DAI_ASSIGNMENT_1_Q2_0to500_MLP.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DAI_ASSIGNMENT_1_Q2_0to500_MLP.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": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yNxN-SY-ZMUz"
      },
      "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": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OYMmd4xXew4B",
        "outputId": "680ae9ae-0be2-4367-dfa9-8a86b68bcf13"
      },
      "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": 41,
      "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": "4ca0052e-8625-46f6-fb3e-ab61dc4093b6"
      },
      "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": 42,
      "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": "e8e2b016-5e8d-4f1b-ef72-921c0b7bb26e"
      },
      "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[0:500]))\n",
        "# To take different samples just change teh indices from 0:500 and then 1500:2000\n",
        "print(X_train.shape) # Expecting 6500\n",
        "Y_train = np.concatenate((Ytrain_pullover, Ytrain_coat[0:500]))\n",
        "# To take different samples just change teh indices from 0:500 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": 43,
      "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": [
        "# Multi Layer Perceptron - Sklearn"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1cuP04v5s8Kr",
        "outputId": "62556245-3cfb-43ba-9d85-a72df79e5d11"
      },
      "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": 44,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(6500, 784)\n",
            "(6500,)\n",
            "(2000, 784)\n",
            "(2000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 568
        },
        "id": "Rq6vpp9A6mNW",
        "outputId": "e66abe38-057b-4dc1-ac1b-11a999f2b7d2"
      },
      "source": [
        "\n",
        "from sklearn.neural_network import MLPClassifier\n",
        "\n",
        "\n",
        "classifier = MLPClassifier(alpha=0.0001,hidden_layer_sizes=(128,128,128,64,1), activation='relu',max_iter=400)\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": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Testing accuracy: 0.792\n",
            "----------------------------------------\n",
            "Classification Report: \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           2       0.71      0.98      0.83      1000\n",
            "           4       0.97      0.60      0.74      1000\n",
            "\n",
            "    accuracy                           0.79      2000\n",
            "   macro avg       0.84      0.79      0.78      2000\n",
            "weighted avg       0.84      0.79      0.78      2000\n",
            "\n",
            "Confusion Matrix: \n",
            " [[984  16]\n",
            " [400 600]]\n",
            "----------------------------------------\n",
            "----------------------------------------\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQYAAAEGCAYAAACHNTs8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAASMklEQVR4nO3de3RU5b3G8e8vCSEQUIkUQUFFjLZqxaIFy0WprWIPWmu7OFWwtaIC3qjWC7a0AtYu0RorraKiHu+Xom0tLipYqUYEKyCoVNBiEZWbBAKEixCS/M4fM2DwjcPWsmcPyfNZizXZ757JPFmzfNz7nX0xd0dEpL68pAOISO5RMYhIQMUgIgEVg4gEVAwiEihIOsBn2bZ6sb4u2YO02L9P0hHkC6ipXmYNjWuLQUQCKgYRCagYRCSgYhCRgIpBRAIqBhEJqBhEJKBiEJGAikFEAioGEQmoGEQkoGIQkYCKQUQCKgYRCagYRCSgYhCRgIpBRAIqBhEJqBhEJKBiEJGAikFEAioGEQmoGEQkoGIQkYCKQUQCKgYRCagYRCSgYhCRgIpBRAIqBhEJqBhEJKBiEJGAikFEAioGEQmoGEQkoGIQkYCKQUQCKgYRCagYRCSgYhCRgIpBRAIqBhEJqBhEJKBiEJGAikFEAioGEQmoGEQkoGIQkYCKQUQCKgYRCagYYvDwxKf53jnDOGPQUB7+418AePvf/2HghZfzg3Mv4X8HD2f+gnd2es38he/Q9YT+PPfC9CQiS9o9E8pYvvQNXp83bafxSy4+j3/NL+eN1//B2BtHJpQue1QMu9mixUv406QpPH7vbfzpwfGUz5zFB0uXUzb+Pi4aPIg/PXgHl15wDmXj79vxmtraWn43/n56fr1bgskF4KGHJtL/tEE7jfU9sSffPb0f3Y49ma7HnETZrXcllC57CuL6xWb2ZeAM4ID00DJgkrsvjOs9c8HiJR/y1SMPp0VREQDHHfNVni+fgZmxcdNmADZu2ky7tvvueM1jT03i5L69+NfCfyeSWT4x/eVXOeigjjuNDR36Y27+7R1UV1cDUFGxJoloWRXLFoOZjQCeAAyYlf5nwONmdm0c75krDj3kIOa+8Rbr1lfx8ZYtTH9lNis/qmDET4dSNv4+vnXmj7jl9nu5fNhPAPioYjXTXprJD8/sn2xw+UylpYfQu3d3Zr78DP94/imOO7Zr0pFiF9cWw/nAke6+rf6gmd0KvAWMbehFZjYEGAIwvuwGLvjx2THFi0+Xgw9k8KABDLliJC2Kiji89BDy8vL4418mM+KyIZz8zd5MmfYS1914G/eOu5Gbxt3NFRcNJi9Pe3W5qqAgnzZt9qFn79P5+nHH8Phjd1F6+DeSjhWruIqhDtgfeP9T4x3S6xrk7hOACQDbVi/2mLLF7gen9+MHp/cD4La7HqB9u7bcdtf9/PzyYQD0O6kPo8beBsBbby/i6lGpnly7vorpr8wmPz+fb53QM5nwEli2dAVPP/0sALPnvE5dXR1t25awenVlwsniE1cxXA5MM7NFwIfpsQOBQ4FLY3rPnLFm7Tr2bbMPK1auYlr5DB6d8DsefWoSs+fNp3u3o3n1tdc5qFNq6mXqUw/seN3IG8o4sVd3lUKO+eukqfTt25MXy2dSWnoIhYWFjboUIKZicPcpZnYY0J2dJx9nu3ttHO+ZS674xQ2sq6qioKCAkVdezF6tWzFmxHDGjrubmtpamhcWMuqa4UnHlAY88vAdnHjCN2jbtoQli+cw5vpbuP+BJ7j3njJenzeN6uptDD7/8qRjxs7cc3OLfU/elWiKWuzfJ+kI8gXUVC+zhsY14yUiARWDiARUDCISUDGISEDFICIBFYOIBFQMIhJQMYhIQMUgIgEVg4gEVAwiElAxiEhAxSAiARWDiAQ+83oMZjYfaOjUZwPc3Y+OLZWIJCrThVpOy1oKEckpn1kM7r7jeo1mdhBQ6u7Pm1mLTK8TkT3fLucYzOxC4Cng7vRQR+DpOEOJSLKiTD5eAvQCqgDcfRHQLs5QIpKsKMWw1d2rty+YWQENT0qKSCMRpRjKzewXQAszOxl4Engm3lgikqQoxXAtUAHMB4YCfwN+GWcoEUnWLr9dcPc6M3sQeJXULsQ7nqvXnBeR3WKXxWBm/YG7gP+QOrips5kNdfdn4w4nIsmIcjxCGfBNd38XwMy6AJMBFYNIIxVljmHD9lJIWwxsiCmPiOSATOdKfD/94xwz+xswkdQcwwBgdhayiUhCMu1KnF7v54+AE9M/VwAtYkskIonLdK7EedkMIiK5I8q3EkXA+cCRQNH2cXcfHGMuEUlQlMnHh4H2QD+gnNRJVJp8FGnEohTDoe7+K2CTuz8I9Ad6xBtLRJIUpRi2pR/XmdlRwN7o7EqRRi3KAU4TzKwN8CtgEtAKuC7WVCKSKMvV0x62rV6cm8GkQS3275N0BPkCaqqXWUPjmQ5w+lmmX+jut/63oUQkN2XalWidtRQiklMyHeA0JptBRCR36IYzIhJQMYhIQMUgIgF9KyEigSjfShwOfJ3UwU2QOh17VpyhRCRZu/xWwsxeArq5+4b08mhSl3YTkUYqyhzDfkB1veXq9JiINFJRzpV4CJhlZn9JL38PeDC+SCKStCj3lfiNmT0LbD8Y/jx3nxdvLBFJUtSvK1sCVe4+DlhqZp1jzCQiCdtlMZjZKGAE8PP0UDPgkThDiUiyoswxnAl8DZgL4O7LzSz2E6we7apLPuxJlvU6NOkIshtF2ZWoTt+r0gHMrDjeSCKStCjFMNHM7gb2MbMLgeeBe+ONJSJJivKtxC1mdjJQReooyOvc/e+xJxORxES5r8RN7j4C+HsDYyLSCEXZlTi5gbHv7O4gIpI7Mp1deRFwMdDFzN6st6o1MDPuYCKSnEy7Eo8BzwI3AtfWG9/g7pWxphKRRH3mroS7r3f3JcA4oNLd33f394EaM9OdqEQasShzDHcCG+stb0yPiUgjFaUYzOvdlcbd64h2xKSI7KGiFMNiMxtuZs3S/34KLI47mIgkJ0oxDAN6AsuApaTudD0kzlAikqwoRz6uAs7KQhYRyRGZjmO4xt1vNrM/kD6Bqj53Hx5rMhFJTKYthoXpxznZCCIiuSPTVaKfST/q+o4iTUymXYlnaGAXYjt3/24siUQkcZl2JW5JP34faM8nl3M7G/gozlAikqxMuxLlAGZW5u7H1Vv1jJlp3kGkEYtyHEOxmR2yfSF9hWhd3k2kEYtyaPMVwItmthgw4CBgaKypRCRRUQ5wmmJmpcCX00Nvu/vWeGOJSJKi3FeiJXA1cKm7vwEcaGanxZ5MRBITZY7hflI3sv1GenkZcENsiUQkcVGKoYu73wxsA3D3zaTmGkSkkYp0wxkza8EnN5zpAmiOQaQRi/KtxChgCtDJzB4FegE/iTOUiCQrYzGYWR7QhtTRj8eT2oX4qbuvzkI2EUlIxmJw97r06dcTgclZyiQiCYsyx/C8mV1lZp3MrGT7v9iTiUhioswx/DD9eEm9MQcOaeC5ItIIRDnysXM2gohI7ohyU9siUreq601qS2E6cJe7b4k5m4gkJMquxEPABuAP6eWBwMPAgLhCiUiyohTDUe5+RL3lF8xsQVyBRCR5Ub6VmGtmx29fSN+3UhdqEWnEomwxHAvMNLMP0ssHAu+Y2XzA3f3o2NKJSCKiFMOpsacQkZwS5evK97MRRERyR5Q5BhFpYlQMIhJQMYhIQMUgIgEVg4gEVAwiElAxiEhAxSAiARWDiARUDCISiHKuhHwBlmec9uyv2bxyLdPOLaNVpy9x4vhLaN6mNWvmv8f04XdSt62WvMIC+owbxr5f7czWtRsov+h2Ni7VRbizzYpbsdfVV1PQuTM4VN18EzUffsDe140mv317aleuZP2YUfjGjQC0vmw4hT164Fu2UnXTjdQsWpTwX7B7aYshJl+54FTWL1q+Y/nYkWex4J4p/Ln3lVSv30Tp2X0BKD27L9XrN/Hn3ley4J4pHDvyrIQSN22tL7uM6lmzWHPuj1lzwWBq3n+f4oGDqJ77Gmt+lHosHjgIgMIePcg/oCNrzhnEhrJb2OuKnyWcfvdTMcSgZYcSOn7rGP79+Is7xjr0OoIlk2cB8O6T0zmw37EAHHhKN959cjoASybPokPvI7Oet6mz4mIKj+7Kx39L3yGhpgbftJHmPXuxZeoUALZMnULzXr0BaN6rN1uemwrAtoULsOJW5JU0rgunqxhi0H3MObx2w+NQ5wA0b9OK6vWb8do6ADatqKRl+zYAtGzfhk3LKwHw2jqqqzbTvE2rZII3UfntO1C3bh17jbiWkgn3stdVV0NREXklbairTH02dZWV5JWkPrP8tm2pXbVqx+trV1eQ1/