DAI_ASSIGNMENT_1_Q2_1500to2000_MLP.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DAI_ASSIGNMENT_1_Q2_1500to2000_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": 142,
      "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": 143,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OYMmd4xXew4B",
        "outputId": "ff80a243-9048-4920-bacb-6fe387224341"
      },
      "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": 144,
      "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": "dee64c8c-86cf-4255-ca2a-b9e539ed38d9"
      },
      "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": 145,
      "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": "a32ed0e9-c21f-454d-d9be-df1b464c7d9e"
      },
      "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[1500:2000]))\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[1500:2000]))\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": 146,
      "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": "ef289b4d-ce04-4f10-975b-b1a0051234fe"
      },
      "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": 147,
      "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": 622
        },
        "id": "Rq6vpp9A6mNW",
        "outputId": "3728974f-fabe-41b9-f929-b4641d531a99"
      },
      "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": 148,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Testing accuracy: 0.5\n",
            "----------------------------------------\n",
            "Classification Report: \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           2       0.50      1.00      0.67      1000\n",
            "           4       0.00      0.00      0.00      1000\n",
            "\n",
            "    accuracy                           0.50      2000\n",
            "   macro avg       0.25      0.50      0.33      2000\n",
            "weighted avg       0.25      0.50      0.33      2000\n",
            "\n",
            "Confusion Matrix: \n",
            " [[1000    0]\n",
            " [1000    0]]\n",
            "----------------------------------------\n",
            "----------------------------------------\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQYAAAEGCAYAAACHNTs8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAARKklEQVR4nO3de7SVdZ3H8feXpCDlohwuopamTtFoaoJCXkbzCl7A1gq0MUshNW1JTXkZlqaNaU6jptaEeRnzMhI0qxm8gCbm5I0E1LyEOimmAQHniMCxaQL1N3/sDRz9HTZb5dnP8Zz3a6299n5+z7589jry8bnuJ1JKSFJb3coOIKnjsRgkZSwGSRmLQVLGYpCU2azsABuypmWBu0veR3oO3q/sCHoXXl+9KNobd4lBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGCRlLAZJGYtBUsZikJSxGDaBcy++nP2POJYxx5+6bmzlqlYmTJzEqHHjmTBxEitXtQKQUuLiH0xm5NiTOOaErzL/uefXvWb6jHsYNW48o8aNZ/qMexr+PZQ77NAD+N3T9/Ps/Ac568zTy47TMBbDJjBm1CFcffl33zJ23c3TGD50d2ZMvZ7hQ3fn+lumAfDA7Lm8vHAxM6ZezwVnncGFl/4IqBTJ5BtuZcq1VzDl2iuYfMOt68pE5ejWrRtXXXkRRx51PLvudiDjxo1hyJCdy47VEIUVQ0R8IiLOjoirqrezI2JIUZ9XpqG770qf3r3eMnbfA7MZPfJgAEaPPJhf3T+7Mv7gbzj68IOICHbbZQitra/R3LKchx55lBHD9qBP71706d2LEcP24KFHHm34d9F6ew3bgxde+AMvvvgya9asYdq06Rx91GFlx2qIQoohIs4GfgYEMKd6C2BKRJxTxGd2NK+8uoL+TVsB0NRvS155dQUAS5tfYdCApnXPGzigiaXNLSxtbmHQgP7rx/tXxlWewdsM4o8LF6+bXrjoTwwePKjERI1T1BLDeGBYSumSlNIt1dslwF7Vee2KiJMjYl5EzLvupikFRWu8iCAiyo4h1a2oYngTGNzO+NbVee1KKV2TUhqaUho64YTjCorWGP227Etzy3IAmluWs1XfPgAM7N+PJcvWLwksXdbCwP5NDOzfxJJlzevHmyvjKs/iRUvYbtv1/xlvu83WLF68pMREjVNUMXwduDciZkbENdXbXcC9wMSCPrNDOWDf4UyfOQuA6TNnceB+I9aN33bXvaSUeOLpZ9hii83p37QV++y9Jw/PeYyVq1pZuaqVh+c8xj5771nmV+jy5s77LTvttAPbb78d3bt3Z+zY0dx+xy/LjtUQkVIq5o0julFZddimOrQImJtSeqOe169pWVBMsAKcef4lzH38SVasWEW/rfpy2vgvctD+I/jmeRfzp6XNDB40gMsunESf3r1IKXHR5T/mwd/Mo2ePHlw46RvsMuRvAPjFHXdz7U1TATj5S8dyzBGHlvm13pGeg/crO0IhRh7+WS677Dt8oFs3fnrjVL53yVVlR9qkXl+9qN113MKK4b16PxWDOm8xdHYbKgaPY5CUsRgkZSwGSRmLQVLGYpCUsRgkZSwGSRmLQVLGYpCUsRgkZSwGSRmLQVLGYpCUsRgkZTbb0IyIeApo79TnAFJK6VOFpZJUqg0WA3Bkw1JI6lA2WAwppZfWPo6IjwI7p5RmRUTPWq+T9P630W0MEfEV4D+An1SHtgX+q8hQkspVz8bH04F9gFUAKaXfAwOKDCWpXPUUw19TSqvXTkTEZrS/UVJSJ1FPMfw6IiYBPSPiEODnwO3FxpJUpnqK4RygGXgKOAWYAZxbZChJ5dro3oWU0psRcSPwCJVViOdSR/3NeUmbxEaLISKOAK4GXqBycNMOEXFKSmlm0eEklaOe4xEuAw5MKT0PEBE7AncCFoPUSdWzjaF1bSlULQBaC8ojqQOoda7E56oP50XEDGAalW0MnwfmNiCbpJLUWpU4qs3jpcDfVR83Az0LSySpdLXOlTixkUEkdRz17JXoAYwH/hbosXY8pXRSgbkklaiejY83A4OAw4BfUzmJyo2PUidWTzHslFI6D/hzSulG4Ahg72JjSSpTPcWwpnq/IiJ2Afrg2ZVSp1bPAU7XRMSWwHnAbcAWwLcLTSWpVNFRT3tY07KgYwZTu3oO3q/sCHoXXl+9KNobr3WA0z/UesOU0uXvNZSkjqnWqkSvhqWQ1KHUOsDpO40MIqnj8IIzkjIWg6SMxSAp414JSZl69kp8HBhG5eAmqJyOPafIUJLKtdG9EhFxP/DplFJrdfoCKj/tJqmTqmcbw0BgdZvp1dUxSZ1UPedK3ATMiYj/rE6PAW4sLpKkstVzXYmLImImsPZg+BNTSo8XG0tSmerdXflhYFVK6UpgYUTsUGAmSSXbaDFExPnA2cA/Voe6A7cUGUpSuepZYjgGOBr4M0BKaTGeYCV1avUUw+rqtSoTQERsXmwkSWWrpximRcRPgL4R8RVgFnBdsbEklamevRKXRsQhwCoqR0F+O6V0T+HJJJWmnutK/HNK6WzgnnbGJHVC9axKHNLO2MhNHURSx1Hr7MqvAqcBO0bEk21m9QIeLjqYpPLUWpW4FZgJfA84p814a0ppeaGpJJVqg6sSKaWVKaU/AFcCy1NKL6WUXgJejwivRCV1YvVsY5gMvNZm+rXqmKROqp5iiNTmqjQppTep76xMSe9T9RTDgog4IyK6V28TgQVFB5NUnnqK4VTgM8AiYCGVK12fXGQoSeWq58jHZcCxDcgiqYOodRzDWSml70fED6meQNVWSumMQpNJKk2tJYZnqvfzGhFEUsdR61eib6/e+/uOUhdTa1XidtpZhVgrpXR0IYkkla7WqsSl1fvPAYNY/3NuxwFLiwwlqVy1ViV+DRARl6WUhraZdXtEuN1B6sTqOY5h84j42NqJ6i9E+/NuUidWz6HN3wD+OyIWAAF8FDil0FSSSlXPAU53RcTOwCeqQ8+mlP5abCxJZarnuhIfBs4EvpZSegL4SEQcWXgySaWpZxvDDVQuZDuiOr0I+G5hiSSVrp5i2DGl9H1gDUBK6X+pbGuQ1EnVdcGZiOjJ+gvO7Ai4jUHqxOrZK3E+cBewXUT8O7AP8OUiQ0kqV81iiIhuwJZUjn4cTmUVYmJKqaUB2SSVpGYxpJTerJ5+PQ24s0GZJJWsnm0MsyLiWxGxXURstfZWeDJJpalnG8O46v3pbcYS8LF2niupE6jnyMcdGhFEUsdRz0Vte1C5VN2+VJYUHgCuTin9X8HZJJWknlWJm4BW4IfV6S8ANwOfLyqUpHLVUwy7pJQ+2Wb6voiYX1QgSeWrZ6/EYxExfO1E9bqV/lCL1InVs8SwJ/BwRLxcnf4I8FxEPAWklNKnCksnqRT1FMPhhaeQ1KHUs7vypUYEkdRx1LONQVIXYzFIylgMkjIWg6SMxSApYzFIylgMkjIWg6SMxSApYzFIylgMm8C5F1/O/kccy5jjT103tnJVKxMmTmLUuPFMmDiJlataAUgpcfEPJjNy7Ekcc8JXmf/c8+teM33GPYwaN55R48YzfcY9Df8eyh126AH87un7eXb+g5x15ukbf0EnYTFsAmNGHcLVl7/1qn3X3TyN4UN3Z8bU6xk+dHeuv2UaAA/MnsvLCxczY+r1XHDWGVx46Y+ASpFMvuFWplx7BVOuvYLJN9y6rkxUjm7dunHVlRdx5FHHs+tuBzJu3BiGDNm57FgNYTFsAkN335U+vXu9Zey+B2YzeuTBAIweeTC/un92ZfzB33D04QcREey2yxBaW1+juWU5Dz3yKCOG7UGf3r3o07sXI4btwUOPPNrw76L19hq2By+88AdefPFl1qxZw7Rp0zn6qMPKjtUQFkNBXnl1Bf2bKr+y39RvS155dQUAS5tfYdCApnXPGzigiaXNLSxtbmHQgP7rx/tXxlWewdsM4o8LF6+bXrjoTwwePKjERI3T8GKIiBNrzDs5IuZFxLzrbprSyFiFiggivA6w3j/KWGL4zoZmpJSuSSkNTSkNnXDCcY3