Robustness-HDFS.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1><center>Robustness - HDFS</center></h1>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 327,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training ratio 0.9 - Reduction percentage 0.75\n",
      "Reduction percentage 0.0\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.904, recall: 0.999, F1-measure: 0.950\n",
      "\n",
      "MCC: 0.949, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.904, recall: 0.999, F1-measure: 0.950\n",
      "\n",
      "MCC: 0.949, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.989, recall: 0.998, F1-measure: 0.993\n",
      "\n",
      "MCC: 0.993, ROC_AUC: 0.999\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 41-by-41\n",
      "SPE threshold: 2.6409544615360288\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.818, recall: 0.681, F1-measure: 0.743\n",
      "\n",
      "MCC: 0.745, ROC_AUC: 0.840\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.958, recall: 0.692, F1-measure: 0.803\n",
      "\n",
      "MCC: 0.810, ROC_AUC: 0.845\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.253, recall: 1.000, F1-measure: 0.404\n",
      "\n",
      "MCC: 0.498, ROC_AUC: 0.989\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[-1 -1  1 ...  1  1  1]\n",
      "Precision: 0.578, recall: 1.000, F1-measure: 0.733\n",
      "\n",
      "MCC: 0.752, ROC_AUC: 0.989\n",
      "\n",
      "Reduction percentage 0.25\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.876, recall: 1.000, F1-measure: 0.934\n",
      "\n",
      "MCC: 0.935, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.876, recall: 1.000, F1-measure: 0.934\n",
      "\n",
      "MCC: 0.935, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.985, recall: 0.995, F1-measure: 0.990\n",
      "\n",
      "MCC: 0.990, ROC_AUC: 0.997\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 41-by-41\n",
      "SPE threshold: 2.6409544615360288\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.818, recall: 0.681, F1-measure: 0.743\n",
      "\n",
      "MCC: 0.745, ROC_AUC: 0.840\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.946, recall: 0.705, F1-measure: 0.808\n",
      "\n",
      "MCC: 0.813, ROC_AUC: 0.852\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.253, recall: 1.000, F1-measure: 0.404\n",
      "\n",
      "MCC: 0.498, ROC_AUC: 0.989\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.507, recall: 1.000, F1-measure: 0.673\n",
      "\n",
      "MCC: 0.704, ROC_AUC: 0.989\n",
      "\n",
      "Reduction percentage 0.5\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.825, recall: 1.000, F1-measure: 0.904\n",
      "\n",
      "MCC: 0.907, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.999, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.825, recall: 1.000, F1-measure: 0.904\n",
      "\n",
      "MCC: 0.907, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.978, recall: 0.995, F1-measure: 0.986\n",
      "\n",
      "MCC: 0.986, ROC_AUC: 0.997\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 41-by-41\n",
      "SPE threshold: 2.6409544615360288\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.818, recall: 0.681, F1-measure: 0.743\n",
      "\n",
      "MCC: 0.745, ROC_AUC: 0.840\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.921, recall: 0.705, F1-measure: 0.799\n",
      "\n",
      "MCC: 0.804, ROC_AUC: 0.852\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.253, recall: 1.000, F1-measure: 0.404\n",
      "\n",
      "MCC: 0.498, ROC_AUC: 0.989\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.407, recall: 1.000, F1-measure: 0.578\n",
      "\n",
      "MCC: 0.631, ROC_AUC: 0.989\n",
      "\n",
      "Reduction percentage 0.75\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.702, recall: 0.998, F1-measure: 0.824\n",
      "\n",
      "MCC: 0.836, ROC_AUC: 0.997\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.702, recall: 0.998, F1-measure: 0.824\n",
      "\n",
      "MCC: 0.836, ROC_AUC: 0.997\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.957, recall: 0.998, F1-measure: 0.977\n",
      "\n",
      "MCC: 0.977, ROC_AUC: 0.999\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 41-by-41\n",
      "SPE threshold: 2.6409544615360288\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.818, recall: 0.681, F1-measure: 0.743\n",
      "\n",
      "MCC: 0.745, ROC_AUC: 0.840\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.849, recall: 0.682, F1-measure: 0.756\n",
      "\n",
      "MCC: 0.759, ROC_AUC: 0.840\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.253, recall: 1.000, F1-measure: 0.404\n",
      "\n",
      "MCC: 0.498, ROC_AUC: 0.989\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.255, recall: 1.000, F1-measure: 0.407\n",
      "\n",
      "MCC: 0.500, ROC_AUC: 0.989\n",
      "\n",
