.        , 0.18459655]), array([1.        , 1.        , 0.31978155]), array([1.        , 0.        , 0.98521922]), array([1.        , 1.        , 0.30204968]), array([1.        , 1.        , 0.09237662]), array([1.        , 0.        , 0.82628864]), array([1.        , 1.        , 0.55649618]), array([1.        , 0.        , 0.98867772]), array([1.       , 0.       , 0.9541622]), array([1.        , 0.        , 0.95157428]), array([1.        , 0.        , 0.94111844]), array([1.        , 0.        , 0.93461636]), array([1.        , 0.        , 0.91315697]), array([1.        , 0.        , 0.96988623]), array([1.        , 1.        , 0.12727931]), array([1.        , 1.        , 0.38625764]), array([1.       , 0.       , 0.9161456]), array([1.        , 1.        , 0.66555514]), array([1.        , 1.        , 0.41876788]), array([1.        , 0.        , 0.99949891]), array([1.        , 1.        , 0.61539315]), array([1.        , 1.        , 0.47851492]), array([1.        , 0.        , 0.97268499]), array([1.        , 1.        , 0.00628837]), array([1.        , 0.        , 0.94312096]), array([1.        , 1.        , 0.44723702]), array([1.        , 0.        , 0.84278873]), array([1.       , 1.       , 0.2011731]), array([1.        , 1.        , 0.66354555]), array([1.        , 0.        , 0.79907088]), array([1.        , 1.        , 0.13808019]), array([1.        , 0.        , 0.73007085]), array([1.        , 1.        , 0.21457378]), array([1.        , 0.        , 0.73356712]), array([1.        , 1.        , 0.47284829]), array([1.        , 1.        , 0.25984317]), array([1.       , 0.       , 0.7590516]), array([1.        , 1.        , 0.66653036]), array([1.        , 0.        , 0.96083693]), array([1.        , 1.        , 0.33691076]), array([1.        , 0.        , 0.71425517]), array([1.        , 0.        , 0.98090815]), array([1.        , 1.        , 0.15631163]), array([1.        , 1.        , 0.62830434]), array([1.        , 0.        , 0.89563069]), array([1.        , 0.        , 0.81665156]), array([1.        , 1.        , 0.12705347]), array([1.        , 1.        , 0.66109933]), array([1.        , 0.        , 0.90848268]), array([1.       , 0.       , 0.7483353]), array([1.        , 1.        , 0.55693673]), array([1.        , 0.        , 0.98192967]), array([1.        , 1.        , 0.16521133]), array([1.        , 1.        , 0.37630178]), array([1.       , 0.       , 0.9802853]), array([1.        , 0.        , 0.92440151]), array([1.       , 0.       , 0.8150617]), array([1.        , 1.        , 0.37272827]), array([1.        , 1.        , 0.42685864]), array([1.        , 0.        , 0.99429883]), array([1.        , 0.        , 0.94629589]), array([1.        , 1.        , 0.52398205]), array([1.        , 1.        , 0.40001857]), array([1.        , 1.        , 0.44940689]), array([1.        , 1.        , 0.01721209]), array([1.        , 1.        , 0.20766824]), array([1.        , 0.        , 0.77355209]), array([1.        , 1.        , 0.56815036]), array([1.        , 1.        , 0.60203176]), array([1.       , 1.       , 0.1440685]), array([1.        , 0.        , 0.84742402]), array([1.        , 1.        , 0.09674025]), array([1.        , 1.        , 0.35366037]), array([1.        , 0.        , 0.87385559]), array([1.        , 0.        , 0.92586668]), array([1.        , 0.        , 0.91866251]), array([1.        , 1.        , 0.44394605]), array([1.        , 1.        , 0.47069159]), array([1.        , 1.        , 0.53636417]), array([1.        , 0.        , 0.69394973]), array([1.        , 0.        , 0.98231553]), array([1.        , 1.        , 0.63779729]), array([1.        , 0.        , 0.84318628]), array([1.        , 0.        , 0.87008968]), array([1.        , 0.        , 0.75797729]), array([1.        , 1.        , 0.63325213]), array([1.        , 0.        , 0.67861315]), array([1.        , 0.        , 0.91553494]), array([1.        , 1.        , 0.50992272]), array([1.        , 0.        , 0.97492718]), array([1.        , 1.        , 0.20120141]), array([1.        , 1.        , 0.57630042]), array([1.        , 0.        , 0.83684829]), array([1.        , 1.        , 0.44427069]), array([1.       , 0.       , 0.9156467]), array([1.       , 0.       , 0.9925618]), array([1.        , 0.        , 0.93309831]), array([1.        , 1.        , 0.51923643]), array([1.        , 1.        , 0.54629569]), array([1.        , 0.        , 0.85359653]), array([1. , 1. , 0.5]), array([1.        , 0.        , 0.85454908]), array([1.        , 0.        , 0.82139467]), array([1.        , 0.        , 0.86156548]), array([1.        , 1.        , 0.35736046]), array([1.        , 0.        , 0.94132743]), array([1.       , 1.       , 0.3838089]), array([1.        , 1.        , 0.62376185]), array([1.       , 0.       , 0.9783068]), array([1.        , 0.        , 0.82394086]), array([1.        , 0.        , 0.82536648]), array([1.        , 1.        , 0.48990015]), array([1.        , 1.        , 0.50879651]), array([1.        , 0.        , 0.99744277]), array([1.        , 1.        , 0.61847806]), array([1.        , 1.        , 0.66301747]), array([1.        , 1.        , 0.05173015]), array([1.        , 0.        , 0.90658672]), array([1.        , 1.        , 0.11986451]), array([1.        , 1.        , 0.26475908]), array([1.        , 0.        , 0.86851381]), array([1.        , 1.        , 0.57223115]), array([1.        , 0.        , 0.98434471]), array([1.        , 1.        , 0.27130789]), array([1.        , 0.        , 0.97640451]), array([1.        , 0.        , 0.79893212]), array([1.        , 1.        , 0.12132718]), array([1.        , 1.        , 0.48917875]), array([1.        , 0.        , 0.93199847]), array([1.        , 0.        , 0.67711134]), array([1.        , 0.        , 0.95443873]), array([1.        , 0.        , 0.86154885]), array([1.        , 1.        , 0.57031588]), array([1.        , 1.        , 0.39292553]), array([1.        , 0.        , 0.91370303]), array([1.        , 1.        , 0.62522289]), array([1.        , 0.        , 0.95045426]), array([1.        , 0.        , 0.93223822]), array([1.       , 0.       , 0.9534925]), array([1.        , 1.        , 0.49030498]), array([1.        , 1.        , 0.15874136]), array([1.       , 0.       , 0.9568466]), array([1.        , 0.        , 0.90489472]), array([1.        , 0.        , 0.89195207]), array([1.        , 1.        , 0.19295718]), array([1.       , 0.       , 0.9049484]), array([1.        , 0.        , 0.97821142]), array([1.        , 1.        , 0.18034308]), array([1.       , 1.       , 0.4670336]), array([1.        , 0.        , 0.90493001]), array([1.        , 0.        , 0.91274053]), array([1.        , 0.        , 0.76759292]), array([1.        , 0.        , 0.88100324]), array([1.        , 0.        , 0.70858575]), array([1.        , 0.        , 0.96809713]), array([1.        , 1.        , 0.25623472]), array([1.        , 1.        , 0.42789209]), array([1.        , 1.        , 0.51496162]), array([1.        , 1.        , 0.27311078]), array([1.        , 0.        , 0.74160754]), array([1.        , 0.        , 0.90911803]), array([1.        , 0.        , 0.97745197]), array([1.        , 0.        , 0.73111283]), array([1.        , 1.        , 0.25401566]), array([1.        , 0.        , 0.80038811]), array([1.        , 0.        , 0.84486435]), array([1.        , 0.        , 0.81273766]), array([1.       , 1.       , 0.4143157]), array([1.        , 0.        , 0.98330288]), array([1.        , 0.        , 0.84189561]), array([1.        , 1.        , 0.65861372]), array([1.        , 0.        , 0.68416637]), array([1.        , 1.        , 0.26338818]), array([1.        , 0.        , 0.85521568]), array([1.        , 0.        , 0.91140133]), array([1.        , 0.        , 0.91663126]), array([1.        , 1.        , 0.47161534]), array([1.        , 1.        , 0.61649481]), array([1.        , 1.        , 0.19519685]), array([1.        , 1.        , 0.62244421]), array([1.        , 1.        , 0.33016224]), array([1.        , 1.        , 0.39252508]), array([1.        , 0.        , 0.97084279]), array([1.        , 0.        , 0.82919446]), array([1.        , 1.        , 0.19479181]), array([1.        , 0.        , 0.80859108]), array([1.        , 1.        , 0.36451345]), array([1.        , 0.        , 0.90486467]), array([1.        , 0.        , 0.90070253]), array([1.        , 1.        , 0.59918974]), array([1.        , 1.        , 0.63101137]), array([1.        , 1.        , 0.43356635]), array([1.        , 1.        , 0.60727571]), array([1.        , 0.        , 0.93188919]), array([1.        , 1.        , 0.43958606]), array([1.       , 0.       , 0.9561173]), array([1.        , 1.        , 0.58193976]), array([1.        , 1.        , 0.56010424]), array([1.        , 0.        , 0.97617032]), array([1.        , 1.        , 0.66181189]), array([1.        , 1.        , 0.36702097]), array([1.       , 0.       , 0.8503028]), array([1.        , 1.        , 0.60684687]), array([1.        , 0.        , 0.84582073]), array([1.        , 1.        , 0.31503833]), array([1.        , 1.        , 0.49267169]), array([1.       , 1.       , 0.5256675]), array([1.        , 1.        , 0.46372259]), array([1.        , 1.        , 0.64651464]), array([1.        , 0.        , 0.97559624]), array([1.        , 1.        , 0.59978265]), array([1.        , 0.        , 0.67151102]), array([1.        , 1.        , 0.62106022]), array([1.        , 0.        , 0.84418852]), array([1.        , 0.        , 0.98148452]), array([1.        , 1.        , 0.17928802]), array([1.        , 0.        , 0.93120041]), array([1.        , 0.        , 0.73221123]), array([1.        , 1.        , 0.16069224]), array([1.        , 0.        , 0.98542757]), array([1.        , 1.        , 0.51710526]), array([1.        , 1.        , 0.48353543]), array([1.        , 0.        , 0.86767782]), array([1.        , 0.        , 0.80916867]), array([1.      , 1.      , 0.537387]), array([1.        , 1.        , 0.48582831]), array([1.        , 0.        , 0.81213655]), array([1.        , 0.        , 0.86761718]), array([1.        , 1.        , 0.55206964]), array([1.        , 1.        , 0.16625407]), array([1.        , 1.        , 0.20201427]), array([1.        , 1.        , 0.35731825]), array([1.        , 0.        , 0.81140848]), array([1.        , 1.        , 0.42737397]), array([1.       , 1.       , 0.3193008]), array([1.        , 1.        , 0.39223758]), array([1.        , 0.        , 0.89287598]), array([1.        , 0.        , 0.97044084]), array([1.        , 0.        , 0.80247723]), array([1.        , 1.        , 0.39333572]), array([1.        , 1.        , 0.63398494]), array([1.        , 0.        , 0.76486422]), array([1.        , 1.        , 0.64526369]), array([1.        , 0.        , 0.89801044]), array([1.        , 1.        , 0.30075405]), array([1.       , 1.       , 0.6135859]), array([1.        , 0.        , 0.96229982]), array([1.        , 1.        , 0.46555119]), array([1.        , 1.        , 0.56093224]), array([1.        , 0.        , 0.95570991]), array([1.        , 0.        , 0.76686273]), array([1.        , 0.        , 0.99198041]), array([1.        , 1.        , 0.50629321]), array([1.       , 1.       , 0.3165974]), array([1.        , 0.        , 0.71690647]), array([1.       , 0.       , 0.8366378]), array([1.        , 0.        , 0.83544892]), array([1.        , 1.        , 0.56190176]), array([1.        , 0.        , 0.99246856]), array([1.       , 0.       , 0.9994375]), array([1.        , 0.        , 0.91158074]), array([1.        , 0.        , 0.77580362]), array([1.        , 0.        , 0.90744699]), array([1.        , 1.        , 0.01317484]), array([1.        , 1.        , 0.34913873]), array([1.        , 1.        , 0.16366242]), array([1.        , 1.        , 0.59592102]), array([1.        , 1.        , 0.18223045]), array([1.        , 1.        , 0.13442232]), array([1.        , 0.        , 0.96331652]), array([1.        , 0.        , 0.71900649]), array([1.       , 1.       , 0.3060441]), array([1.        , 0.        , 0.84011409]), array([1.        , 0.        , 0.98228381]), array([1.        , 1.        , 0.15307801]), array([1.        , 1.        , 0.55138766]), array([1.        , 0.        , 0.98658505]), array([1.        , 0.        , 0.87034296]), array([1.        , 1.        , 0.60601003]), array([1.        , 0.        , 0.71908752]), array([1.        , 0.        , 0.97919135]), array([1.       , 1.       , 0.6035181]), array([1.        , 0.        , 0.69468469]), array([1.        , 1.        , 0.55285526]), array([1.        , 1.        , 0.51987405]), array([1.        , 1.        , 0.60451711]), array([1.        , 0.        , 0.81669622]), array([1.        , 1.        , 0.38575586]), array([1.        , 0.        , 0.96921899]), array([1.        , 0.        , 0.97349542]), array([1.        , 0.        , 0.96960607]), array([1.        , 1.        , 0.27167958]), array([1.        , 1.        , 0.57849959]), array([1.        , 0.        , 0.99361466]), array([1.        , 0.        , 0.70198942]), array([1.        , 1.        , 0.64244436]), array([1.        , 0.        , 0.95200682]), array([1. , 1. , 0.5]), array([1.       , 1.       , 0.4369672]), array([1.        , 0.        , 0.79360665]), array([1.       , 0.       , 0.9636843]), array([1.        , 0.        , 0.93982158]), array([1.        , 1.        , 0.27485904]), array([1.        , 0.        , 0.87877758]), array([1.        , 0.        , 0.93931666]), array([1.        , 0.        , 0.98469229]), array([1.        , 1.        , 0.10905429]), array([1.        , 1.        , 0.07264591]), array([1.        , 0.        , 0.98034404]), array([1.       , 1.       , 0.2599554]), array([1.        , 1.        , 0.32419301]), array([1.       , 1.       , 0.5173413]), array([1.        , 0.        , 0.74850996]), array([1.       , 0.       , 0.7650425]), array([1.        , 1.        , 0.46038158]), array([1.        , 1.        , 0.51568752]), array([1.        , 0.        , 0.92759828]), array([1.       , 1.       , 0.6630629]), array([1.        , 1.        , 0.21906146]), array([1.        , 0.        , 0.81886619]), array([1.        , 1.        , 0.41840078]), array([1.        , 0.        , 0.88244818]), array([1.        , 1.        , 0.43550861]), array([1.        , 0.        , 0.87451477]), array([1.        , 0.        , 0.96518947]), array([1.        , 0.        , 0.96781716]), array([1.        , 1.        , 0.07930574]), array([1.        , 1.        , 0.58676761]), array([1.        , 0.        , 0.76104113]), array([1.       , 1.       , 0.4631966]), array([1.        , 0.        , 0.82169876]), array([1.        , 1.        , 0.58455519]), array([1.        , 0.        , 0.89765205]), array([1.        , 1.        , 0.46467177]), array([1.       , 0.       , 0.8526917]), array([1.        , 1.        , 0.45287444]), array([1.       , 0.       , 0.8192429]), array([1.       , 0.       , 0.7885061]), array([1.        , 0.        , 0.68349099]), array([1.       , 1.       , 0.1854671]), array([1.        , 0.        , 0.92316977]), array([1.       , 1.       , 0.3524352]), array([1.        , 0.        , 0.99860392]), array([1.        , 0.        , 0.84479428]), array([1.        , 0.        , 0.94718828]), array([1.        , 0.        , 0.95796528]), array([1.        , 0.        , 0.75384162]), array([1.        , 1.        , 0.58791762]), array([1.        , 1.        , 0.51051348]), array([1.        , 0.        , 0.72955459]), array([1.        , 0.        , 0.89283532]), array([1.        , 1.        , 0.55512451]), array([1.        , 1.        , 0.19252465]), array([1.        , 1.        , 0.58527275]), array([1.        , 1.        , 0.45435488]), array([1.        , 0.        , 0.89329944]), array([1.        , 1.        , 0.64727453]), array([1.       , 0.       , 0.9697519]), array([1.        , 0.        , 0.97159879]), array([1.        , 1.        , 0.66529855]), array([1.        , 0.        , 0.97594313]), array([1.        , 0.        , 0.84272884]), array([1.        , 0.        , 0.93010464]), array([1.        , 0.        , 0.97506132]), array([1.        , 1.        , 0.62705431]), array([1.        , 1.        , 0.28647693]), array([1.        , 1.        , 0.27033931]), array([1.        , 1.        , 0.56890829]), array([1.        , 0.        , 0.71245729]), array([1.        , 0.        , 0.72054523]), array([1.       , 0.       , 0.8406625]), array([1.        , 0.        , 0.92890876]), array([1.        , 0.        , 0.87643889]), array([1.        , 0.        , 0.91341191]), array([1.        , 1.        , 0.41286777]), array([1.        , 1.        , 0.14295568]), array([1.       , 0.       , 0.9813647]), array([1.        , 0.        , 0.90353028]), array([1.        , 0.        , 0.92805897]), array([1.        , 1.        , 0.19277785]), array([1. , 1. , 0.5]), array([1.        , 1.        , 0.63726579]), array([1.        , 1.        , 0.58164117]), array([1.        , 0.        , 0.80053692]), array([1.        , 1.        , 0.65910492]), array([1.        , 1.        , 0.56696981]), array([1.        , 0.        , 0.75894744]), array([1.        , 1.        , 0.66038726]), array([1.        , 0.        , 0.97204005]), array([1.        , 0.        , 0.79507189]), array([1.        , 0.        , 0.86931142]), array([1.        , 1.        , 0.02766421]), array([1.        , 0.        , 0.71701124]), array([1.        , 0.        , 0.94306724]), array([1.        , 1.        , 0.21470815]), array([1.        , 0.        , 0.85694018]), array([1.        , 1.        , 0.09153688]), array([1.        , 0.        , 0.77950769]), array([1.        , 1.        , 0.31131353]), array([1.        , 1.        , 0.34580819]), array([1.       , 0.       , 0.7860371]), array([1.        , 1.        , 0.53920055]), array([1.        , 0.        , 0.92758581]), array([1.        , 0.        , 0.82871852]), array([1.        , 1.        , 0.43223994]), array([1.        , 1.        , 0.15859836]), array([1.        , 0.        , 0.81966337]), array([1.        , 1.        , 0.01303229]), array([1.        , 0.        , 0.82446226]), array([1.        , 0.        , 0.94595527]), array([1.        , 0.        , 0.87830709]), array([1.        , 0.        , 0.82307222]), array([1.        , 0.        , 0.72406568]), array([1.        , 1.        , 0.19741448]), array([1.       , 1.       , 0.6447033]), array([1.        , 1.        , 0.56069732]), array([1.        , 1.        , 0.47892012]), array([1.        , 1.        , 0.19630569]), array([1.        , 1.        , 0.48821789]), array([1.        , 0.        , 0.99037387]), array([1.        , 1.        , 0.34018349]), array([1.        , 1.        , 0.37928261]), array([1.       , 0.       , 0.8934674]), array([1.        , 1.        , 0.29628521]), array([1.       , 0.       , 0.8168144]), array([1.        , 1.        , 0.29408121]), array([1.        , 0.        , 0.98634586]), array([1.        , 0.        , 0.76472718]), array([1.        , 1.        , 0.61956064]), array([1.        , 1.        , 0.46720631]), array([1.        , 0.        , 0.72279974]), array([1.        , 1.        , 0.23248985]), array([1.        , 0.        , 0.92680459]), array([1.        , 1.        , 0.41224875]), array([1.       , 0.       , 0.8637605])]\n",
            "\n",
            "\n",
            "pullover_2\n",