ZLiWSPS9aLwczOy7BuiJnNMbM5L27aM/fZOn77GLasrmLN/CVJR5Go8vMpOKyUzZP+SuWQC/AtWyg+e2D4PM9+tKQkMfk4Bri/oRXuPgGYAPDAAefskR9Du+MOo9Mp3eh4UlfymzejWesWdL/+RxTu3RLLz8Nr6yjuUMLmlWsB2LxyLcX7l7B5RSWWn0fhXi3ZunZjwn9F01JXUUFdRQU1CxcCsKW8nOKBA6mrXEteSUl6a6GEurWpz6x29Wry27VL3f4dyG/7JepWVySUPh6xbDGY2Zuf8W8+sF8c75kr5o6dyJPHDeep46+g/OI7WDFjAdMvu5OVMxdwcP/uABw6oA8fPDcXgA+fm8uhA/oAcHD/7qyYoevsZlvd2kpqV1WQ36kTAIXdulGzZAlbZ86gqF/qAmZF/U5l68wZAKnxU/oB0OwrR+CbNu3Y5Wgs4tpi2A/oB6z91LgBM2N6z5w25zdPcOL4S/naNQOofGsJi9ITk4ueKKfP74fx/ZfL2LpuI+UX355s0CZqw+/HsffIX0JBM2pXLKfqprGQl8feo0bT4n/6U/vRStaPGQ1A9T//SfMex7PvI4/hW7emntvImPvu32I3s/uA+9395QbWPebuDezA7WxP3ZVoqr5z2IdJR5AvYL8Xyq2h8Vi2GNz9/AzrdlkKIpIsfV0pIgEVg4gEVAwiElAxiEhAxSAiARWDiARUDCISUDGISEDFICIBFYOIBFQMIhJQMYhIQMUgIgEVg4gEVAwiElAxiEhAxSAiARWDiARUDCISUDGISEDFICIBFYOIBFQMIhJQMYhIQMUgIgEVg4gEVAwiElAxiEhAxSAiARWDiARUDCISUDGISEDFICIBFYOIBFQMIhJQMYhIQMUgIgEVg4gEVAwiElAxiEhAxSAiARWDiARUDCISUDGISEDFICIBFYOIBFQMIhJQMYhIQMUgIgFz96QzNDlmNsTdJySdQ6Jpip+XthiSMSTpAPK5NLnPS8UgIgEVg4gEVAzJaFL7q41Ak/u8NPkoIgFtMYhIQMUgIgEVQxaZ2alm9o6ZvWtm1yadRzIzs/8zs1Vm9q+ks2SbiiFLzCwfuAP4DnAEcLaZHZFsKtmFB4BTkw6RBBVD9nQH3nX3xe5eDTwBnJFwJsnA3V8CKpPOkQQVQ/YcAHxYb3lpekwk56gYRCSgYsieZUCnessd02MiOUfFkD2zgVIz62xmhcBZwKSEM4k0SMWQJe5eA1wKTAUWAhPd/a1kU0kmZvY48ApwuJktNbPzk86ULTokWkQC2mIQkYCKQUQCKgYRCagYRCSgYhCRgIqhCTGzfczs4hh//0/M7PZdPGe0mV31OX/vxv8umXxeKoamZR+gwWIws4IsZ5EcpmJoWsYCXczsdTP7rZn1NbPpZjYJWGBmB9e/9oCZXWVmo9M/dzGzKWb2Wvo1X870RmZ2upm9ambzzOx5M9uv3uquZvaKmS0yswvrveZqM5ttZm+a2Zjd+6fL56H/SzQt1wJHufsxAGbWF+iWHnvPzA7O8NoJwDB3X2RmPYDxwEkZnv8ycLy7u5ldAFwDXJledzRwPFAMzDOzycBRQCmp09MNmGRmJ6RPfZYsUzHILHd/L9MTzKwV0BN40sy2Dzffxe/tCPzRzDoAhUD99/iru38MfGxmL5Aqg97AKcC89HNakSoKFUMCVAyyqd7PNey8e1mUfswD1m3f0ojoD8Ct7j4pvWUyut66Tx+H76S2Em5097s/x3tITDTH0LRsAFpnWP8R0M7M9jWz5sBpAO5eBbxnZgMALKXrLt5rbz45rfzcT607w8yKzGxfoC+pM0+nAoPTWyeY2QFm1i76nya7k7YYmhB3X2NmM9ITjM8Ckz+1fpuZXQ/MIvUf9dv1Vg8C7jSzXwLNSF2a7o0Mbzea1K7HWuAfQOd6694EXgDaAr929+XAcjP7CvBKendlI3AOsOoL/rnyX9DZlSIS0K6EiARUDCISUDGISEDFICIBFYOIBFQMIhJQMYhI4P8BkJW90eYxtOsAAAAASUVORK5CYII=\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": "3adbdb07-0a89-4947-cf83-8a1fc7309ede"
      },
      "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": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mean / STD:\n",
            "79.2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1_e75za46LMu"
      },
      "source": [
        ""
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "MTsjovHOh3Et",
        "outputId": "89080bbe-3bf5-41ca-875d-9596893b31b4"
      },
      "source": [
        "fpr, tpr, thresholds = roc_curve(Y_test, Y_pred, pos_label=2)\n",
        "plt.plot(tpr, fpr)\n",
        "print(thresholds)"
      ],
      "execution_count": 47,
      "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": "a7ddadcd-fa5f-42a8-94cc-c7e2e39d002f"
      },
      "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": 48,
      "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": 49,
      "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": 50,
      "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": "79d0d258-b19a-40b4-f8e3-43f5037184cd"
      },
      "source": [
        "prb = classifier.predict_proba(X_test/255)\n",
        "print(prb.shape)"
      ],
      "execution_count": 55,
      "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": "69c31edd-b59e-4025-8c79-f6c6668dbf36"
      },
      "source": [
        "prb\n",
        "plt.plot(prb)"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7fa8fea7c9e8>,\n",
              " <matplotlib.lines.Line2D at 0x7fa8fea7cb00>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        },
        {
          "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": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3-h_YbkoKVXj"
      },
      "source": [
        ""
      ],
      "execution_count": 59,
      "outputs": []
    },
    {
      "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": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C3PCvfyM_lJW",
        "outputId": "93bd2a16-2341-435d-fe32-0a3af76747ec"
      },
      "source": [
        "roc"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.        , 0.        , 1.        ],\n",
              "       [0.        , 0.        , 1.        ],\n",
              "       [0.        , 0.        , 1.        ],\n",
              "       ...,\n",
              "       [1.        , 0.        , 0.99987397],\n",
              "       [1.        , 1.        , 0.1361828 ],\n",
              "       [1.        , 0.        , 1.        ]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9kZD5LHKAuAQ",
        "outputId": "caddde53-a7bb-4d1a-864f-6d1131468931"
      },
      "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": 62,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "coat_4\n",
            "[array([1.        , 0.        , 0.99524552]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.8682755]), array([1., 0., 1.]), array([1.        , 0.        , 0.99999998]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99880501]), array([1.        , 0.        , 0.63887561]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99997515]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.30313871]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.96719258]), array([1.        , 0.        , 0.99982203]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.94749774]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.17785589]), array([1.        , 1.        , 0.20727149]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.18640224]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.99410195]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.77460941]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9882865]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.67302393]), array([1.        , 0.        , 0.99931297]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.88781686]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.79577023]), array([1.        , 0.        , 0.94319962]), array([1.        , 1.        , 0.19941922]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.95952394]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99862159]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.40633398]), array([1.        , 0.        , 0.97906998]), array([1.        , 1.        , 0.13831515]), array([1.        , 1.        , 0.40542892]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99932875]), array([1.        , 0.        , 0.99726928]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.37240262]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.62713418]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.6403044]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99998173]), array([1.        , 0.        , 0.99133908]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.87016441]), array([1.        , 0.        , 0.95558071]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.81435408]), array([1.        , 0.        , 0.89533305]), array([1., 0., 1.]), array([1.        , 0.        , 0.99937384]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.46547844]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99979301]), array([1.        , 0.        , 0.99963678]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999715]), array([1.        , 0.        , 0.98829749]), array([1.        , 1.        , 0.41773077]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.36373704]), array([1.        , 0.        , 0.74767706]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.86225423]), array([1., 0., 1.]), array([1.       , 0.       , 0.9999995]), array([1.        , 0.        , 0.96086384]), array([1.        , 1.        , 0.42605839]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.38030655]), array([1.        , 0.        , 0.99999004]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999926]), array([1.        , 0.        , 0.56073754]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.45059232]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.53201135]), array([1.        , 0.        , 0.98014528]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.79533356]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.16741108]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.75140018]), array([1.        , 0.        , 0.99999997]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.71511329]), array([1.       , 0.       , 0.9970924]), array([1.        , 1.        , 0.43376401]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99972458]), array([1.        , 1.        , 0.35699274]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.63108633]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999998]), array([1.        , 0.        , 0.82982593]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.79241883]), array([1.        , 0.        , 0.99053771]), array([1.        , 0.        , 0.99999769]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.99998776]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.40827909]), array([1.        , 0.        , 0.71613822]), array([1.        , 0.        , 0.99817826]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.93100502]), array([1.        , 0.        , 0.99694292]), array([1.        , 1.        , 0.49469998]), array([1.        , 0.        , 0.99989791]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99960651]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99992859]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.81113839]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999984]), array([1.        , 0.        , 0.75475985]), array([1.        , 0.        , 0.52115139]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99381529]), array([1.       , 0.       , 0.8976022]), array([1.      , 0.      , 0.687212]), array([1.        , 0.        , 0.99999999]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999664]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999651]), array([1.        , 0.        , 0.99607791]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999912]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9999796]), array([1.       , 0.       , 0.9999966]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.39490653]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98710557]), array([1.        , 0.        , 0.89638573]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.99900107]), array([1.        , 0.        , 0.93737211]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9949872]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999534]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99705439]), array([1.        , 0.        , 0.84511673]), array([1.        , 0.        , 0.73061952]), array([1.        , 0.        , 0.99678977]), array([1.        , 0.        , 0.91569676]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.18179848]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.17228228]), array([1.        , 0.        , 0.71904772]), array([1.        , 0.        , 0.98423125]), array([1.        , 0.        , 0.99999305]), array([1.        , 0.        , 0.82724841]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99649568]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99989938]), array([1.        , 0.        , 0.82078727]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99971586]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.14240514]), array([1.        , 0.        , 0.99999949]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.56323673]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.28810459]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99698676]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99990229]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99518987]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99952303]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.43029257]), array([1.       , 1.       , 0.2822396]), array([1.        , 0.        , 0.99970122]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99991813]), array([1.        , 0.        , 0.96756892]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.84876484]), array([1., 0., 1.]), array([1.        , 0.        , 0.56393245]), array([1.        , 0.        , 0.99998391]), array([1.        , 1.        , 0.15763274]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9587353]), array([1.        , 0.        , 0.99831918]), array([1.        , 0.        , 0.68365338]), array([1.        , 0.        , 0.99999984]), array([1.        , 1.        , 0.49740542]), array([1., 0., 1.]), array([1.        , 1.        , 0.38912381]), array([1.        , 0.        , 0.90046794]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.60891238]), array([1.        , 0.        , 0.99999995]), array([1.        , 0.        , 0.99999872]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.90821531]), array([1.       , 0.       , 0.9909926]), array([1., 0., 1.]), array([1.        , 1.        , 0.29582268]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.77719625]), array([1.        , 1.        , 0.38751859]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.76737808]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.16618311]), array([1.        , 0.        , 0.99150019]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99988313]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99992234]), array([1.        , 0.        , 0.76917646]), array([1.        , 1.        , 0.43073779]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.52231873]), array([1.        , 0.        , 0.97481621]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.26163753]), array([1.        , 0.        , 0.99971163]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9993255]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99974814]), array([1.        , 0.        , 0.99828964]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.36572065]), array([1.        , 0.        , 0.99999996]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.55851887]), array([1.        , 0.        , 0.98466995]), array([1.        , 0.        , 0.99591684]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.83730295]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.97122505]), array([1.        , 0.        , 0.97601714]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.97906793]), array([1.        , 0.        , 0.99931467]), array([1.        , 1.        , 0.17577371]), array([1.       , 0.       , 0.9999999]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.95773409]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.45505075]), array([1.        , 0.        , 0.99999882]), array([1.        , 0.        , 0.99696089]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.94231314]), array([1.        , 0.        , 0.99999701]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.68627779]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.38519554]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.22051012]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.93415515]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999839]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98537147]), array([1.        , 1.        , 0.43165913]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.61767678]), array([1.        , 0.        , 0.66738723]), array([1.        , 0.        , 0.91601671]), array([1.       , 0.       , 0.9989025]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99929928]), array([1.        , 1.        , 0.29887078]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98650202]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99772968]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99998831]), array([1.        , 0.        , 0.66930374]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.35826353]), array([1.        , 0.        , 0.97235145]), array([1.        , 0.        , 0.99996891]), array([1.        , 0.        , 0.98996101]), array([1.        , 1.        , 0.35659465]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999999]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99953031]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.6822295]), array([1.        , 0.        , 0.99975064]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 1.        , 0.28118727]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99970862]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98682153]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99948442]), array([1.        , 0.        , 0.97956528]), array([1.        , 0.        , 0.92829673]), array([1.        , 0.        , 0.99792306]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98552178]), array([1.       , 1.       , 0.1587772]), array([1.        , 0.        , 0.98060973]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98728717]), array([1.        , 1.        , 0.21939957]), array([1.        , 0.        , 0.99381234]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99987994]), array([1.        , 0.        , 0.99999997]), array([1.        , 1.        , 0.42889331]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99688294]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.      , 0.      , 0.999999]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99647161]), array([1.        , 0.        , 0.99947217]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99644797]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99995582]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98389119]), array([1.        , 0.        , 0.99948507]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.65922787]), array([1.        , 0.        , 0.99803027]), array([1.        , 0.        , 0.87287526]), array([1.        , 1.        , 0.15690425]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99975486]), array([1.        , 0.        , 0.98232489]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.80853195]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.4540447]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.8902454]), array([1.        , 0.        , 0.99999212]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.34909111]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.17049815]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99893543]), array([1.        , 0.        , 0.99996978]), array([1., 0., 1.]), array([1.        , 0.        , 0.94811088]), array([1.        , 0.        , 0.95433462]), array([1.        , 0.        , 0.74332261]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99665572]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999964]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9269446]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999662]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99467033]), array([1.       , 0.       , 0.5944496]), array([1.       , 0.       , 0.9886083]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99959221]), array([1.        , 0.        , 0.52677842]), array([1.        , 1.        , 0.16781525]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999497]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99997351]), array([1.        , 0.        , 0.99999879]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98362046]), array([1.        , 1.        , 0.14911595]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999924]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9953587]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99986826]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.99990365]), array([1.        , 0.        , 0.96280953]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.31337468]), array([1.        , 0.        , 0.69452348]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.91322516]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.97660942]), array([1.        , 0.        , 0.61941907]), array([1.        , 0.        , 0.99999995]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.32885027]), array([1.        , 1.        , 0.47339972]), array([1.        , 0.        , 0.95491657]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.86901309]), array([1.        , 0.        , 0.99999532]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.36768131]), array([1.        , 0.        , 0.89329112]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.97903588]), array([1.        , 1.        , 0.17823131]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98661945]), array([1.        , 0.        , 0.93425018]), array([1.        , 0.        , 0.99999841]), array([1., 0., 1.]), array([1.        , 0.        , 0.81261717]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999851]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1526413]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999992]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.80305338]), array([1.        , 0.        , 0.99999992]), array([1.        , 1.        , 0.19372912]), array([1.        , 1.        , 0.30886684]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.2634676]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999973]), array([1.        , 0.        , 0.52044527]), array([1.        , 0.        , 0.99996975]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9971538]), array([1.        , 0.        , 0.83392312]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99829185]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99325729]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.86990989]), array([1.        , 0.        , 0.99992981]), array([1.        , 0.        , 0.99999992]), array([1.        , 0.        , 0.99925747]), array([1.        , 0.        , 0.99999992]), array([1.        , 0.        , 0.99866703]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.86884205]), array([1.      , 0.      , 0.937942]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.51887005]), array([1.        , 0.        , 0.90498882]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999993]), array([1.        , 0.        , 0.74961548]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99991966]), array([1.        , 0.        , 0.99995657]), array([1.        , 1.        , 0.48965242]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999694]), array([1.        , 0.        , 0.99427779]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99979694]), array([1., 0., 1.]), array([1.        , 1.        , 0.16462613]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.93218158]), array([1.       , 0.       , 0.9997227]), array([1.        , 1.        , 0.17080472]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.40156742]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999918]), array([1.        , 0.        , 0.63792561]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.82038041]), array([1.        , 0.        , 0.99999022]), array([1.        , 0.        , 0.79454621]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.40396895]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99676838]), array([1.        , 1.        , 0.29438638]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99995254]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.56208246]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99979783]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.86313856]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.23402579]), array([1.        , 1.        , 0.21555638]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.97676434]), array([1.        , 0.        , 0.96527026]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.96010687]), array([1.        , 0.        , 0.97802775]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999142]), array([1.        , 0.        , 0.83138358]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99952703]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.72547522]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999971]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99962581]), array([1.        , 0.        , 0.79781576]), array([1.        , 0.        , 0.99973146]), array([1.        , 1.        , 0.32293077]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999007]), array([1.        , 0.        , 0.71705209]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.55902487]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99994741]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99695448]), array([1.       , 0.       , 0.8800056]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.72085207]), array([1.        , 0.        , 0.99820502]), array([1.        , 1.        , 0.47038677]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.72385072]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99970193]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1., 0., 1.]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.92346465]), array([1.        , 0.        , 0.74904329]), array([1.       , 0.       , 0.9980841]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.84867618]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.94838125]), array([1.        , 0.        , 0.86098519]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99992914]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99978431]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.85245662]), array([1.        , 0.        , 0.99102439]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999953]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.98204793]), array([1.        , 0.        , 0.99222582]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99844405]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9972333]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.96138158]), array([1.        , 1.        , 0.31663624]), array([1.        , 0.        , 0.99989623]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.96471514]), array([1.        , 0.        , 0.69945698]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99996073]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.19209245]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9999995]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9972692]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.88219621]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.62905087]), array([1.        , 0.        , 0.99992819]), array([1.        , 0.        , 0.96725505]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99212603]), array([1.        , 0.        , 0.93021267]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99902944]), array([1.        , 0.        , 0.50018311]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9999987]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.58721558]), array([1.       , 1.       , 0.1361828]), array([1.        , 1.        , 0.35154569]), array([1.        , 0.        , 0.99999959]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.]), array([1.        , 0.        , 0.99970552]), array([1.        , 0.        , 0.97570108]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9999645]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.88132357]), array([1.        , 0.        , 0.87883088]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.75334728]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.95425757]), array([1.        , 0.        , 0.99644201]), array([1.        , 0.        , 0.77694143]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.7424958]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99998389]), array([1.        , 0.        , 0.98429018]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.9996031]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99992739]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999612]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.97388926]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99312334]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99153993]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99999992]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.75625495]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.      , 0.      , 0.999755]), array([1.        , 0.        , 0.58402929]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99995434]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.95681784]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.95063668]), array([1.        , 0.        , 0.79983359]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.89834904]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.66065348]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 0.       , 0.5318299]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.92666812]), array([1.        , 0.        , 0.99822657]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.85682067]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.       , 1.       , 0.1361828]), array([1.        , 0.        , 0.99987397]), array([1.       , 1.       , 0.1361828]), array([1., 0., 1.])]\n",
            "\n",
            "\n",
            "pullover_2\n",