      "Reduction percentage 0.99\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.082, recall: 1.000, F1-measure: 0.152\n",
      "\n",
      "MCC: 0.287, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.082, recall: 1.000, F1-measure: 0.152\n",
      "\n",
      "MCC: 0.287, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.457, recall: 1.000, F1-measure: 0.627\n",
      "\n",
      "MCC: 0.676, ROC_AUC: 1.000\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 41-by-41\n",
      "SPE threshold: 2.6409544615360288\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.818, recall: 0.681, F1-measure: 0.743\n",
      "\n",
      "MCC: 0.745, ROC_AUC: 0.840\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.164, recall: 0.625, F1-measure: 0.260\n",
      "\n",
      "MCC: 0.320, ROC_AUC: 0.812\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.253, recall: 1.000, F1-measure: 0.404\n",
      "\n",
      "MCC: 0.498, ROC_AUC: 0.989\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.013, recall: 1.000, F1-measure: 0.025\n",
      "\n",
      "MCC: 0.112, ROC_AUC: 0.989\n",
      "\n",
      "Reduction percentage 0.999\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.006, recall: 1.000, F1-measure: 0.011\n",
      "\n",
      "MCC: 0.075, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.999, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.006, recall: 1.000, F1-measure: 0.011\n",
      "\n",
      "MCC: 0.075, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.050, recall: 1.000, F1-measure: 0.095\n",
      "\n",
      "MCC: 0.224, ROC_AUC: 1.000\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 41-by-41\n",
      "SPE threshold: 2.6409544615360288\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.818, recall: 0.681, F1-measure: 0.743\n",
      "\n",
      "MCC: 0.745, ROC_AUC: 0.840\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.019, recall: 1.000, F1-measure: 0.038\n",
      "\n",
      "MCC: 0.139, ROC_AUC: 1.000\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 506188-by-41\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-41\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.253, recall: 1.000, F1-measure: 0.404\n",
      "\n",
      "MCC: 0.498, ROC_AUC: 0.989\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.028, ROC_AUC: 0.989\n",
      "\n",
      "Training ratio 0.9 - Reduction percentage 0.99\n",
      "Reduction percentage 0.0\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.850, recall: 1.000, F1-measure: 0.919\n",
      "\n",
      "MCC: 0.837, ROC_AUC: 0.912\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.145, recall: 0.997, F1-measure: 0.253\n",
      "\n",
      "MCC: 0.344, ROC_AUC: 0.910\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.904, recall: 0.999, F1-measure: 0.950\n",
      "\n",
      "MCC: 0.949, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.998, recall: 0.980, F1-measure: 0.989\n",
      "\n",
      "MCC: 0.989, ROC_AUC: 0.990\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 22-by-22\n",
      "SPE threshold: 1.7492366241825603\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.151, recall: 0.695, F1-measure: 0.248\n",
      "\n",
      "MCC: 0.324, ROC_AUC: 0.847\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.960, recall: 0.726, F1-measure: 0.827\n",
      "\n",
      "MCC: 0.831, ROC_AUC: 0.863\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.020\n",
      "\n",
      "MCC: 0.099, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.504, recall: 1.000, F1-measure: 0.670\n",
      "\n",
      "MCC: 0.699, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.25\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.850, recall: 1.000, F1-measure: 0.919\n",
      "\n",
      "MCC: 0.837, ROC_AUC: 0.912\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.112, recall: 0.995, F1-measure: 0.202\n",
      "\n",
      "MCC: 0.303, ROC_AUC: 0.909\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.876, recall: 0.995, F1-measure: 0.932\n",
      "\n",
      "MCC: 0.932, ROC_AUC: 0.996\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.998, recall: 0.956, F1-measure: 0.976\n",
      "\n",
      "MCC: 0.976, ROC_AUC: 0.978\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 22-by-22\n",
      "SPE threshold: 1.7492366241825603\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.151, recall: 0.695, F1-measure: 0.248\n",
      "\n",
      "MCC: 0.324, ROC_AUC: 0.847\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.948, recall: 0.740, F1-measure: 0.831\n",
      "\n",
      "MCC: 0.834, ROC_AUC: 0.869\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.020\n",
      "\n",
      "MCC: 0.099, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.433, recall: 1.000, F1-measure: 0.604\n",
      "\n",
      "MCC: 0.648, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.5\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.850, recall: 1.000, F1-measure: 0.919\n",