            "[array([0.        , 0.        , 0.95353829]), array([0.        , 0.        , 0.99693415]), array([0.        , 0.        , 0.99991678]), array([0.        , 0.        , 0.99498426]), array([0.        , 0.        , 0.96649154]), array([0.        , 0.        , 0.99097818]), array([0.       , 0.       , 0.9795637]), array([0.        , 0.        , 0.99676889]), array([0.       , 0.       , 0.9375742]), array([0.        , 0.        , 0.99648042]), array([0.        , 0.        , 0.97144656]), array([0.        , 0.        , 0.99794784]), array([0.       , 0.       , 0.9979726]), array([0.        , 0.        , 0.99181296]), array([0.        , 0.        , 0.99828384]), array([0.        , 0.        , 0.99843068]), array([0.        , 0.        , 0.98340278]), array([0.        , 0.        , 0.98825672]), array([0.        , 0.        , 0.99973668]), array([0.        , 0.        , 0.99045587]), array([0.      , 0.      , 0.988915]), array([0.        , 0.        , 0.99036782]), array([0.        , 0.        , 0.99245066]), array([0.       , 0.       , 0.8855089]), array([0.        , 0.        , 0.99839448]), array([0.       , 0.       , 0.9892451]), array([0.        , 0.        , 0.91126439]), array([0.        , 0.        , 0.98094847]), array([0.        , 0.        , 0.95445279]), array([0.        , 0.        , 0.89412545]), array([0.        , 0.        , 0.95581463]), array([0.        , 0.        , 0.88446578]), array([0.        , 0.        , 0.99034882]), array([0.        , 0.        , 0.94959093]), array([0.        , 0.        , 0.98182396]), array([0.        , 0.        , 0.73482929]), array([0.        , 0.        , 0.85086634]), array([0.        , 0.        , 0.92637571]), array([0.        , 0.        , 0.94954752]), array([0.        , 0.        , 0.98350712]), array([0.        , 0.        , 0.99996366]), array([0.        , 0.        , 0.99681128]), array([0.        , 0.        , 0.99377411]), array([0.        , 0.        , 0.97993785]), array([0.        , 0.        , 0.99629932]), array([0.        , 0.        , 0.98910149]), array([0.        , 0.        , 0.99684613]), array([0.        , 0.        , 0.67972068]), array([0.        , 0.        , 0.98860493]), array([0.        , 0.        , 0.99768786]), array([0.       , 0.       , 0.9913354]), array([0.        , 0.        , 0.97572111]), array([0.        , 0.        , 0.98627562]), array([0.        , 0.        , 0.99694102]), array([0.       , 0.       , 0.9189303]), array([0.        , 0.        , 0.96095712]), array([0.        , 0.        , 0.98722319]), array([0.        , 0.        , 0.99743142]), array([0.        , 0.        , 0.99752239]), array([0.        , 0.        , 0.98076529]), array([0.        , 0.        , 0.88648339]), array([0.        , 0.        , 0.99185641]), array([0.        , 0.        , 0.97468774]), array([0.        , 0.        , 0.95904836]), array([0.        , 0.        , 0.99587492]), array([0.        , 0.        , 0.97750125]), array([0.        , 0.        , 0.98664785]), array([0.        , 0.        , 0.98809212]), array([0.        , 0.        , 0.93196758]), array([0.        , 0.        , 0.98920645]), array([0.        , 0.        , 0.76003834]), array([0.        , 0.        , 0.99155478]), array([0.       , 0.       , 0.9835089]), array([0.        , 0.        , 0.97367413]), array([0.        , 0.        , 0.99188698]), array([0.        , 0.        , 0.85814558]), array([0.        , 0.        , 0.99810801]), array([0.        , 0.        , 0.93956874]), array([0.        , 0.        , 0.99737494]), array([0.        , 0.        , 0.99028122]), array([0.        , 0.        , 0.99880425]), array([0.        , 0.        , 0.89341918]), array([0.        , 0.        , 0.99775149]), array([0.        , 0.        , 0.99365138]), array([0.       , 0.       , 0.9956587]), array([0.        , 0.        , 0.94077711]), array([0.        , 0.        , 0.99146726]), array([0.        , 0.        , 0.96776125]), array([0.        , 0.        , 0.91748079]), array([0.        , 0.        , 0.95450393]), array([0.        , 0.        , 0.99706169]), array([0.        , 0.        , 0.97500568]), array([0.        , 0.        , 0.84147048]), array([0.        , 0.        , 0.99719581]), array([0.        , 0.        , 0.99763768]), array([0.        , 0.        , 0.95207475]), array([0.        , 0.        , 0.98410771]), array([0.        , 0.        , 0.98589635]), array([0.        , 0.        , 0.99554076]), array([0.        , 0.        , 0.93761447]), array([0.        , 0.        , 0.75479436]), array([0.        , 0.        , 0.98390005]), array([0.       , 0.       , 0.9953143]), array([0.        , 0.        , 0.99521958]), array([0.        , 0.        , 0.99793379]), array([0.        , 0.        , 0.91916424]), array([0.        , 0.        , 0.99256139]), array([0.        , 0.        , 0.98847104]), array([0.        , 0.        , 0.84418076]), array([0.        , 0.        , 0.92575473]), array([0.        , 0.        , 0.99391632]), array([0.        , 0.        , 0.96858586]), array([0.        , 0.        , 0.89569073]), array([0.        , 0.        , 0.99175012]), array([0.        , 0.        , 0.98450632]), array([0.        , 0.        , 0.99650259]), array([0.        , 0.        , 0.98842088]), array([0.        , 0.        , 0.99242676]), array([0.        , 0.        , 0.99619273]), array([0.        , 0.        , 0.99331543]), array([0.        , 0.        , 0.90382122]), array([0.        , 0.        , 0.99757004]), array([0.        , 0.        , 0.97679189]), array([0.        , 0.        , 0.97455689]), array([0.        , 0.        , 0.98273259]), array([0.        , 0.        , 0.94664297]), array([0.        , 0.        , 0.99852325]), array([0.        , 0.        , 0.99154458]), array([0.        , 0.        , 0.96664044]), array([0.        , 0.        , 0.99881237]), array([0.        , 0.        , 0.99298553]), array([0.        , 0.        , 0.99977873]), array([0.        , 0.        , 0.95806456]), array([0.        , 0.        , 0.99453108]), array([0.        , 0.        , 0.99101834]), array([0.        , 0.        , 0.99823458]), array([0.        , 0.        , 0.98863116]), array([0.        , 0.        , 0.88538285]), array([0.        , 0.        , 0.99389242]), array([0.        , 0.        , 0.92315914]), array([0.        , 0.        , 0.97790833]), array([0.        , 0.        , 0.90834355]), array([0.        , 1.        , 0.57076621]), array([0.        , 0.        , 0.91986114]), array([0.        , 0.        , 0.98963924]), array([0.        , 0.        , 0.99097615]), array([0.        , 0.        , 0.97424542]), array([0.        , 0.        , 0.99981041]), array([0.        , 0.        , 0.90970284]), array([0.        , 0.        , 0.99963281]), array([0.        , 0.        , 0.87652201]), array([0.        , 0.        , 0.87575089]), array([0.        , 0.        , 0.96482668]), array([0.        , 0.        , 0.97513951]), array([0.        , 0.        , 0.93736696]), array([0.        , 0.        , 0.98678876]), array([0.        , 0.        , 0.99496103]), array([0.       , 0.       , 0.8813847]), array([0.        , 0.        , 0.98442073]), array([0.        , 1.        , 0.57654202]), array([0.        , 0.        , 0.98644255]), array([0.        , 0.        , 0.99429603]), array([0.        , 0.        , 0.91470097]), array([0.      , 0.      , 0.999478]), array([0.        , 0.        , 0.99887602]), array([0.        , 0.        , 0.86278489]), array([0.        , 0.        , 0.98048115]), array([0.        , 0.        , 0.99446965]), array([0.        , 0.        , 0.99045866]), array([0.       , 0.       , 0.8990819]), array([0.        , 0.        , 0.99900997]), array([0.        , 0.        , 0.99861456]), array([0.        , 0.        , 0.96436291]), array([0.        , 0.        , 0.98047434]), array([0.        , 0.        , 0.99965979]), array([0.        , 0.        , 0.99552004]), array([0.        , 0.        , 0.99723981]), array([0.        , 0.        , 0.95319628]), array([0.        , 0.        , 0.89978854]), array([0.        , 0.        , 0.90161294]), array([0.       , 0.       , 0.9766347]), array([0.        , 0.        , 0.93739211]), array([0.        , 0.        , 0.80847938]), array([0.        , 0.        , 0.99900185]), array([0.        , 0.        , 0.99936999]), array([0.       , 0.       , 0.9965811]), array([0.        , 0.        , 0.89007332]), array([0.        , 0.        , 0.94087745]), array([0.        , 0.        , 0.99635488]), array([0.        , 0.        , 0.69221939]), array([0.        , 0.        , 0.98698018]), array([0.        , 0.        , 0.98212695]), array([0.        , 0.        , 0.99669602]), array([0.        , 0.        , 0.99173967]), array([0.        , 0.        , 0.99553657]), array([0.        , 0.        , 0.95813366]), array([0.        , 0.        , 0.99935171]), array([0.        , 0.        , 0.97768525]), array([0.        , 0.        , 0.96612384]), array([0.        , 0.        , 0.90724386]), array([0.        , 0.        , 0.99836197]), array([0.        , 0.        , 0.99215036]), array([0.        , 0.        , 0.98937846]), array([0.        , 0.        , 0.98700506]), array([0.        , 0.        , 0.99569227]), array([0.        , 0.        , 0.98211933]), array([0.       , 0.       , 0.9909084]), array([0.        , 0.        , 0.95930696]), array([0.        , 0.        , 0.97673994]), array([0.        , 0.        , 0.94006786]), array([0.        , 0.        , 0.91826677]), array([0.        , 0.        , 0.99811051]), array([0.        , 0.        , 0.98977163]), array([0.        , 0.        , 0.96868774]), array([0.        , 1.        , 0.31073866]), array([0.       , 0.       , 0.9939237]), array([0.        , 0.        , 0.89321239]), array([0.        , 0.        , 0.98290025]), array([0.        , 0.        , 0.99714041]), array([0.        , 0.        , 0.75621445]), array([0.        , 0.        , 0.97604855]), array([0.        , 0.        , 0.99610584]), array([0.        , 0.        , 0.95524374]), array([0.        , 0.        , 0.99156505]), array([0.        , 0.        , 0.97103323]), array([0.        , 0.        , 0.98582767]), array([0.       , 0.       , 0.8099072]), array([0.        , 0.        , 0.98117085]), array([0.       , 0.       , 0.9873123]), array([0.        , 0.        , 0.92637086]), array([0.        , 0.        , 0.99716356]), array([0.        , 0.        , 0.95899634]), array([0.        , 0.        , 0.99300203]), array([0.        , 0.        , 0.96165291]), array([0.        , 0.        , 0.99920786]), array([0.        , 0.        , 0.99883214]), array([0.        , 0.        , 0.99673965]), array([0.        , 0.        , 0.99625072]), array([0.       , 0.       , 0.9689209]), array([0.        , 0.        , 0.99456456]), array([0.       , 0.       , 0.9982659]), array([0.        , 0.        , 0.94003331]), array([0.        , 0.        , 0.97840327]), array([0.        , 0.        , 0.76077565]), array([0.        , 0.        , 0.81658358]), array([0.        , 0.        , 0.91466395]), array([0.        , 0.        , 0.93191642]), array([0.        , 0.        , 0.98698839]), array([0.        , 0.        , 0.99764198]), array([0.        , 0.        , 0.81027108]), array([0.        , 0.        , 0.68270522]), array([0.        , 0.        , 0.85278208]), array([0.        , 0.        , 0.97587591]), array([0.        , 0.        , 0.99380329]), array([0.        , 0.        , 0.99804946]), array([0.        , 0.        , 0.95665637]), array([0.        , 0.        , 0.99679921]), array([0.        , 0.        , 0.99241123]), array([0.        , 1.        , 0.39762205]), array([0.        , 0.        , 0.99683746]), array([0.        , 0.        , 0.99285497]), array([0.        , 0.        , 0.96795008]), array([0.        , 0.        , 0.96190417]), array([0.        , 0.        , 0.99760082]), array([0.       , 0.       , 0.9528466]), array([0.        , 0.        , 0.99490728]), array([0.       , 0.       , 0.9993853]), array([0.        , 0.        , 0.97374409]), array([0.        , 0.        , 0.91963815]), array([0.        , 0.        , 0.99170116]), array([0.        , 0.        , 0.98484436]), array([0.        , 0.        , 0.90960756]), array([0.        , 0.        , 0.99409919]), array([0.        , 0.        , 0.94528109]), array([0.        , 0.        , 0.93365217]), array([0.        , 0.        , 0.92056689]), array([0.        , 0.        , 0.95860909]), array([0.        , 0.        , 0.93476527]), array([0.       , 0.       , 0.9833614]), array([0.        , 0.        , 0.89168175]), array([0.        , 0.        , 0.99482066]), array([0.        , 0.        , 0.99301277]), array([0.        , 0.        , 0.97271546]), array([0.        , 0.        , 0.99761396]), array([0.        , 0.        , 0.99424722]), array([0.        , 0.        , 0.98635129]), array([0.        , 0.        , 0.99499362]), array([0.        , 0.        , 0.96841749]), array([0.        , 0.        , 0.99247033]), array([0.        , 0.        , 0.99956553]), array([0.        , 0.        , 0.99645486]), array([0.        , 0.        , 0.98200043]), array([0.        , 0.        , 0.99689296]), array([0.        , 0.        , 0.98490262]), array([0.        , 0.        , 0.87543134]), array([0.        , 0.        , 0.96802686]), array([0.        , 0.        , 0.99259495]), array([0.        , 0.        , 0.98057863]), array([0.        , 1.        , 0.66722897]), array([0.        , 0.        , 0.87555042]), array([0.       , 0.       , 0.9921328]), array([0.        , 0.        , 0.99348894]), array([0.        , 0.        , 0.95631975]), array([0.        , 0.        , 0.99913767]), array([0.        , 0.        , 0.73855683]), array([0.        , 0.        , 0.94893194]), array([0.        , 0.        , 0.97490494]), array([0.      , 0.      , 0.994682]), array([0.        , 0.        , 0.98827969]), array([0.      , 0.      , 0.999664]), array([0.        , 0.        , 0.99980582]), array([0.        , 0.        , 0.98209726]), array([0.        , 0.        , 0.96570581]), array([0.        , 0.        , 0.99795583]), array([0.        , 0.        , 0.98007076]), array([0.       , 0.       , 0.7596698]), array([0.        , 0.        , 0.95280282]), array([0.        , 0.        , 0.99809524]), array([0.       , 0.       , 0.9751892]), array([0.        , 0.        , 0.99214696]), array([0.        , 0.        , 0.99319032]), array([0.        , 0.        , 0.94990868]), array([0.        , 0.        , 0.99497939]), array([0.        , 0.        , 0.99889485]), array([0.        , 0.        , 0.98680029]), array([0.        , 0.        , 0.99905869]), array([0.        , 0.        , 0.96174162]), array([0.       , 0.       , 0.9837571]), array([0.        , 0.        , 0.99981706]), array([0.        , 0.        , 0.94171272]), array([0.        , 0.        , 0.98555315]), array([0.        , 0.        , 0.97902227]), array([0.        , 0.        , 0.95214798]), array([0.        , 0.        , 0.99586396]), array([0.        , 0.        , 0.97669874]), array([0.       , 0.       , 0.9758028]), array([0.        , 0.        , 0.93451867]), array([0.        , 0.        , 0.96145967]), array([0.        , 0.        , 0.96997425]), array([0.        , 0.        , 0.91148481]), array([0.        , 0.        , 0.99972105]), array([0.        , 0.        , 0.96561431]), array([0.        , 0.        , 0.96739539]), array([0.        , 0.        , 0.98213405]), array([0.        , 1.        , 0.63230761]), array([0.        , 0.        , 0.98693345]), array([0.        , 0.        , 0.97765494]), array([0.        , 0.        , 0.87424359]), array([0.        , 0.        , 0.99969265]), array([0.        , 1.        , 0.61181041]), array([0.        , 0.        , 0.99043649]), array([0.        , 0.        , 0.97109877]), array([0.        , 0.        , 0.97178686]), array([0.        , 0.        , 0.96849485]), array([0.        , 0.        , 0.98916168]), array([0.        , 0.        , 0.99969556]), array([0.        , 0.        , 0.99397241]), array([0.        , 0.        , 0.95816905]), array([0.        , 0.        , 0.94884363]), array([0.        , 0.        , 0.99816534]), array([0.       , 0.       , 0.9997304]), array([0.        , 1.        , 0.53617226]), array([0.        , 0.        , 0.99713326]), array([0.        , 0.        , 0.91067128]), array([0.        , 0.        , 0.99626759]), array([0.        , 0.        , 0.94726876]), array([0.        , 0.        , 0.98661127]), array([0.        , 0.        , 0.78379894]), array([0.        , 0.        , 0.72577264]), array([0.        , 0.        , 0.92064102]), array([0.        , 1.        , 0.34580246]), array([0.        , 0.        , 0.87462021]), array([0.        , 0.        , 0.97777418]), array([0.        , 0.        , 0.94511986]), array([0.        , 0.        , 0.77761105]), array([0.       , 0.       , 0.9935597]), array([0.        , 0.        , 0.99892774]), array([0.        , 0.        , 0.97689306]), array([0.        , 0.        , 0.98242462]), array([0.        , 0.        , 0.99121239]), array([0.        , 0.        , 0.88041255]), array([0.        , 0.        , 0.96316401]), array([0.        , 0.        , 0.99590697]), array([0.        , 0.        , 0.98205502]), array([0.        , 0.        , 0.99216853]), array([0.       , 0.       , 0.9607025]), array([0.       , 0.       , 0.9847945]), array([0.        , 0.        , 0.98355071]), array([0.        , 0.        , 0.99614705]), array([0.        , 0.        , 0.98274093]), array([0.        , 0.        , 0.99540879]), array([0.        , 0.        , 0.94118806]), array([0.        , 0.        , 0.98612229]), array([0.        , 0.        , 0.98907749]), array([0.        , 0.        , 0.95447922]), array([0.        , 0.        , 0.98690481]), array([0.        , 0.        , 0.99298787]), array([0.        , 0.        , 0.99819501]), array([0.        , 0.        , 0.92162994]), array([0.        , 0.        , 0.92991271]), array([0.       , 0.       , 0.6961923]), array([0.        , 0.        , 0.89805003]), array([0.        , 0.        , 0.99336246]), array([0.        , 0.        , 0.98846501]), array([0.        , 0.        , 0.94955129]), array([0.        , 0.        , 0.98901842]), array([0.        , 0.        , 0.88516883]), array([0.        , 0.        , 0.99737014]), array([0.        , 0.        , 0.82482852]), array([0.        , 0.        , 0.97990005]), array([0.        , 0.        , 0.83180436]), array([0.        , 0.        , 0.99896252]), array([0.       , 0.       , 0.9949732]), array([0.        , 0.        , 0.95241807]), array([0.       , 0.       , 0.9937841]), array([0.        , 0.        , 0.86532193]), array([0.        , 0.        , 0.68365419]), array([0.        , 0.        , 0.98298483]), array([0.        , 0.        , 0.98723773]), array([0.        , 0.        , 0.98795501]), array([0.        , 0.        , 0.97262439]), array([0.        , 0.        , 0.94881923]), array([0.        , 0.        , 0.93626262]), array([0.        , 0.        , 0.97216436]), array([0.        , 1.        , 0.65952453]), array([0.        , 0.        , 0.96141587]), array([0.        , 0.        , 0.96520629]), array([0.        , 0.        , 0.91225563]), array([0.        , 0.        , 0.96154924]), array([0.        , 0.        , 0.76042194]), array([0.        , 0.        , 0.97525704]), array([0.        , 0.        , 0.98773753]), array([0.        , 0.        , 0.99301076]), array([0.        , 0.        , 0.93639121]), array([0.        , 0.        , 0.99729701]), array([0.        , 0.        , 0.99972178]), array([0.        , 0.        , 0.94792072]), array([0.        , 0.        , 0.99306019]), array([0.        , 0.        , 0.98699622]), array([0.        , 0.        , 0.99264131]), array([0.        , 0.        , 0.95141552]), array([0.        , 0.        , 0.99694884]), array([0.        , 0.        , 0.99784293]), array([0.        , 0.        , 0.99943001]), array([0.        , 0.        , 0.96891045]), array([0.        , 0.        , 0.95990483]), array([0.        , 0.        , 0.98552632]), array([0.        , 0.        , 0.96292196]), array([0.        , 0.        , 0.98795766]), array([0.        , 0.        , 0.99948686]), array([0.        , 0.        , 0.99787856]), array([0.        , 0.        , 0.98395139]), array([0.        , 0.        , 0.96968187]), array([0.        , 0.        , 0.96359073]), array([0.        , 0.        , 0.98619756]), array([0.        , 0.        , 0.99906642]), array([0.       , 0.       , 0.9597935]), array([0.        , 0.        , 0.92292678]), array([0.        , 0.        , 0.94890758]), array([0.      , 0.      , 0.991316]), array([0.        , 0.        , 0.90231435]), array([0.        , 0.        , 0.98405623]), array([0.        , 0.        , 0.99793338]), array([0.        , 0.        , 0.94448116]), array([0.        , 0.        , 0.99350011]), array([0.        , 0.        , 0.99339095]), array([0.        , 0.        , 0.93624171]), array([0.        , 0.        , 0.98378867]), array([0.        , 0.        , 0.99751275]), array([0.        , 0.        , 0.99367436]), array([0.        , 0.        , 0.89874286]), array([0.        , 0.        , 0.99207656]), array([0.        , 0.        , 0.99279314]), array([0.        , 0.        , 0.99247544]), array([0.        , 0.        , 0.97743612]), array([0.        , 0.        , 0.97370977]), array([0.        , 0.        , 0.82866461]), array([0.        , 0.        , 0.98520986]), array([0.        , 0.        , 0.96842865]), array([0.        , 0.        , 0.99773512]), array([0.        , 0.        , 0.99970916]), array([0.        , 0.        , 0.98085097]), array([0.        , 0.        , 0.99973381]), array([0.        , 0.        , 0.93809696]), array([0.        , 0.        , 0.93072861]), array([0.        , 0.        , 0.91535242]), array([0.        , 0.        , 0.99601026]), array([0.        , 0.        , 0.93543859]), array([0.        , 0.        , 0.99818619]), array([0.       , 0.       , 0.9467171]), array([0.       , 0.       , 0.8056418]), array([0.        , 0.        , 0.92287107]), array([0.        , 0.        , 0.99739161]), array([0.        , 0.        , 0.83311872]), array([0.        , 0.        , 0.99798807]), array([0.        , 0.        , 0.89411436]), array([0.        , 0.        , 0.91618147]), array([0.       , 0.       , 0.9702318]), array([0.        , 0.        , 0.74812653]), array([0.        , 0.        , 0.99420133]), array([0.        , 0.        , 0.97933154]), array([0.        , 0.        , 0.87779437]), array([0.        , 0.        , 0.70083826]), array([0.        , 0.        , 0.86596725]), array([0.        , 0.        , 0.99307781]), array([0.        , 1.        , 0.59397887]), array([0.        , 0.        , 0.99075653]), array([0.        , 0.        , 0.69609554]), array([0.        , 0.        , 0.94387962]), array([0.        , 0.        , 0.99244104]), array([0.        , 0.        , 0.95052963]), array([0.        , 0.        , 0.96952747]), array([0.        , 0.        , 0.99934791]), array([0.        , 0.        , 0.94093887]), array([0.        , 0.        , 0.91840071]), array([0.        , 0.        , 0.99802603]), array([0.        , 0.        , 0.98292575]), array([0.        , 0.        , 0.98767288]), array([0.        , 0.        , 0.94258834]), array([0.        , 0.        , 0.98889854]), array([0.        , 0.        , 0.98589665]), array([0.        , 0.        , 0.97294836]), array([0.       , 1.       , 0.5769203]), array([0.        , 0.        , 0.99618786]), array([0.        , 0.        , 0.96918119]), array([0.        , 0.        , 0.99269698]), array([0.        , 0.        , 0.99814783]), array([0.        , 0.        , 0.99102407]), array([0.        , 0.        , 0.97273841]), array([0.        , 0.        , 0.95417941]), array([0.        , 0.        , 0.95679444]), array([0.        , 0.        , 0.99580465]), array([0.        , 0.        , 0.98745647]), array([0.        , 0.        , 0.99179711]), array([0.        , 0.        , 0.86563355]), array([0.       , 0.       , 0.9870547]), array([0.        , 0.        , 0.92610125]), array([0.        , 0.        , 0.94205207]), array([0.        , 0.        , 0.99452868]), array([0.        , 0.        , 0.93391691]), array([0.        , 0.        , 0.98984274]), array([0.        , 0.        , 0.97565109]), array([0.       , 0.       , 0.9937399]), array([0.        , 0.        , 0.95509717]), array([0.        , 0.        , 0.95043933]), array([0.        , 0.        , 0.99902493]), array([0.        , 0.        , 0.99404658]), array([0.        , 0.        , 0.93449628]), array([0.        , 0.        , 0.93360002]), array([0.        , 0.        , 0.99628554]), array([0.        , 0.        , 0.98506822]), array([0.        , 0.        , 0.97803748]), array([0.        , 0.        , 0.99332321]), array([0.        , 0.        , 0.95562391]), array([0.       , 0.       , 0.9918249]), array([0.        , 0.        , 0.93290555]), array([0.        , 0.        , 0.99967583]), array([0.        , 0.        , 0.99923486]), array([0.        , 0.        , 0.99902673]), array([0.        , 0.        , 0.96910718]), array([0.        , 0.        , 0.79862488]), array([0.        , 0.        , 0.95551776]), array([0.        , 0.        , 0.80606586]), array([0.        , 0.        , 0.98537421]), array([0.       , 0.       , 0.9941705]), array([0.        , 0.        , 0.90298121]), array([0.       , 0.       , 0.9791367]), array([0.        , 0.        , 0.99876266]), array([0.       , 0.       , 0.9991697]), array([0.        , 0.        , 0.96287124]), array([0.        , 0.        , 0.93540376]), array([0.       , 0.       , 0.9854036]), array([0.        , 0.        , 0.99696837]), array([0.        , 0.        , 0.95332583]), array([0.        , 0.        , 0.99829818]), array([0.        , 0.        , 0.98891711]), array([0.        , 0.        , 0.97870176]), array([0.        , 0.        , 0.94313779]), array([0.        , 0.        , 0.98800311]), array([0.        , 0.        , 0.95124153]), array([0.       , 0.       , 0.9656211]), array([0.        , 0.        , 0.99177802]), array([0.        , 0.        , 0.99768735]), array([0.        , 0.        , 0.98695539]), array([0.        , 0.        , 0.99813947]), array([0.        , 0.        , 0.99861303]), array([0.        , 0.        , 0.99760416]), array([0.        , 0.        , 0.99944053]), array([0.      , 0.      , 0.682531]), array([0.       , 0.       , 0.8179624]), array([0.        , 0.        , 0.99371073]), array([0.        , 0.        , 0.68432985]), array([0.        , 0.        , 0.93189751]), array([0.        , 0.        , 0.99970158]), array([0.        , 0.        , 0.99731023]), array([0.        , 0.        , 0.75297967]), array([0.        , 0.        , 0.99774028]), array([0.        , 0.        , 0.99450915]), array([0.        , 0.        , 0.85155516]), array([0.        , 0.        , 0.97969295]), array([0.        , 0.        , 0.98809166]), array([0.        , 0.        , 0.99046689]), array([0.        , 0.        , 0.99355499]), array([0.        , 0.        , 0.96753742]), array([0.        , 0.        , 0.99097373]), array([0.        , 0.        , 0.98576083]), array([0.       , 1.       , 0.6508155]), array([0.        , 0.        , 0.99594954]), array([0.        , 0.        , 0.99788673]), array([0.        , 0.        , 0.99089857]), array([0.        , 0.        , 0.97045614]), array([0.        , 0.        , 0.99891726]), array([0.        , 0.        , 0.94806657]), array([0.        , 0.        , 0.78009632]), array([0.        , 0.        , 0.99153735]), array([0.        , 0.        , 0.96988138]), array([0.        , 0.        , 0.88729753]), array([0.        , 0.        , 0.96935369]), array([0.        , 0.        , 0.99540907]), array([0.        , 0.        , 0.96565217]), array([0.        , 0.        , 0.99421035]), array([0.        , 0.        , 0.96486717]), array([0.        , 0.        , 0.94006512]), array([0.        , 0.        , 0.99720527]), array([0.        , 0.        , 0.97415954]), array([0.        , 0.        , 0.97574284]), array([0.        , 0.        , 0.99877105]), array([0.        , 1.        , 0.61579724]), array([0.        , 0.        , 0.94276662]), array([0.       , 0.       , 0.9962949]), array([0.        , 0.        , 0.90280264]), array([0.        , 0.        , 0.97830232]), array([0.        , 0.        , 0.98187875]), array([0.        , 0.        , 0.99407683]), array([0.        , 0.        , 0.97453033]), array([0.        , 0.        , 0.95944476]), array([0.        , 0.        , 0.99147404]), array([0.        , 0.        , 0.99369459]), array([0.        , 0.        , 0.91266729]), array([0.       , 0.       , 0.9876513]), array([0.        , 0.        , 0.97940557]), array([0.        , 0.        , 0.97967337]), array([0.      , 0.      , 0.932849]), array([0.        , 0.        , 0.99111083]), array([0.        , 0.        , 0.99966176]), array([0.        , 0.        , 0.98006676]), array([0.        , 0.        , 0.93777125]), array([0.        , 0.        , 0.99731424]), array([0.       , 0.       , 0.9351842]), array([0.        , 0.        , 0.99669239]), array([0.        , 0.        , 0.99102748]), array([0.        , 0.        , 0.99014558]), array([0.        , 0.        , 0.87665074]), array([0.        , 0.        , 0.97650357]), array([0.        , 0.        , 0.97197901]), array([0.        , 0.        , 0.97505903]), array([0.        , 0.        , 0.99829181]), array([0.        , 0.        , 0.99724262]), array([0.       , 0.       , 0.9861279]), array([0.        , 0.        , 0.99938463]), array([0.        , 0.        , 0.99505044]), array([0.       , 0.       , 0.9482812]), array([0.        , 0.        , 0.99706986]), array([0.        , 0.        , 0.98681839]), array([0.        , 0.        , 0.99853881]), array([0.        , 0.        , 0.99248986]), array([0.        , 0.        , 0.84434814]), array([0.     , 0.     , 0.84607]), array([0.        , 0.        , 0.99884682]), array([0.        , 0.        , 0.97925194]), array([0.        , 0.        , 0.99835707]), array([0.        , 0.        , 0.99428776]), array([0.        , 0.        , 0.96616484]), array([0.        , 0.        , 0.96446964]), array([0.        , 0.        , 0.99911497]), array([0.        , 0.        , 0.93482214]), array([0.       , 0.       , 0.9797843]), array([0.        , 0.        , 0.99850672]), array([0.        , 0.        , 0.99826935]), array([0.        , 0.        , 0.90241111]), array([0.        , 0.        , 0.99516307]), array([0.        , 0.        , 0.95590285]), array([0.        , 1.        , 0.43349766]), array([0.        , 0.        , 0.89717372]), array([0.        , 0.        , 0.94233797]), array([0.        , 0.        , 0.96360727]), array([0.        , 0.        , 0.99440617]), array([0.        , 0.        , 0.98729407]), array([0.        , 0.        , 0.96346646]), array([0.        , 0.        , 0.97926273]), array([0.        , 0.        , 0.97561031]), array([0.        , 0.        , 0.99625259]), array([0.        , 0.        , 0.99403932]), array([0.        , 0.        , 0.96948048]), array([0.        , 0.        , 0.98326866]), array([0.        , 0.        , 0.99675852]), array([0.        , 0.        , 0.94191167]), array([0.        , 0.        , 0.99062668]), array([0.        , 0.        , 0.96679357]), array([0.        , 0.        , 0.96403573]), array([0.        , 0.        , 0.99049731]), array([0.        , 0.        , 0.99841206]), array([0.       , 0.       , 0.9888119]), array([0.        , 0.        , 0.85118569]), array([0.        , 0.        , 0.91786242]), array([0.        , 0.        , 0.99600281]), array([0.       , 0.       , 0.9047361]), array([0.        , 0.        , 0.99643072]), array([0.        , 0.        , 0.92070648]), array([0.        , 0.        , 0.97297408]), array([0.        , 0.        , 0.76850822]), array([0.        , 0.        , 0.99017358]), array([0.        , 0.        , 0.67825407]), array([0.        , 0.        , 0.98466946]), array([0.        , 0.        , 0.99279241]), array([0.        , 0.        , 0.98339426]), array([0.        , 0.        , 0.99255287]), array([0.        , 0.        , 0.81621255]), array([0.     , 0.     , 0.85729]), array([0.        , 0.        , 0.97575832]), array([0.        , 0.        , 0.92081112]), array([0.        , 0.        , 0.99809367]), array([0.        , 0.        , 0.99857314]), array([0.        , 0.        , 0.94674232]), array([0.       , 0.       , 0.8920877]), array([0.        , 0.        , 0.96553689]), array([0.        , 0.        , 0.97644882]), array([0.      , 0.      , 0.987606]), array([0.        , 0.        , 0.99805301]), array([0.        , 0.        , 0.99514803]), array([0.        , 0.        , 0.91664554]), array([0.        , 0.        , 0.99112416]), array([0.        , 0.        , 0.94868295]), array([0.        , 1.        , 0.55368446]), array([0.        , 0.        , 0.96927087]), array([0.        , 0.        , 0.90408043]), array([0.       , 0.       , 0.9762914]), array([0.       , 0.       , 0.9999772]), array([0.        , 0.        , 0.76223686]), array([0.        , 0.        , 0.99375907]), array([0.        , 0.        , 0.99624556]), array([0.        , 0.        , 0.97175681]), array([0.        , 0.        , 0.99838366]), array([0.        , 0.        , 0.96423091]), array([0.        , 0.        , 0.99970453]), array([0.        , 0.        , 0.99391135]), array([0.        , 0.        , 0.99857859]), array([0.        , 0.        , 0.80123847]), array([0.        , 0.        , 0.99722704]), array([0.        , 0.        , 0.96979165]), array([0.        , 0.        , 0.96039809]), array([0.        , 0.        , 0.99394003]), array([0.        , 0.        , 0.96632099]), array([0.        , 0.        , 0.99293858]), array([0.        , 0.        , 0.88484306]), array([0.        , 0.        , 0.83961857]), array([0.        , 0.        , 0.99997346]), array([0.        , 0.        , 0.99766221]), array([0.        , 0.        , 0.99609024]), array([0.        , 0.        , 0.99149056]), array([0.        , 0.        , 0.86026252]), array([0.        , 0.        , 0.89876288]), array([0.        , 0.        , 0.95266939]), array([0.        , 0.        , 0.98790745]), array([0.        , 0.        , 0.99141712]), array([0.        , 0.        , 0.98343425]), array([0.        , 0.        , 0.98953667]), array([0.        , 0.        , 0.95042619]), array([0.        , 0.        , 0.99564659]), array([0.      , 0.      , 0.975988]), array([0.        , 0.        , 0.91931725]), array([0.        , 0.        , 0.99624693]), array([0.        , 0.        , 0.93934141]), array([0.        , 0.        , 0.98423569]), array([0.        , 0.        , 0.99949202]), array([0.        , 0.        , 0.99634996]), array([0.       , 0.       , 0.9987154]), array([0.        , 0.        , 0.96020002]), array([0.        , 0.        , 0.86133775]), array([0.        , 0.        , 0.99410421]), array([0.        , 0.        , 0.99558319]), array([0.        , 0.        , 0.99046814]), array([0.        , 0.        , 0.99736584]), array([0.        , 0.        , 0.99698278]), array([0.       , 0.       , 0.9807203]), array([0.        , 0.        , 0.99213603]), array([0.        , 0.        , 0.99754025]), array([0.       , 0.       , 0.9870817]), array([0.        , 0.        , 0.94918187]), array([0.        , 0.        , 0.90791246]), array([0.        , 0.        , 0.99520417]), array([0.        , 0.        , 0.98301998]), array([0.        , 0.        , 0.99741858]), array([0.        , 0.        , 0.76231614]), array([0.        , 0.        , 0.98881362]), array([0.        , 0.        , 0.97472608]), array([0.        , 0.        , 0.98999003]), array([0.        , 0.        , 0.94279425]), array([0.        , 0.        , 0.97745683]), array([0.        , 0.        , 0.94346686]), array([0.       , 0.       , 0.9686161]), array([0.        , 0.        , 0.94084295]), array([0.        , 0.        , 0.96028306]), array([0.        , 0.        , 0.99745375]), array([0.        , 0.        , 0.93147817]), array([0.        , 0.        , 0.92129586]), array([0.        , 0.        , 0.99241611]), array([0.        , 0.        , 0.94253062]), array([0.        , 0.        , 0.99885916]), array([0.        , 0.        , 0.99197152]), array([0.        , 0.        , 0.92000165]), array([0.        , 0.        , 0.97812112]), array([0.        , 0.        , 0.96316487]), array([0.        , 0.        , 0.83746861]), array([0.        , 0.        , 0.99257251]), array([0.        , 0.        , 0.69329564]), array([0.        , 0.        , 0.99807744]), array([0.        , 0.        , 0.98281657]), array([0.        , 0.        , 0.99766867]), array([0.        , 0.        , 0.99701096]), array([0.        , 1.        , 0.60469016]), array([0.        , 0.        , 0.97641395]), array([0.       , 0.       , 0.9968483]), array([0.        , 0.        , 0.69144427]), array([0.       , 0.       , 0.9986151]), array([0.        , 1.        , 0.61406576]), array([0.       , 0.       , 0.9922674]), array([0.        , 0.        , 0.85906216]), array([0.        , 0.        , 0.99459577]), array([0.        , 0.        , 0.98086832]), array([0.        , 0.        , 0.98924238]), array([0.        , 0.        , 0.79155998]), array([0.        , 0.        , 0.98814459]), array([0.        , 0.        , 0.99531783]), array([0.        , 0.        , 0.95565576]), array([0.        , 0.        , 0.86839359]), array([0.        , 0.        , 0.96619381]), array([0.        , 0.        , 0.99850415]), array([0.        , 0.        , 0.98352825]), array([0.        , 0.        , 0.98040755]), array([0.        , 0.        , 0.91040726]), array([0.        , 0.        , 0.98343879]), array([0.        , 0.        , 0.91519447]), array([0.        , 0.        , 0.99516592]), array([0.        , 0.        , 0.97693226]), array([0.        , 0.        , 0.99853929]), array([0.        , 0.        , 0.98169705]), array([0.        , 0.        , 0.99664405]), array([0.        , 0.        , 0.98743311]), array([0.        , 0.        , 0.99987932]), array([0.        , 0.        , 0.96773904]), array([0.        , 0.        , 0.91150869]), array([0.       , 0.       , 0.9483738]), array([0.        , 0.        , 0.88312501]), array([0.        , 0.        , 0.98145751]), array([0.        , 0.        , 0.98147494]), array([0.        , 0.        , 0.96840267]), array([0.        , 0.        , 0.98962116]), array([0.        , 0.        , 0.99963239]), array([0.        , 0.        , 0.89126038]), array([0.        , 0.        , 0.88673847]), array([0.        , 0.        , 0.92701706]), array([0.        , 0.        , 0.97463044]), array([0.        , 0.        , 0.99713369]), array([0.        , 0.        , 0.98707737]), array([0.        , 0.        , 0.99630842]), array([0.        , 0.        , 0.99744119]), array([0.       , 0.       , 0.9769818]), array([0.        , 0.        , 0.99426155]), array([0.        , 0.        , 0.98166102]), array([0.        , 0.        , 0.96449809]), array([0.        , 0.        , 0.99748054]), array([0.        , 0.        , 0.97042346]), array([0.        , 0.        , 0.99064559]), array([0.        , 0.        , 0.94506469]), array([0.        , 0.        , 0.99874723]), array([0.        , 0.        , 0.99273227]), array([0.        , 0.        , 0.96994797]), array([0.        , 0.        , 0.93970592]), array([0.       , 0.       , 0.9542468]), array([0.        , 0.        , 0.99034697]), array([0.        , 0.        , 0.99729503]), array([0.       , 0.       , 0.9754612]), array([0.        , 0.        , 0.97259218]), array([0.        , 0.        , 0.99078262]), array([0.        , 0.        , 0.99948306]), array([0.        , 0.        , 0.97023124]), array([0.        , 0.        , 0.86103914]), array([0.        , 0.        , 0.94180443]), array([0.        , 0.        , 0.99946186]), array([0.        , 0.        , 0.98683769]), array([0.        , 0.        , 0.99722639]), array([0.        , 0.        , 0.90235936]), array([0.        , 0.        , 0.80204514]), array([0.        , 0.        , 0.99993984]), array([0.        , 0.        , 0.94157056]), array([0.        , 0.        , 0.98977577]), array([0.        , 0.        , 0.97053063]), array([0.        , 0.        , 0.88401857]), array([0.        , 0.        , 0.99942577]), array([0.        , 0.        , 0.98421005]), array([0.        , 0.        , 0.78631156]), array([0.       , 0.       , 0.9721032]), array([0.        , 0.        , 0.83199039]), array([0.        , 0.        , 0.96299061]), array([0.        , 0.        , 0.99452051]), array([0.        , 0.        , 0.80999563]), array([0.        , 0.        , 0.99683443]), array([0.        , 0.        , 0.99579706]), array([0.        , 1.        , 0.23805962]), array([0.        , 0.        , 0.99572537]), array([0.        , 0.        , 0.86989571]), array([0.        , 0.        , 0.99722169]), array([0.        , 0.        , 0.97592993]), array([0.        , 0.        , 0.97380523]), array([0.        , 0.        , 0.85231298]), array([0.        , 0.        , 0.98694531]), array([0.        , 0.        , 0.95426323]), array([0.        , 0.        , 0.98987594]), array([0.       , 0.       , 0.9734235]), array([0.        , 0.        , 0.98743532]), array([0.        , 0.        , 0.99917799]), array([0.        , 0.        , 0.93383024]), array([0.        , 0.        , 0.98760114]), array([0.        , 0.        , 0.99990751]), array([0.        , 0.        , 0.99577323]), array([0.       , 0.       , 0.9482522]), array([0.        , 0.        , 0.72541805]), array([0.        , 0.        , 0.99443614]), array([0.        , 0.        , 0.91440318]), array([0.        , 0.        , 0.86071934]), array([0.        , 0.        , 0.97992287]), array([0.        , 0.        , 0.98318611]), array([0.        , 0.        , 0.99989629]), array([0.        , 0.        , 0.98779939]), array([0.        , 0.        , 0.98517267]), array([0.        , 0.        , 0.99995976]), array([0.        , 0.        , 0.99942708]), array([0.        , 0.        , 0.92912124]), array([0.        , 0.        , 0.98662647]), array([0.        , 0.        , 0.99643752]), array([0.        , 0.        , 0.92845259]), array([0.        , 0.        , 0.93982787]), array([0.        , 0.        , 0.99842084]), array([0.        , 0.        , 0.95105521]), array([0.        , 0.        , 0.99859615]), array([0.        , 0.        , 0.97382353]), array([0.        , 0.        , 0.98080811]), array([0.        , 0.        , 0.95183881]), array([0.        , 0.        , 0.98617405]), array([0.        , 0.        , 0.96096557]), array([0.        , 0.        , 0.99861347]), array([0.        , 0.        , 0.99352135]), array([0.        , 0.        , 0.85295567]), array([0.        , 0.        , 0.99273586]), array([0.        , 0.        , 0.99968885]), array([0.       , 0.       , 0.9615684]), array([0.        , 0.        , 0.99097861]), array([0.        , 1.        , 0.41774747]), array([0.        , 0.        , 0.99895142]), array([0.        , 0.        , 0.99351167]), array([0.        , 0.        , 0.95090277]), array([0.       , 0.       , 0.9955954]), array([0.        , 0.        , 0.80597891]), array([0.        , 1.        , 0.21007602]), array([0.        , 0.        , 0.99780173]), array([0.        , 0.        , 0.91324581]), array([0.        , 0.        , 0.99826446]), array([0.        , 0.        , 0.98664237]), array([0.        , 0.        , 0.95889631]), array([0.       , 0.       , 0.9949894]), array([0.        , 0.        , 0.98385996]), array([0.        , 0.        , 0.98516509]), array([0.        , 0.        , 0.99149742]), array([0.        , 0.        , 0.75459633]), array([0.        , 0.        , 0.99959085]), array([0.       , 0.       , 0.8952707]), array([0.        , 0.        , 0.90353683]), array([0.        , 0.        , 0.98383867]), array([0.        , 0.        , 0.94163181]), array([0.        , 0.        , 0.98301806]), array([0.        , 0.        , 0.97457272]), array([0.        , 0.        , 0.98240847]), array([0.        , 0.        , 0.93742604]), array([0.        , 0.        , 0.96828639]), array([0.       , 0.       , 0.9975675]), array([0.        , 0.        , 0.96357651]), array([0.        , 0.        , 0.99530574]), array([0.        , 0.        , 0.89238099]), array([0.        , 0.        , 0.99483355]), array([0.        , 0.        , 0.97540808]), array([0.        , 0.        , 0.99420802]), array([0.        , 0.        , 0.97734383]), array([0.        , 0.        , 0.99775794]), array([0.        , 0.        , 0.94549621]), array([0.        , 0.        , 0.93129622]), array([0.        , 0.        , 0.97594228]), array([0.        , 0.        , 0.99546875]), array([0.        , 0.        , 0.99371966]), array([0.        , 0.        , 0.95400282])]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "unDPIOW9A2tl"
      },
      "source": [
        "def tpr_fpr(arr,thresh):\n",
        "  x=[]\n",
        "  for k in range(len(thresh)):\n",
        "    t=thresh[k]\n",
        "    tp=0\n",
        "    fp=0\n",
        "    tn=0\n",
        "    fn=0\n",
        "    for i in range(len(arr)):  \n",
        "      if (arr[i,0]==0 and (arr[i,2]==t or arr[i,2]>t )):\n",
        "        tp+=1\n",
        "      elif (arr[i,0]==1 and (arr[i,2]==t or arr[i,2]>t)):\n",
        "        fp+=1\n",
        "      elif (arr[i,0]==1 and arr[i,2]<t):\n",
        "        tn+=1\n",
        "      else :\n",
        "        fn+=1\n",
        "    x.append(tp/(tp+fn))\n",
        "    x.append(fp/(tn+fp))\n",
        "  return(x)\n",
        "\n",
        "def precision_recall(arr,thresh):\n",
        "  x=[]\n",
        "  for k in range(len(thresh)):\n",
        "    t=thresh[k]\n",
        "    tp=0\n",
        "    fp=0\n",
        "    tn=0\n",
        "    fn=0\n",
        "    for i in range(len(arr)):  \n",
        "      if (arr[i,0]==0 and (arr[i,2]==t or arr[i,2]>t )):\n",
        "        tp+=1\n",
        "      elif (arr[i,0]==1 and (arr[i,2]==t or arr[i,2]>t)):\n",
        "        fp+=1\n",
        "      elif (arr[i,0]==1 and arr[i,2]<t):\n",
        "        tn+=1\n",
        "      else :\n",
        "        fn+=1\n",
        "    x.append(tp/(tp+fp))\n",
        "    x.append(tp/(tp+fn))\n",
        "  return(x)"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DVswi7tSEmIy"
      },
      "source": [
        "final_plot_roc=np.asarray(tpr_fpr(sequence_arr,thresh_matrix))\n",
        "roc_ma=final_plot_roc.reshape((int(len(final_plot_roc)/2),2))"
      ],
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "3IjVeNquFRaw",
        "outputId": "8877998a-0d56-4d44-bb06-227acce2815c"
      },
      "source": [
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.title('ROC Curve for SVM Training Data')\n",
        "plt.xlabel('True Positive Rate')\n",
        "plt.ylabel('False Positive Rate')\n",
        "plt.plot(roc_ma[:,1],roc_ma[:,0],'r')   \n",
        "plt.show()"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CW5mWg5KFbdy"
      },
      "source": [
        "final_plot_pr=np.asarray(precision_recall(sequence_arr,thresh_matrix))\n",
        "pr=final_plot_pr.reshape((int(len(final_plot_pr)/2),2))\n"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Nbht4yD0HMkx",
        "outputId": "d171c035-1c9b-42d1-bab7-89eed634cad2"
      },
      "source": [
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.title('Precision Recall Curve for SVM')\n",
        "plt.xlabel('Precision ')\n",
        "plt.ylabel('Recall')\n",
        "plt.plot(pr[:,0],pr[:,1],'r')  \n",
        "plt.show()"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd5xV1bn/8c9XmgVEpVhAiogFSyyIGmPvJWIsieYnsUVjLJFoTMw1iV4TY8vVaMSusRAxotGLuRg0KrZYADsqSrCBqIiCEEWKz++Ptcc5DDNwgDlnz8z+vl+veZ29197nnGfPwHnOWmuvtRQRmJlZca2QdwBmZpYvJwIzs4JzIjAzKzgnAjOzgnMiMDMrOCcCM7OCcyKwRiHp/0l6oIzzrpH062rEVGmSjpb0RMl+SFo/z5iWlaTfSfpY0gd5x2LV50RQAJLelvSFpNmSPpR0s6T2jfkeEfGXiNirjPNOjIjfNuZ7A0g6V9K87BpnSPqXpO0b+32Wh6S9JT0maZakaZIelXRgE4irB3AG0C8i1mqk1xwo6QVJn2UJ5mFJvSUdnv17VJ3zW0v6SNIBknbJkuo9dc75RlY+ujFitFpOBMXx7YhoD2wF9Ad+VfcESa2rHlXj+mt2jZ2BR4DhOcfzNUmHkuK5FegOrAn8Bvj2MryWJDXm/90ewPSI+GgZYlnk30xWK7qVlFw6Ar2BIcAC4F5gNWDnOk/bBwjgH9n+NGB7SZ1KzjkKeGNpY7QlcyIomIiYAtwPbApfN2ecLOlN4M2s7IDs21zNN+vNa54vaV1Jf8u+0U6XdGVW/nUzSfZBdVn2De8zSS9Lqnm/myX9ruT1jpc0UdInkkZIWqfkWEg6UdKbWSxD6n6TbOAa5wN/AbpJ6pK9VkdJN0qaKmlK1hTSqk4cr2Xf1l+VtFVWfpakf5eUf2dpf+dZzJcCv42IGyJiZkR8FRGPRsTx2TnnShpa8pxe2fW3zvZHSzpf0pPA58CZksbWeZ+fShqRbbeT9AdJ72a1wGskrVRPbHsADwLrZLWpm7PyAyWNz37voyVtXPKctyX9QtJLwH/qSQZbAG9FxEORzIqIuyPi3YiYA9wJ/KDOc34A3J797QDmkpLG4dl7tgK+R/q7WiNzIigYSesC+wHPlxQfBGwL9JO0JXAT8COgE3AtMCL7YGkF/B14B+gFdAPuqOdt9gJ2AjYgfSP8LjC9nlh2Ay7Ijq+dvW7d1zsA2AbYPDtv7zKusS3pg2U68GlWfDMwH1gf2DKL8YfZ+YcB52bPWRU4sCTefwM7Ztfx38BQSWsvKYY6NgTWBe5ayufVNQg4AegAXANsKKlvyfHvA7dn2xeSfv9bkK65G6kGspCI+CewL/B+RLSPiKMlbQAMAwYDXYCRwH3Z77XGEcD+wGolH941ngM2yr4M7KpFmyFvAQ6tSUySOpJqRrfUOe9WahPG3sArwPv1/mZs+USEf1r4D/A2MBuYQfqwvQpYKTsWwG4l515N+uZa+vwJpKr89qQqe+t63uNo4IlsezdSFX47YIU6590M/C7bvhG4uORYe2Ae0Ksktm+VHL8TOKuBazyX9C1yBqkJYjqwS3ZsTeDLmmvOyo4AHsm2RwGnlfm7fAEYWPeaS+Jdv57n7JAdW3Exr3suMLRkv1f2nNbZ/mjgvDrPGQr8JtvuC8wCVgYE/AfoU3Lu9qRv6fW99y7A5JL9XwN3luyvAEwp+X2+DRy7hN/TdtnfaxowJ/u7ty85/ibw/Wz7eODF+uLJztuQ9AXh/5GS9+i8/0+1tB/XCIrjoIhYLSJ6RsRJEfFFybH3SrZ7AmdkTQIzJM0gfZtdJ3t8Jxb9BriQiHgYuJLULvyRpOskrVrPqeuQElPN82aTPsC7lZxTehfL56Rk0ZA7I2I10gf/K8DWJdfUBphack3XAl2z4+uSvvkvQtIPSprJZpCa1DovJob61NQulrYmUdd7dfZvJyU0SLWBeyPic9K3+JWBcSVx/yMrL0fdv8tX2XuX/l3qxrKQiHg6Ir4bEV1INaqdgLNLTin9tj8o26/PbcApwK7APQ2cY8vJicAgffOs8R5wfpY0an5Wjohh2bEe5XQqR8QVEbE10I/URHFmPae9T/qQBkDSKqTmqCnLcS1ExMekJpRzs2ac90g1gs4l17RqRGySPeU9oE/d15HUE7ie9EHUKUsyr5C+cS+NCdl7HLKYc/5D+vCuUd/dO3WnCn4Q6CJpC1JCqGkW+hj4Atik5Ho7RupIL0fdv4tIybL071L2tMURMQb4G1m/VOY2YHelO7u2o+G2/9uAk4CRWZKzCnAisLquB06UtG3W6buKpP0ldQCeBaYCF2blK0raoe4LSNome34b0gfcHOCret5rGHCMpC0ktQN+DzwTEW8v70VExARSk8/PI2Iq8ADwP5JWlbSCpD6Sau5cuQH4maSts2teP0sCq5A+8KZl13UMC3+YlRtLAKcDv5Z0TEkM35J0XXbaC8BOknpkbea/LON155HuRLoEWIOUGGq+wV8PXCapaxZ7N0lL7F/J3AnsL2n37G94BimR/qucJ2fXdXzJe29E6nd5uiT2t4EnSP8GHoyIescvRMRbpGbJs+s7bo3DicAWEhFjSW22V5I6WieS2sKJiAWkTr31gXeByaQ7OepalfRB9CmpiWE66cOq7nv9k9QefTcpwfQhu0ukkVwCnJB9IP0AaAu8msV1F1lTTUQMB84nfaOeRbpbZY2IeBX4H+Ap4ENgM+DJZQkkIu4i/a6OJX3j/hD4HfC/2fEHgb8CLwHjSJ3y5bgd2AMYXqfJ7hekv93Tkj4D/klqay8n1gnAkcCfSLWLb5NuP55bZkwzSB/8L0uaTWqWuge4uM55t5BqHg01C9XE80REuJO4gpS+rJiZWVG5RmBmVnBOBGZmBedEYGZWcE4EZmYF1+wmGevcuXP06tUr7zDMzJqVcePGfZwN8FtEs0sEvXr1YuzYsUs+0czMvibpnYaOuWnIzKzgnAjMzArOicDMrOCcCMzMCs6JwMys4CqWCCTdpLRU4SsNHJekK5SWKXxJ2dKAZmZWXZWsEdxMWpC6IfuSVlXqS5o7/uoKxmJmZg2o2DiCiHhMUq/FnDIQuDWbq/1pSatJWjubO77xPfEEPPDAsj9/wQJ4913YZRfo0AE22AA6N7BQVdu20LVr/cfMzJqYPAeUdWPh5e4mZ2WLJAJJJ5BqDfTo0WPZ3u2pp+B3v1u25wLUTNc9dGj5z+nSBaZNS9t77gmtWi18/P334Y034Mc/ho8+goMPTu/TujVsscWyx1o3hpVXXvJ5ZlZYFV2PIKsR/D0iFlnVSdLfgQsj4ols/yHgF9nCKA3q379/5Day+M03YfZsePLJ9K1/hXpa1mbPhr/9DXr2TDWHe++FqVNhwIBFz3322crHDNCpE0zPls3dcUdYccW0/cYb8M47MGYMrLFGKuvcGVatb3lhM2vOJI2LiP71HcuzRjCFtA5qje4s51q1Fde3b3rccsvFnzd4cO32VVct+XXnzoXXX0/bU6bAxInQvtzlZRfjwQdTDaNTJ/jnP2HCBJg3D+ZnC1m9k40432abhZ+32mowY0baXnvtlMguuijVWEqtsko6bmbNWp41gv1Ji4LvB2wLXBER9XxtXliuNYKWZv58+POfU+1GSs1nn3wCa62VaitPP73k14BU8wGYNSs99u2bak8nnZRqH59+mt5jxx3hP/+B/faD7t0rc01mVq/F1QgqlggkDQN2ATqT1mc9B2gDEBHXSBJpXdx9gM+BY5bULAROBLkYOTI1LUm1ZVOnwqOPQp8+tX0fQ4em/o2pWTdP69YpEcyevehrXn89DBoE7dpVPn4zyycRVIoTQTM0cyZ8+CHMmQPf+MbCx848Ey6uu6a5mTU2JwJrOiLg7rvhsMNqy9q0WfS8efPS4957p5pIv36w2Wa1/TMrrZSaoEprKWbWICcCa5puvTX1Q6y22qLHLrigvNfo3j2N73BCMFssJwJr3mbMgJdeSmMtWrVKTU3DhtUOEHz4Ydh555QMnBDM6uVEYC1TfR/6XbumhLH11rDmmilBfOtbqW9ixRUXHdRnVhBNdRyB2fJ57rlUM2jfHoYPT/0GNSO5x41LjyNHLvycjTZKd0BNmwb77pue89VXcOed9fdVmBWAawTWMs2dmz7wn3sOHn8c3n4bnn8+dTq/+SaMH5/GP9SMfYA00rpm0KBZC+OmIbOGvPwybL557b6U7my69NLaO5Qi0vbKK6eBcWbNkJuGzBqy2Wbw+edpFPTUqTBqVCo//fSGn3PBBWnuqJ4904A6s2bONQKzUhFpDqa33qqtHYwenQbEXXtt/c/ZaafUrPTf/w2HHlo7gZ9ZE+KmIbPGEJFmap0zBx57DH796/rP+/vfYf/9qxub2RI4EZhV2oQJ6Y6kuubM8XxK1iQsLhF48XqzxrDhhvDll6nv4Hvfqy1fc820foVZE+bOYrPG0rYtPPNM2j700DSf0syZaUBbz55pnMLOO6fV6tq1g/XWW/iOJbOcuGnIrFIi4Le/hfPOS2te12fHHVN/g1mFuWnILA8S/OY3aQGgiPTz+utpYNuQIemcxx+Hbt1g/fXhhhtqlxQ1qyLXCMzycuyxaTnRyZMXLv/00/pnZDVbDq4RmDVFN90E772XagqTJtWWn3oq/Pvf+cVlheNEYNYU9O4NJ5+ctocOTU1FgwfnG5MVhhOBWVNxxRVphHLNYLTLL0/9DA8/nG9c1uI5EZg1FSuskGY//fvf4cUXa8t33x023TS/uKzFcyIwa4o23zz1Hdx3X9ofPz7VDi67LN+4rEVyIjBryg44AG6/HbbaKu2ffnqatmLBgvTz1Vf5xmctghOBWVN3xBHpNtMaK60ErVunn1atUk1h6NA0XsFsGXiKCbPmYI01UvPQ8OGpLwHScpt/+lPaHjQo/bz9dhqg1tr/ta18HlBm1pxFwGuvwSabLFy+5ZZwySWpo9kMDygza7mktA7zvHmpdrDddqn8+edhjz3Smgn33psShlkDXCMwa2ki0gC1d95ZuHyttdJynFZIrhGYFYmUpqiYNQvGjq0dg/DBB3DQQfnGZk2SE4FZS9SqFbRvD1tvDS+/XHvX0dtv5xqWNU1OBGZFMHx4erzqqnzjsCbJ95iZFUHNfEU77AB//CP06pU6mL/5TVhnnVxDs/xVtEYgaR9JEyRNlHRWPcd7SHpE0vOSXpK0XyXjMSus0nWTBw9OfQWHHZbGHEgpGZxxBsyenV+MlpuKJQJJrYAhwL5AP+AISf3qnPYr4M6I2BI4HHC91awSunZNdxN98AE89xyMGweXXlp7fOrUtN+hA9x9N3z2WX6xWtVVskYwAJgYEZMiYi5wBzCwzjkBrJptdwTer2A8Zrbmmmmw2VZbwU9/WruE5vTpqYMZ4NBDoU+ffOO0qqpkIugGvFeyPzkrK3UucKSkycBI4NT6XkjSCZLGSho7bdq0SsRqVmxrrJHmKvrXv9L+xx/nG49VVd53DR0B3BwR3YH9gNskLRJTRFwXEf0jon+XLl2qHqRZYWy/fe32+66gF0UlE8EUYN2S/e5ZWanjgDsBIuIpYEWgcwVjMrNydatbgbeWqpKJYAzQV1JvSW1JncEj6pzzLrA7gKSNSYnAbT9meSq9c0hK8xZZi1axRBAR84FTgFHAa6S7g8ZLOk/SgdlpZwDHS3oRGAYcHc1t8iOzlmaVVdL6BjWOPz6/WKwqPOmcmdXvwQdhr73SdjP7nLBFedI5M1t6PXrkHYFViROBmdXvww9rt2+6Ka2RbC2SE4GZ1W/dkpv+jjsuLX+58cZuJmqBnAjMrH69e6cP/eefhwMOSGWvv57WTH7rrXxjs0blRGBmi7fFFnDfffDqq7Vl662XXzzW6JwIzKw8G28Mc+fW7k+fnl8s1qicCMysfG3a1G63b59fHNaonAjMbNmsuCL06wdz5uQdiS0nJwIzWzp33QV9+6bt116DHXf0WsjNnBOBmS2dQw6BN96AW29N+2PHwr335huTLRcnAjNbNoMGwT77pO2f/hRmzsw3HltmTgRmtuy+//3a7dNOyy8OWy5OBGa27AYNgj/9KW1vs02+sdgya513AGbWzI0alR5POQXGjIE990z7ffvCgAH5xWVl8zTUZrZ8Zs5MH/5jxix67KSTYMiQ6sdki/A01GZWOR07wrPPwuTJ6W6iN96An/88HbvqKrj88nzjsyVyIjCzxtGtW2oO6tsXLroITjghlQ8eDPvvn8YcWJPkRGBmlXHppWnqaoCRI+FnP8s3HmuQE4GZVcYqq8C8efDXv6b9kSMXnrTOmgwnAjOrrF12qd1+773cwrCGORGYWWV17QprrJG2118fJkzINx5bhBOBmVXeP/5Ru73RRvCHP8CsWfnFYwtxIjCzyttmm9RfUOPMM+HJJ/OLxxbiRGBm1dG6NSxYAK1apf3dd883HvuaE4GZVc+8eSkZAHzySb6x2NecCMysetq1q91eay3o2TO/WOxrTgRmVl0zZsDFF6ftd9/NNxYDnAjMrNo6dky3kYInpGsinAjMrPquvTY9/uQncM898MUX+cZTcE4EZlZ9Z5+dHhcsgIMPhpVXhr/9Ld+YCqyiiUDSPpImSJoo6awGzvmupFcljZd0eyXjMbMmYscdUy3g2WehfftUdsghsP32tXcVWdVULBFIagUMAfYF+gFHSOpX55y+wC+BHSJiE2BwpeIxsyZmxRXTQLPXXkurmwE8/XRa/tKqqpI1ggHAxIiYFBFzgTuAgXXOOR4YEhGfAkTERxWMx8yaou7d07rHDzyQ9ocN891EVVbJRNANKJ1qcHJWVmoDYANJT0p6WtI+9b2QpBMkjZU0dtq0aRUK18xyteeesO66abtnT9hsM5g+Pd+YCiLvzuLWQF9gF+AI4HpJq9U9KSKui4j+EdG/S5cuVQ7RzKrmgQdqp61+5RV4881cwymK1hV87SnAuiX73bOyUpOBZyJiHvCWpDdIiaGeVbDNrMXbaCO4//40zqBHD9h227wjKoRK1gjGAH0l9ZbUFjgcGFHnnHtJtQEkdSY1FU2qYExm1tTdeitMmQJPPQVffpl3NIVQsUQQEfOBU4BRwGvAnRExXtJ5kg7MThsFTJf0KvAIcGZEuFHQrMg23bR2+7vfTbeYWkUpIvKOYan0798/xo4dm3cYZlZJv/hF7XxEAJMmQe/e+cXTAkgaFxH96zuWd2exmdmiLrwQPvigdn/27PxiKQAnAjNreiRYc83a/c03hwsuyC+eFs6JwMyarvnzYddd03avXrmG0pIt9vZRSbOA+joRBERErFqRqMzMIC1rOXp02t5rr1xDackWmwgiokO1AjEzq1fNDS2ffQadOuUbSwu1pBrBGos7HhFedNTMqmO99VJTUatWeUfS4iypj2AcMDZ7rPvjezjNrPLuuKN2+8or84ujBVtS05Bv3DWzfH3vezBiBNx+O2y5Zd7RtEhlzzUkaXXSPEAr1pRFxGOVCMrM7Gsffwz33Zf6B0aNgg8/hMMOyzuqFqWsRCDph8BppInjXgC2A54CdqtcaGZmwCefQJs2MGMG/P73sPHGTgSNrNxxBKcB2wDvRMSuwJbAjIpFZWZWY4MN0roE996b9l97DbwuSaMqNxHMiYg5AJLaRcTrwIaVC8vMrI7SkcZdu9auaGbLrdxEMDlbMOZe4EFJ/wu8U7mwzMzq2GYbmDMHVsvWrtp7b/jLX/KNqYUoKxFExHciYkZEnAv8GrgROKiSgZmZLaJdu7RWwQ47pP0jj4S11vIax8uprEQgaTtJHQAi4lFgNKmfwMysulZeGR55BP7nf6BDB5g7Fzp2zDuqZq3cpqGrgdJ5YGdnZWZm1demDWy9NcyalfoLbrkl74iatXITgaJkBZuI+IrKrndsZrZ4c+fCOuvAhAlw+eV5R9OslZsIJkn6iaQ22c9peG1hM8vTnnvW1gQm+eNoeZSbCE4EvglMASYD2wInVCooM7OydOuWHtu1yzeOZq6s5p2I+Ag4vMKxmJktnZdfTo+rrAKrrgorrADPP+/1jZdSuXcNbSDpIUmvZPubS/pVZUMzM1uCfv3g8MNht91Sx/HMmbDvvnlH1eyU2zR0PfBLYB5ARLyEawhmlrdNN4Vhw2D4cBg3LpVNmJDWPD7ppLSYjS1RuYlg5Yh4tk7Z/MYOxsxsmfXoAT/6EWyySdq/+uo0vuD11/ONqxko9xbQjyX1IVu/WNKhwNSKRWVmtrQ6d4ZrrknbH31UOzfRxhunW03btMkvtiau3BrBycC1wEaSpgCDSXcSmZk1PV27pnUMarRtC089lV88TVy5cw1Niog9gC7ARsDOwLcqGZiZ2XLp1GnhZiGvddygxSYCSatK+qWkKyXtCXwOHAVMBL5bjQDNzJZZ587pttIDD4QBA/KOpslaUh/BbcCnpNXIjgfOBgR8JyJeqHBsZmbL56KL0m2l55+fdyRN2pISwXoRsRmApBtIHcQ9ahapMTNrst56C664AgYNSreZWoOW1Ecwr2YjIhYAk50EzKxZOPPM1C/w+9/nHUmTt6QawTck1YzIELBSti8gImLVikZnZrYsRo+Gu++G3/62dj4ia9BiawQR0SoiVs1+OkRE65LtJSYBSftImiBpoqSzFnPeIZJCUv9luQgzs68tWACnnQY9e8IZZ+QdTbNQ7jiCpSapFTAE2BfoBxwhqV8953UATgOeqVQsZlYgI0fCSy/BO+/A++/nHU2zULFEAAwAJmZjEOYCdwAD6znvt8BFgPsezGz5bbtt7fY++6T9nXaC8ePzi6mJq2Qi6Aa8V7I/OSv7mqStgHUj4v8W90KSTpA0VtLYadOmNX6kZtZydO0KF18M22wD66+fFq15/HF44gl4770lP7+AKpkIFkvSCsClwBIb8SLiuojoHxH9u3TpUvngzKx5O/NMePZZuP/+tFYBwIknpnUKvvgi39iaoEomginAuiX73bOyGh2ATYHRkt4GtgNGuMPYzBrVqFG1S1ouWAA33wyff55rSE1NJRPBGKCvpN6S2pLWLxhRczAiZkZE54joFRG9gKeBAyNibAVjMrOi2XBD2GuvdBcRpHUKTj8935iamIolgoiYD5wCjAJeA+6MiPGSzpN0YKXe18xsEWutBW+/nUYZQ1q4xr5W7noEyyQiRgIj65T9poFzd6lkLGZmtM4+8q65Ji1cY0COncVmZlV35pm12/fck18cTYwTgZkVR5cuaVpqgEMO8YCzjBOBmRVH585w++1pO2LhVcwKzInAzIplyy1rtz3ADHAiMLMiqrlraNy4fONoIip615CZWZPTpQtsthl88onHE2ScCMysWC6/PM1Oes890L593tE0CW4aMrPimDcPzjknbf/lL/nG0oQ4EZhZcbRqBUcfnbYfegieey7XcJoKJwIzK44VVoDvfz9tf/opbL01jBmTb0xNgBOBmRXLDjvAo4/W7g8YAI89ll88TYATgZkVz047wb/+Vbt/9935xdIEOBGYWTENGJCWsVxjDfiv/8o7mlz59lEzK6bLLoNnnoGhQ2HNNfOOJleuEZhZ8SxYAL/+ddo+5hiYPj3feHLmRGBmxbPCCrXrEcyblyajK/CtpE4EZlY8UhpPMGxYbdnWW6fynXdO008UiPsIzKy4Dj881QaeeAJuuw0mTYJXX4U2bfKOrKqcCMys2PbYI/2stBKcdRYMGQIdOuQdVVW5acjMbNy41Hl8yCFw2GF5R1N1TgRmVmyzZqUmojXXhGuvrV2roECcCMys2E4+GSZOhCOPTOMKRo2Czz/PO6qqUkTkHcNS6d+/f4wdOzbvMMysJZgxA1ZffdHyq6+GE0+sfjwVJGlcRPSv75g7i82suFZbDd54I81EOn067LdfKj/iiHzjqjInAjMrtr594auv4IADoG3bdCtpx455R1VV7iMwMzv/fLj/fth777SmccE4EZiZXXNNerzvPujdGzbYICWGgnAiMDN78UV48kno0SPtv/lmoaaZcCIwM+vcGb75Tbj00rR/3HG1S1oWgBOBmRnAhAlw1FFpQFnr1uluooKoaCKQtI+kCZImSjqrnuOnS3pV0kuSHpLUs5LxmJk16N13YZVVICKNMB43Lu+IqqZiiUBSK2AIsC/QDzhCUr86pz0P9I+IzYG7gIsrFY+Z2WLtuSdccknaPu64Qo0lqGSNYAAwMSImRcRc4A5gYOkJEfFIRNSM5X4a6F7BeMzMGvbKK/DDH6btG29Mi9ccfzwMHgyffZZvbBVWyUTQDXivZH9yVtaQ44Di3K9lZk1L+/YwcCB8+9u1ZTfckKabaOFLWTaJzmJJRwL9gUsaOH6CpLGSxk6bNq26wZlZMfTqBcOHw4gRaaTxMcek8htuSGMLWrBKJoIpwLol+92zsoVI2gM4GzgwIr6s74Ui4rqI6B8R/bsUcNSfmVVRBPzsZ/DnP8M558CgQXlHVHGVTARjgL6SektqCxwOjCg9QdKWwLWkJPBRBWMxMyvP73+fxhOcempKBAVQsUQQEfOBU4BRwGvAnRExXtJ5kg7MTrsEaA8Ml/SCpBENvJyZWWVFwLnnwq9+ldYm+OMfC7NITUVnH42IkcDIOmW/Kdneo5Lvb2ZWlprmoEsvhYMOSusWr9AkulCrwtNQm5ndfXft9BL33pumoT755HQn0b77ws475xtfhTkRmJntvjtcdRXMnw8/+UkqGzIkPb7+eotPBMWp+5iZNWT11eHHP04dxF99lW4ZBfjGN+D66/ONrQqcCMzMakTAhRemEca77QaPP16IhWrcNGRmBjB7dhpEdtddaZ6hP/8Z2rXLO6qqcI3AzOzLL2HHHVMSaNsWPvwQjj029RkUgBOBmZmUppjYfnuYNw8efhjGj0/9BQXgRGBm1rYtXHkldOiQ+gkGDUpLV7Ztm3dkVeFEYGbFFgE33wybbJI6h6+5Bm65JS1SUxBOBGZWXG+8AfvtlzqJN98cXnoJfvSjwkwtUcOJwMyKZ+bMNKXEppumJqA//hFGj4b11887slz49lEzK47Zs9OI4T/8IS02c+yxcP75sOaaeUeWKycCM2v5Pv8crriiNgGsvTZcdlkaL9C1a97R5aLF1aIAAAk7SURBVM5NQ2bW8l15Jfzyl7VLTk6dmtYibuFzCJXLicDMWr6TT06zi5YuOdm6dVqk3tw0ZGYt3NNPw9lnp0FikBLAt78N3brBBhvkG1sT4URgZi3b1VfXJgFI00bcc0/a3nVXOPjgfOJqQtw0ZGYt2znnwM9/nmoANQ45BB55BL7znfziakKcCMys5Ro8GPr0gYsvhilTasvvvjvVBl54Ib/YmhAnAjNrubbbLj22b1//8QKsNVAOJwIza3nefx9uvRX+7/9grbXSQLJWrWqXpJw6Nc0x1L173pE2Ce4sNrOW46GHYI896j/21Vfp+Msvw2GHVTeuJs41AjNrOb74ouFjEenxo4/SmgP2NdcIzKx5mzULXn0VXnwRxoxJs4i+8krtojJ9+qQmod12g112Kfy8QvVxIjCz5uGzz9K00ePHL/zzzju156y+OgwYAAMHpsdttvEHfxmcCMys6Zg+HYYPh0mT0s+//50eP/ts0XPbtoUNN0xjAo48Mk0p3adP4dYSaAxOBGbWdOy6a+rMLcfcuencl1+Go44q7FoCjcGJwMyajmHD0uRwffrATTelGkE5Bg5snPWFFyxIU1DUaN8+rV7WFLRtC6efDuus0+gv7URgZk3HJpvUzgi6226phtCpE/TsCT16pGkiWlfwY+vLL9NqZTVmz4ahQyv3fuX66qu0qlrfvmkpzUbmRGBmTdN22y3+dtBKueyy6r/nknzwQVpMp+YW2EbmcQRmZgVX0UQgaR9JEyRNlHRWPcfbSfprdvwZSb0qGY+ZmS2qYolAUitgCLAv0A84QlK/OqcdB3waEesDlwEXVSoeMzOrXyVrBAOAiRExKSLmAncAA+ucMxC4Jdu+C9hd8k3AZmbVVMlE0A14r2R/clZW7zkRMR+YCXSq+0KSTpA0VtLYadOmVShcM7Mmql07OPRQWG+9irx8s7hrKCKuA64D6N+/f2W6zc3MmqrVV08jriukkjWCKcC6Jfvds7J6z5HUGugITK9gTGZmVkclE8EYoK+k3pLaAocDI+qcMwI4Kts+FHg4okI3ypqZWb0q1jQUEfMlnQKMAloBN0XEeEnnAWMjYgRwI3CbpInAJ6RkYWZmVVTRPoKIGAmMrFP2m5LtOYCXCjIzy5FHFpuZFZwTgZlZwTkRmJkVnBOBmVnBqbndrSlpGvDOEk+sX2fg40YMpznwNReDr7kYlueae0ZEl/oONLtEsDwkjY2I/nnHUU2+5mLwNRdDpa7ZTUNmZgXnRGBmVnBFSwTX5R1ADnzNxeBrLoaKXHOh+gjMzGxRRasRmJlZHU4EZmYF1yITgaR9JE2QNFHSWfUcP1rSNEkvZD8/zCPOxrSka87O+a6kVyWNl3R7tWNsbGX8nS8r+Ru/IWlGHnE2pjKuuYekRyQ9L+klSfvlEWdjKeN6e0p6KLvW0ZK65xFnY5J0k6SPJL3SwHFJuiL7nbwkaavlftOIaFE/pCmv/w2sB7QFXgT61TnnaODKvGOt8jX3BZ4HVs/2u+Ydd6Wvuc75p5KmQs899gr/na8Dfpxt9wPezjvuCl/vcOCobHs34La8426E694J2Ap4pYHj+wH3AwK2A55Z3vdsiTWCAcDEiJgUEXOBO4CBOcdUaeVc8/HAkIj4FCAiPqpyjI1taf/ORwDDqhJZ5ZRzzQGsmm13BN6vYnyNrZzr7Qc8nG0/Us/xZiciHiOtz9KQgcCtkTwNrCZp7eV5z5aYCLoB75XsT87K6jokq1bdJWndeo43J+Vc8wbABpKelPS0pH2qFl1llPt3RlJPoDe1HxjNVTnXfC5wpKTJpLVATq1OaBVRzvW+CBycbX8H6CCpUxViy1PZ//bL1RITQTnuA3pFxObAg8AtOcdTDa1JzUO7kL4dXy9ptVwjqp7DgbsiYkHegVTBEcDNEdGd1IRwm6SW/P/8Z8DOkp4Hdiatg16Ev3Ojaon/QKYApd/wu2dlX4uI6RHxZbZ7A7B1lWKrlCVeM+lbw4iImBcRbwFvkBJDc1XONdc4nObfLATlXfNxwJ0AEfEUsCJporLmqJz/y+9HxMERsSVwdlbW7G8KWIKl+bdflpaYCMYAfSX1ltSW9CEwovSEOu1pBwKvVTG+SljiNQP3kmoDSOpMaiqaVM0gG1k514ykjYDVgaeqHF8llHPN7wK7A0jamJQIplU1ysZTzv/lziU1nl8CN1U5xjyMAH6Q3T20HTAzIqYuzwtWdM3iPETEfEmnAKNIdx3cFBHjJZ0HjI2IEcBPJB0IzCd1yhydW8CNoMxrHgXsJelVUtX5zIiYnl/Uy6fMa4b04XFHZLdbNGdlXvMZpGa/n5I6jo9urtde5vXuAlwgKYDHgJNzC7iRSBpGuq7OWV/POUAbgIi4htT3sx8wEfgcOGa537OZ/hsxM7NG0hKbhszMbCk4EZiZFZwTgZlZwTkRmJkVnBOBmVnBORFY4UhakM1I+oqk4ZJWboTXPE/SHos5fqKkHyzv+5hVgm8ftcKRNDsi2mfbfwHGRcSlJcdbR8T83AI0qzLXCKzoHgfWl7SLpMcljQBeldRK0iWSxmSTE/6o5gmSfiHpZUkvSrowK7tZ0qHZ9oXZug8vSfpDVnaupJ9l21tkE/+9JOkeSatn5aMlXSTp2Wz9hB2r/cuwYmpxI4vNyiWpNbAv8I+saCtg04h4S9IJpKH720hqBzwp6QFgI9I0wNtGxOeS1qjzmp1Is2BuFBHRwMR+twKnRsSj2SjZc4DB2bHWETEgW1DmHKDB5iazxuIagRXRSpJeAMaS5ua5MSt/NpuQD2Av0nwuLwDPAJ1Ik/TtAfw5Ij4HiIi688bPBOYAN0o6mDQFwNckdQRWi4hHs6JbSAuR1Phb9jgO6LU8F2lWLtcIrIi+iIgtSgskAfyntIj0rX1UnfP2XtwLZ/PjDCBN/HYocApp5axy1cyKuwD//7QqcY3ArH6jgB9LagMgaQNJq5DWrzim5k6jepqG2gMdI2Ik8FPgG6XHI2Im8GlJ+/8g4FHMcuRvHGb1u4HUNPOcUnVhGnBQRPxD0hbAWElzSTNB/lfJ8zoA/ytpRVKt4vR6Xvso4JosmUyiEWaPNFsevn3UzKzg3DRkZlZwTgRmZgXnRGBmVnBOBGZmBedEYGZWcE4EZmYF50RgZlZw/x8BhwBjYEoetAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Wexz5NvKHVHJ",
        "outputId": "08a9d6ef-d7b3-478c-c7c0-bbf1c85ebe05"
      },
      "source": [
        "def fpr_fnr(arr,thresh):\n",
        "  x=[]\n",
        "  for k in range(len(thresh)):\n",
        "    t=thresh[k]\n",
        "    tp=0\n",
        "    fp=0\n",
        "    tn=0\n",
        "    fn=0\n",
        "    for i in range(len(arr)):  \n",
        "      if (arr[i,0]==0 and (arr[i,2]==t or arr[i,2]>t )):\n",
        "        tp+=1\n",
        "      elif (arr[i,0]==1 and (arr[i,2]>=t)):\n",
        "        fp+=1\n",
        "      elif (arr[i,0]==1 and arr[i,2]<t):\n",
        "        tn+=1\n",
        "      else :\n",
        "        fn+=1\n",
        "    x.append(fp/(tn+fp))\n",
        "    x.append(fn/(tp+fn))\n",
        "  return(x)\n",
        "EER=np.asarray(fpr_fnr(sequence_arr,thresh_matrix))\n",
        "plt.title(\"Equal Error Rate | Blue : FPR, Orange : FNR  \")\n",
        "plt.xlabel('Samples')\n",
        "plt.ylabel('Rate')\n",
        "plt.plot(EER.reshape((int(len(EER)/2),2))[:,0])\n",
        "plt.plot(EER.reshape((int(len(EER)/2),2))[:,1])\n",
        "plt.show()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1aAJsqiIIw2a"
      },
      "source": [
        ""
      ],
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "22E85c4qN6N7"
      },
      "source": [
        "# Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IfJaQLBfN7xs"
      },
      "source": [
        ""
      ],
      "execution_count": 37,
      "outputs": []
    }
  ]
}