MtMn127IvzS3LAWhuWc5WffsAMLB/P5YsW78ksHRZCwP7NzGwfxNLljWvH2+ujKs8ixctYbttB6+b3nabrVm8eEmJiRqnkGKIiCc3cHsKGFjEZ3Y0B+w7nOkzZwEwfeYsDtxvxLrx2+66l5QSTzz9DFtssTn9m7Zin7335OE5j7FyVSsrV7Xy8JzH2GfvPcv8Cl3e3Hm/ZaeddmD77beje/fujB07mtvv+GXZsRqinp92ezcGAocBr75tPICHC/rM0px5/iXMffxJVqxYxUFjjue08V9kwhfH8s3zLuYXd9zN4EEDuOzCSQDsP2IYD8yey8ixJ9GzRw8unPQNAPr07sUpXz6OYydMBODUE7+QbdBUY73xxhtM/Pq5zLjzVj7QrRs/vXEq8+f/T9mxGiJSSpv+TSOuB25IKT3YzrxbU0pf2Nh7rGlZsOmDqTA9B+9XdgS9C6+vXtTuxq9ClhhSSuNrzNtoKUgql7srJWUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUsBkkZi0FSxmKQlLEYJGUipVR2hi4nIk5OKV1Tdg7Vpyv+vVxiKMfJZQfQO9Ll/l4Wg6SMxSApYzGUo0utr3YCXe7v5cZHSRmXGCRlLAZJGYuhgSLi8Ih4LiKej4hzys6j2iLi3yJiWUQ8XXaWRrMYGiQiPgD8KzAS+CRwXER8stxU2oifAoeXHaIMFkPj7AU8n1JakFJaDfwMGF1yJtWQUrofWF52jjJYDI2zDfDHNtMLq2NSh2MxSMpYDI2zCNiuzfS21TGpw7EYGmcusHNE7BARHwSOBW4rOZPULouhQVJKrwNfA+4GngGmpZR+V24q1RIRU4DZwMcjYmFEjC87U6N4SLSkjEsMkjIWg6SMxSApYzFIylgMkjIWQxcSEX0j4rQC3//LEfGjjTzngoj41jt839feWzK9UxZD19IXaLcYImKzBmdRB2YxdC2XADtGxG8j4l8i4oCIeCAibgPmR8T2bX97ICK+FREXVB/vGBF3RcSj1dd8otYHRcRREfFIRDweEbMiYmCb2btFxOyI+H1EfKXNa86MiLkR8WREfGfTfnW9E/5foms5B9glpbQ7QEQcAHy6OvZiRGxf47XXAKemlH4fEXsDPwY+W+P5DwLDU0opIiYAZwHfrM77FDAc2Bx4PCLuBHYBdqZyenoAt0XE/tVTn9VgFoPmpJRerPWEiNgC+Azw84hYO/yhjbzvtsDUiNga+CDQ9jOmp5T+AvwlIu6jUgb7AocCj1efswWVorAYSmAx6M9tHr/OW1cve1TvuwEr1i5p1OmHwOUppduqSyYXtJn39uPwE5WlhO+llH7yDj5DBXEbQ9fSCvSqMX8pMCAi+kXEh4AjAVJKq4AXI+LzAFGx20Y+qw/rTyv/0tvmjY6IHhHRDziAypmndwMnVZdOiIhtImJA/V9Nm5JLDF1ISumViHiouoFxJnDn2+aviYh/AuZQ+Uf9bJvZfw9Mjohzge5UfpruiRofdwGVVY9XgV8BO7SZ9yRwH9AEXJhSWgwsjoghwOzq6sprwPHAsnf5dfUeeHalpIyrEpIyFoOkjMUgKWMxSMpYDJIyFoOkjMUgKfP/JtxeMLUTZc4AAAAASUVORK5CYII=\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": "51216525-213c-444c-cec7-a77b88b1e092"
      },
      "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": 149,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mean / STD:\n",
            "50.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1_e75za46LMu"
      },
      "source": [
        ""
      ],
      "execution_count": 149,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "MTsjovHOh3Et",
        "outputId": "f93116bd-f4cb-4705-f2cb-0d90c5123215"
      },
      "source": [
        "fpr, tpr, thresholds = roc_curve(Y_test, Y_pred, pos_label=2)\n",
        "plt.plot(tpr, fpr)\n",
        "print(thresholds)"
      ],
      "execution_count": 150,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[3 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": "f95afabb-21ae-4eb6-d516-d7916698f192"
      },
      "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": 151,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0 0 0 ... 1 1 1]\n",
            "[0 0 0 ... 0 0 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": 152,
      "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": 153,
      "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": "272b1ca7-50b4-492b-d385-ffeee9f53b76"
      },
      "source": [
        "prb = classifier.predict_proba(X_test/255)\n",
        "print(prb.shape)"
      ],
      "execution_count": 154,