      "\n",
      "MCC: 0.837, ROC_AUC: 0.912\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.078, recall: 0.998, F1-measure: 0.145\n",
      "\n",
      "MCC: 0.253, ROC_AUC: 0.910\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.825, recall: 0.996, F1-measure: 0.903\n",
      "\n",
      "MCC: 0.905, ROC_AUC: 0.997\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.966, F1-measure: 0.981\n",
      "\n",
      "MCC: 0.981, ROC_AUC: 0.983\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 22-by-22\n",
      "SPE threshold: 1.7492366241825603\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.151, recall: 0.695, F1-measure: 0.248\n",
      "\n",
      "MCC: 0.324, ROC_AUC: 0.847\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.925, recall: 0.747, F1-measure: 0.827\n",
      "\n",
      "MCC: 0.829, ROC_AUC: 0.873\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.020\n",
      "\n",
      "MCC: 0.099, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[ 1  1  1 ... -1  1  1]\n",
      "Precision: 0.337, recall: 1.000, F1-measure: 0.504\n",
      "\n",
      "MCC: 0.572, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.75\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.850, recall: 1.000, F1-measure: 0.919\n",
      "\n",
      "MCC: 0.837, ROC_AUC: 0.912\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.041, recall: 0.995, F1-measure: 0.078\n",
      "\n",
      "MCC: 0.182, ROC_AUC: 0.909\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.702, recall: 0.998, F1-measure: 0.824\n",
      "\n",
      "MCC: 0.836, ROC_AUC: 0.997\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.993, recall: 0.964, F1-measure: 0.978\n",
      "\n",
      "MCC: 0.978, ROC_AUC: 0.982\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 22-by-22\n",
      "SPE threshold: 1.7492366241825603\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.151, recall: 0.695, F1-measure: 0.248\n",
      "\n",
      "MCC: 0.324, ROC_AUC: 0.847\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.857, recall: 0.727, F1-measure: 0.787\n",
      "\n",
      "MCC: 0.788, ROC_AUC: 0.863\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.020\n",
      "\n",
      "MCC: 0.099, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.203, recall: 1.000, F1-measure: 0.337\n",
      "\n",
      "MCC: 0.443, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.99\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.850, recall: 1.000, F1-measure: 0.919\n",
      "\n",
      "MCC: 0.837, ROC_AUC: 0.912\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.002, recall: 1.000, F1-measure: 0.003\n",
      "\n",
      "MCC: 0.036, ROC_AUC: 0.911\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.082, recall: 1.000, F1-measure: 0.152\n",
      "\n",
      "MCC: 0.287, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.812, recall: 0.812, F1-measure: 0.812\n",
      "\n",
      "MCC: 0.812, ROC_AUC: 0.906\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 22-by-22\n",
      "SPE threshold: 1.7492366241825603\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.151, recall: 0.695, F1-measure: 0.248\n",
      "\n",
      "MCC: 0.324, ROC_AUC: 0.847\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.177, recall: 0.688, F1-measure: 0.282\n",
      "\n",
      "MCC: 0.349, ROC_AUC: 0.843\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.020\n",
      "\n",
      "MCC: 0.099, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.019\n",
      "\n",
      "MCC: 0.096, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.999\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.850, recall: 1.000, F1-measure: 0.919\n",
      "\n",
      "MCC: 0.837, ROC_AUC: 0.912\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.000, recall: 1.000, F1-measure: 0.000\n",
      "\n",
      "MCC: 0.009, ROC_AUC: 0.911\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.996, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.006, recall: 1.000, F1-measure: 0.011\n",
      "\n",
      "MCC: 0.075, ROC_AUC: 0.998\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.250, recall: 1.000, F1-measure: 0.400\n",
      "\n",
      "MCC: 0.500, ROC_AUC: 1.000\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 22-by-22\n",
      "SPE threshold: 1.7492366241825603\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.151, recall: 0.695, F1-measure: 0.248\n",
      "\n",
      "MCC: 0.324, ROC_AUC: 0.847\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.019, recall: 1.000, F1-measure: 0.038\n",
      "\n",
      "MCC: 0.139, ROC_AUC: 1.000\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502551-by-22\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-22\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.020\n",
      "\n",