            "[array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999641]), array([0.        , 0.        , 0.99999989]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.90073561]), array([0., 0., 1.]), array([0.        , 0.        , 0.95353165]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999997]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99913934]), array([0.       , 0.       , 0.9936314]), array([0.        , 0.        , 0.99978581]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999925]), array([0.        , 0.        , 0.93986382]), array([0.       , 0.       , 0.9999998]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999988]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999348]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999911]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99748885]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999809]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999965]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999989]), array([0.        , 0.        , 0.99999996]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99988658]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999997]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99993339]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999993]), array([0., 0., 1.]), array([0.        , 0.        , 0.99994784]), array([0.        , 0.        , 0.99979846]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99965282]), array([0.        , 0.        , 0.99996561]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99852382]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99996046]), array([0.        , 0.        , 0.97201309]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99994686]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.97260961]), array([0.        , 0.        , 0.99882504]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.67041431]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999996]), array([0.        , 0.        , 0.99993202]), array([0.        , 0.        , 0.99999961]), array([0.        , 0.        , 0.75530899]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.97251442]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999868]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99929779]), array([0.       , 0.       , 0.9773749]), array([0.        , 0.        , 0.99825936]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999238]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.93013089]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.87326663]), array([0., 0., 1.]), array([0.        , 0.        , 0.95588759]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 0.       , 0.6278899]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999998]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99998892]), array([0.        , 0.        , 0.99999998]), array([0., 0., 1.]), array([0.        , 0.        , 0.99958296]), array([0.        , 0.        , 0.93072072]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.98868088]), array([0.        , 0.        , 0.99999879]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999962]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999904]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999974]), array([0.       , 0.       , 0.9999975]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999835]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999966]), array([0.        , 0.        , 0.99999816]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999998]), array([0.        , 0.        , 0.99999996]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998607]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99648302]), array([0.        , 0.        , 0.86067951]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999998]), array([0.       , 1.       , 0.1361828]), array([0., 0., 1.]), array([0.        , 0.        , 0.99671763]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99990559]), array([0.        , 0.        , 0.99518765]), array([0., 0., 1.]), array([0.        , 0.        , 0.98935527]), array([0.        , 0.        , 0.99999987]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 0.       , 0.9997835]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99804567]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99306543]), array([0.        , 0.        , 0.98040194]), array([0.        , 0.        , 0.58724794]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999981]), array([0., 0., 1.]), array([0.        , 0.        , 0.99551593]), array([0.        , 0.        , 0.99979422]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999301]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.58259728]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999737]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999285]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999989]), array([0.        , 0.        , 0.99999137]), array([0.        , 0.        , 0.99994994]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999929]), array([0.        , 0.        , 0.99997133]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999999]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999999]), array([0.        , 0.        , 0.71146211]), array([0.        , 0.        , 0.99999886]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.73965757]), array([0.        , 0.        , 0.96987156]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99991757]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999971]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.96697814]), array([0.        , 0.        , 0.99982369]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999833]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999995]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999981]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99557945]), array([0., 0., 1.]), array([0.        , 0.        , 0.99997643]), array([0.        , 0.        , 0.99909976]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999996]), array([0., 0., 1.]), array([0.        , 0.        , 0.99995683]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998881]), array([0.        , 0.        , 0.93014155]), array([0., 0., 1.]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.99998585]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.59985712]), array([0.        , 0.        , 0.99948943]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99996606]), array([0.        , 0.        , 0.99033878]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.91207343]), array([0.        , 0.        , 0.99999982]), array([0.       , 0.       , 0.9999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99985876]), array([0.        , 1.        , 0.29773272]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.83358811]), array([0.       , 0.       , 0.9578729]), array([0.        , 0.        , 0.99998563]), array([0., 0., 1.]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.99999998]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99086181]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999966]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999861]), array([0., 0., 1.]), array([0.        , 0.        , 0.99984165]), array([0.        , 0.        , 0.99992226]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.99999999]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.99999994]), array([0.        , 0.        , 0.99999997]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99369648]), array([0.        , 0.        , 0.99995764]), array([0.       , 0.       , 0.5563467]), array([0.       , 0.       , 0.9997413]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.64065589]), array([0.        , 1.        , 0.26807589]), array([0.        , 0.        , 0.99999928]), array([0.        , 0.        , 0.99999917]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999218]), array([0., 0., 1.]), array([0.        , 0.        , 0.99980024]), array([0.        , 0.        , 0.99998212]), array([0.        , 0.        , 0.96145001]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99976764]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998915]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99996864]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99988158]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999995]), array([0.        , 0.        , 0.99999965]), array([0.        , 0.        , 0.99998832]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99894341]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999627]), array([0.        , 0.        , 0.99999986]), array([0., 0., 1.]), array([0.        , 0.        , 0.92456185]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999325]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99983684]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999998]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999689]), array([0.        , 0.        , 0.99999774]), array([0.        , 0.        , 0.99983113]), array([0.        , 0.        , 0.99999987]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999675]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.99966099]), array([0.       , 0.       , 0.9999618]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999908]), array([0., 0., 1.]), array([0.        , 0.        , 0.76645215]), array([0.        , 0.        , 0.99386404]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99390013]), array([0., 0., 1.]), array([0.        , 0.        , 0.99976851]), array([0.        , 0.        , 0.99936308]), array([0.        , 0.        , 0.99999996]), array([0.        , 0.        , 0.99997847]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999573]), array([0., 0., 1.]), array([0.        , 0.        , 0.61874168]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.93357118]), array([0., 0., 1.]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.96378607]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99995397]), array([0.        , 0.        , 0.99349058]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.97096043]), array([0.        , 0.        , 0.99997018]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999937]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99995586]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.94037128]), array([0.        , 0.        , 0.99999487]), array([0.        , 0.        , 0.99993729]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999987]), array([0., 0., 1.]), array([0.      , 0.      , 0.999976]), array([0.        , 0.        , 0.99999979]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99990815]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999821]), array([0.        , 0.        , 0.99999992]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999705]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999876]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99828862]), array([0.       , 0.       , 0.9995714]), array([0., 0., 1.]), array([0.        , 0.        , 0.99880008]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0.        , 1.        , 0.48430651]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999829]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99757246]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999879]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998416]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999975]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999991]), array([0.        , 0.        , 0.99994038]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.62955725]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.95993709]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.60806359]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999994]), array([0., 0., 1.]), array([0.        , 0.        , 0.95896047]), array([0.        , 0.        , 0.99997465]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999826]), array([0.       , 0.       , 0.9999071]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.98823112]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998047]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999821]), array([0.        , 0.        , 0.99880614]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.96802428]), array([0., 0., 1.]), array([0.        , 0.        , 0.99844739]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999958]), array([0., 0., 1.]), array([0.       , 0.       , 0.9999999]), array([0.       , 1.       , 0.1361828]), array([0., 0., 1.]), array([0.        , 0.        , 0.99806086]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999561]), array([0.        , 0.        , 0.99999262]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999918]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99989229]), array([0., 0., 1.]), array([0.        , 0.        , 0.99996809]), array([0.        , 0.        , 0.99999982]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999998]), array([0., 0., 1.]), array([0.        , 0.        , 0.99954614]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999991]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999889]), array([0.        , 0.        , 0.99921491]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999995]), array([0.        , 0.        , 0.99899826]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99994883]), array([0., 0., 1.]), array([0.        , 0.        , 0.99131154]), array([0.        , 0.        , 0.99939307]), array([0.        , 0.        , 0.99994954]), array([0.        , 0.        , 0.99999314]), array([0.        , 0.        , 0.90379672]), array([0., 0., 1.]), array([0.      , 0.      , 0.999999]), array([0.        , 0.        , 0.99984435]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999994]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998208]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99944279]), array([0.       , 0.       , 0.9999764]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999914]), array([0.        , 0.        , 0.99999736]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999994]), array([0.        , 0.        , 0.72267568]), array([0.        , 0.        , 0.99964852]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99241324]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999991]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.93868541]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998008]), array([0.      , 0.      , 0.990198]), array([0.        , 0.        , 0.99999486]), array([0.       , 0.       , 0.9940424]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999742]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999857]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99955319]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99995106]), array([0.        , 0.        , 0.99998244]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.92332084]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99997204]), array([0.        , 0.        , 0.99999907]), array([0.        , 0.        , 0.94911906]), array([0.        , 0.        , 0.99962473]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99999948]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999938]), array([0.        , 0.        , 0.99988134]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998916]), array([0.        , 0.        , 0.99996526]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999372]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99941064]), array([0., 0., 1.]), array([0.        , 0.        , 0.99828839]), array([0.       , 0.       , 0.9999995]), array([0.        , 0.        , 0.99999117]), array([0., 0., 1.]), array([0.        , 0.        , 0.99993156]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99987376]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998731]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999873]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999984]), array([0.        , 0.        , 0.99997139]), array([0., 0., 1.]), array([0.        , 0.        , 0.84859357]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 0.       , 0.9858821]), array([0.        , 0.        , 0.99998341]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998646]), array([0., 0., 1.]), array([0.        , 0.        , 0.96585742]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99962821]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999997]), array([0.        , 0.        , 0.99999058]), array([0.       , 1.       , 0.1361828]), array([0.        , 0.        , 0.79392546]), array([0.        , 0.        , 0.99999241]), array([0.        , 0.        , 0.99979893]), array([0.        , 0.        , 0.99914782]), array([0.        , 1.        , 0.16747889]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999998]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999994]), array([0.        , 0.        , 0.96241994]), array([0.        , 0.        , 0.99979137]), array([0.        , 0.        , 0.99959452]), array([0.        , 0.        , 0.99999929]), array([0.        , 0.        , 0.99960339]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.57327885]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999992]), array([0.        , 0.        , 0.99997867]), array([0., 0., 1.]), array([0.        , 0.        , 0.99861136]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998184]), array([0.        , 0.        , 0.73419241]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999706]), array([0.        , 0.        , 0.99502618]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999976]), array([0.       , 1.       , 0.1361828]), array([0.       , 0.       , 0.9977346]), array([0., 0., 1.]), array([0.        , 0.        , 0.99987343]), array([0., 0., 1.]), array([0.        , 0.        , 0.86645137]), array([0.        , 0.        , 0.98889176]), array([0., 0., 1.]), array([0.        , 0.        , 0.99995318]), array([0., 0., 1.]), array([0.        , 0.        , 0.64058624]), array([0.        , 0.        , 0.99980916]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99970439]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99001762]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999997]), array([0., 0., 1.]), array([0.        , 0.        , 0.99997797]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99992903]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.87040487]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999855]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.       , 1.       , 0.1361828]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999997]), array([0.       , 1.       , 0.1361828]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999997]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99806279]), array([0.        , 0.        , 0.99988568]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999981]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.75394301]), array([0., 0., 1.]), array([0.        , 0.        , 0.84622676]), array([0., 0., 1.]), array([0.       , 0.       , 0.9937262]), array([0.        , 0.        , 0.99999994]), array([0.      , 0.      , 0.999917]), array([0.        , 0.        , 0.99925176]), array([0.        , 0.        , 0.99999987]), array([0.        , 0.        , 0.99990629]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.97686366]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999983]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998415]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.90733051]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.77462359]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99999928]), array([0., 0., 1.]), array([0.        , 0.        , 0.99931514]), array([0.        , 0.        , 0.99999885]), array([0.        , 0.        , 0.99999964]), array([0.        , 0.        , 0.98903175]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999999]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0.        , 0.        , 0.99998853]), array([0., 0., 1.]), array([0.        , 0.        , 0.99992353]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999957]), array([0., 0., 1.]), array([0.        , 0.        , 0.99999903]), array([0.        , 0.        , 0.99999999]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.]), array([0., 0., 1.])]\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": 63,
      "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": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "3IjVeNquFRaw",
        "outputId": "5a404dac-a3ba-47e9-f252-5a237f967fe4"
      },
      "source": [
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.title('ROC Curve for MLP 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": 66,
      "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": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Nbht4yD0HMkx",
        "outputId": "bcaf744d-17d1-48f0-a91d-b5f201f1865b"
      },
      "source": [
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.title('Precision Recall Curve for SVM')\n",
        "plt.ylabel('Precision ')\n",
        "plt.xlabel('Recall')\n",
        "plt.plot(pr[:,0],pr[:,1],'r')  \n",
        "plt.show()"
      ],
      "execution_count": 68,
      "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": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Wexz5NvKHVHJ",
        "outputId": "77574c89-5d90-472e-d24b-1a61de2ad075"
      },
      "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": 69,
      "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": 25,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "22E85c4qN6N7"
      },
      "source": [
        "# Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IfJaQLBfN7xs"
      },
      "source": [
        ""
      ],
      "execution_count": 25,
      "outputs": []
    }
  ]
}