      "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": "fd8c222a-a7b7-469c-bab1-4a09830f9f88"
      },
      "source": [
        "prb\n",
        "plt.plot(prb)"
      ],
      "execution_count": 155,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7fa8fe43bf60>,\n",
              " <matplotlib.lines.Line2D at 0x7fa8fe4480b8>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 155
        },
        {
          "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": 156,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3-h_YbkoKVXj"
      },
      "source": [
        ""
      ],
      "execution_count": 156,
      "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": 157,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C3PCvfyM_lJW",
        "outputId": "7efe504c-f4ef-493d-dc9d-1d306512af26"
      },
      "source": [
        "roc"
      ],
      "execution_count": 158,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.        , 0.        , 0.91345805],\n",
              "       [0.        , 0.        , 0.91345805],\n",
              "       [0.        , 0.        , 0.91345805],\n",
              "       ...,\n",
              "       [1.        , 0.        , 0.91345805],\n",
              "       [1.        , 0.        , 0.91345805],\n",
              "       [1.        , 0.        , 0.91345805]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 158
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9kZD5LHKAuAQ",
        "outputId": "6d80a047-b3d4-4d97-9192-b62a79d43a49"
      },
      "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": 159,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "coat_4\n",
            "[array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805]), array([1.        , 0.        , 0.91345805])]\n",
            "\n",
            "\n",
            "pullover_2\n",
            "[array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805]), array([0.        , 0.        , 0.91345805])]\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": 160,
      "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": 161,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "3IjVeNquFRaw",
        "outputId": "ea50f3a9-83e3-4078-c568-c300b930be1f"
      },
      "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": 162,
      "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": 163,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Nbht4yD0HMkx",
        "outputId": "4e2b99d4-a3de-4528-b38b-20b0cf6c46f1"
      },
      "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": 164,
      "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": "9d6af538-6058-4dbc-eac0-5da782b5a7ec"
      },
      "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": 165,
      "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": 165,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "22E85c4qN6N7"
      },
      "source": [
        "# Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IfJaQLBfN7xs"
      },
      "source": [
        ""
      ],
      "execution_count": 165,
      "outputs": []
    }
  ]
}