      "MCC: 0.099, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.001\n",
      "\n",
      "MCC: 0.024, ROC_AUC: 0.985\n",
      "\n",
      "Training ratio 0.9 - Reduction percentage 0.999\n",
      "Reduction percentage 0.0\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.876, recall: 0.658, F1-measure: 0.751\n",
      "\n",
      "MCC: 0.753, ROC_AUC: 0.828\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.997, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.997, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.898, recall: 0.820, F1-measure: 0.857\n",
      "\n",
      "MCC: 0.854, ROC_AUC: 0.909\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 0.822, F1-measure: 0.903\n",
      "\n",
      "MCC: 0.904, ROC_AUC: 0.911\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 21-by-21\n",
      "SPE threshold: 1.7019764758852765\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.013, recall: 0.533, F1-measure: 0.026\n",
      "\n",
      "MCC: 0.084, ROC_AUC: 0.766\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.960, recall: 0.726, F1-measure: 0.827\n",
      "\n",
      "MCC: 0.831, ROC_AUC: 0.863\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57507-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.032, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.508, recall: 1.000, F1-measure: 0.674\n",
      "\n",
      "MCC: 0.703, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.25\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.840, recall: 0.653, F1-measure: 0.735\n",
      "\n",
      "MCC: 0.736, ROC_AUC: 0.825\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.997, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.997, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.866, recall: 0.807, F1-measure: 0.835\n",
      "\n",
      "MCC: 0.832, ROC_AUC: 0.902\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 0.806, F1-measure: 0.893\n",
      "\n",
      "MCC: 0.896, ROC_AUC: 0.903\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 21-by-21\n",
      "SPE threshold: 1.7019764758852765\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.013, recall: 0.533, F1-measure: 0.026\n",
      "\n",
      "MCC: 0.084, ROC_AUC: 0.766\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.948, recall: 0.740, F1-measure: 0.831\n",
      "\n",
      "MCC: 0.834, ROC_AUC: 0.869\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 57086-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.032, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.437, recall: 1.000, F1-measure: 0.608\n",
      "\n",
      "MCC: 0.651, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.5\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.781, recall: 0.664, F1-measure: 0.718\n",
      "\n",
      "MCC: 0.716, ROC_AUC: 0.831\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.779, recall: 0.660, F1-measure: 0.715\n",
      "\n",
      "MCC: 0.713, ROC_AUC: 0.829\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 0.784, F1-measure: 0.879\n",
      "\n",
      "MCC: 0.884, ROC_AUC: 0.892\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 21-by-21\n",
      "SPE threshold: 1.7019764758852765\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.013, recall: 0.533, F1-measure: 0.026\n",
      "\n",
      "MCC: 0.084, ROC_AUC: 0.766\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.925, recall: 0.747, F1-measure: 0.827\n",
      "\n",
      "MCC: 0.829, ROC_AUC: 0.873\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56665-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.032, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[ 1 -1  1 ...  1  1  1]\n",
      "Precision: 0.341, recall: 1.000, F1-measure: 0.508\n",
      "\n",
      "MCC: 0.575, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.75\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.631, recall: 0.639, F1-measure: 0.635\n",
      "\n",
      "MCC: 0.632, ROC_AUC: 0.818\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.997, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.997, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.679, recall: 0.810, F1-measure: 0.739\n",
      "\n",
      "MCC: 0.740, ROC_AUC: 0.904\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 0.781, F1-measure: 0.877\n",
      "\n",
      "MCC: 0.883, ROC_AUC: 0.891\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 21-by-21\n",
      "SPE threshold: 1.7019764758852765\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.013, recall: 0.533, F1-measure: 0.026\n",
      "\n",
      "MCC: 0.084, ROC_AUC: 0.766\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.857, recall: 0.727, F1-measure: 0.787\n",
      "\n",
      "MCC: 0.788, ROC_AUC: 0.863\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 56244-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.032, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[ 1  1  1 ...  1  1 -1]\n",
      "Precision: 0.205, recall: 1.000, F1-measure: 0.341\n",
      "\n",
      "MCC: 0.447, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.99\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.071, recall: 0.750, F1-measure: 0.130\n",
      "\n",
      "MCC: 0.230, ROC_AUC: 0.874\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.997, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.997, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.082, recall: 0.875, F1-measure: 0.150\n",
      "\n",
      "MCC: 0.267, ROC_AUC: 0.936\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 0.688, F1-measure: 0.815\n",
      "\n",
      "MCC: 0.829, ROC_AUC: 0.844\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 21-by-21\n",
      "SPE threshold: 1.7019764758852765\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.013, recall: 0.533, F1-measure: 0.026\n",
      "\n",
      "MCC: 0.084, ROC_AUC: 0.766\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.177, recall: 0.688, F1-measure: 0.282\n",
      "\n",
      "MCC: 0.349, ROC_AUC: 0.843\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55839-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.032, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.010, recall: 1.000, F1-measure: 0.019\n",
      "\n",
      "MCC: 0.097, ROC_AUC: 0.985\n",
      "\n",
      "Reduction percentage 0.999\n",
      "------------------------------ CURRENTLY: LR ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.996, recall: 0.867, F1-measure: 0.927\n",
      "\n",
      "MCC: 0.871, ROC_AUC: 0.932\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.000, recall: 0.000, F1-measure: 0.000\n",
      "\n",
      "MCC: -0.000, ROC_AUC: 0.499\n",
      "\n",
      "------------------------------ CURRENTLY: SVM ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.997, recall: 1.000, F1-measure: 0.998\n",
      "\n",
      "MCC: 0.997, ROC_AUC: 0.998\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.000, recall: 0.000, F1-measure: 0.000\n",
      "\n",
      "MCC: -0.000, ROC_AUC: 0.499\n",
      "\n",
      "------------------------------ CURRENTLY: DecisionTree ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 1.000, recall: 1.000, F1-measure: 1.000\n",
      "\n",
      "MCC: 1.000, ROC_AUC: 1.000\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.000, recall: 0.000, F1-measure: 0.000\n",
      "\n",
      "MCC: 0.000, ROC_AUC: 0.500\n",
      "\n",
      "------------------------------ CURRENTLY: PCA ------------------------------\n",
      "====== Transformed train data summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n",
      "n_components: 4\n",
      "Project matrix shape: 21-by-21\n",
      "SPE threshold: 1.7019764758852765\n",
      "\n",
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.013, recall: 0.533, F1-measure: 0.026\n",
      "\n",
      "MCC: 0.084, ROC_AUC: 0.766\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "Precision: 0.019, recall: 1.000, F1-measure: 0.038\n",
      "\n",
      "MCC: 0.139, ROC_AUC: 1.000\n",
      "\n",
      "------------------------------ CURRENTLY: IsolationForest ------------------------------\n",
      "====== Transformed train data summary ======\n",
      "Train data shape: 502415-by-21\n",
      "\n",
      "====== Model summary ======\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/ensemble/_iforest.py:248: FutureWarning: 'behaviour' is deprecated in 0.22 and will be removed in 0.24. You should not pass or set this parameter.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====== Transformed test data summary ======\n",
      "Test data shape: 55824-by-21\n",
      "\n",
      "Train accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.002\n",
      "\n",
      "MCC: 0.032, ROC_AUC: 0.985\n",
      "\n",
      "Test accuracy:\n",
      "====== Evaluation summary ======\n",
      "[1 1 1 ... 1 1 1]\n",
      "Precision: 0.001, recall: 1.000, F1-measure: 0.001\n",
      "\n",
      "MCC: 0.024, ROC_AUC: 0.985\n",
      "\n",
      "It took: 2627.8755111694336\n"
     ]
    }
   ],
   "source": [
    "from loglizer.models import *\n",
    "from loglizer import dataloader, preprocessing, utils\n",
    "from imblearn.over_sampling import SMOTE, RandomOverSampler\n",
    "import pandas as pd\n",
    "from sklearn.utils import shuffle\n",
    "import numpy as np\n",
    "import time\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from collections import defaultdict\n",
    "from matplotlib.lines import Line2D\n",
    "from sklearn.metrics import precision_recall_fscore_support, matthews_corrcoef, roc_auc_score, confusion_matrix\n",
    "\n",
    "start = time.time()\n",
    "\n",
    "supervised = ['LR', 'SVM', 'DecisionTree']\n",
    "unsupervised = ['PCA', 'IsolationForest', 'InvariantsMiner']\n",
    "semi_supervised = ['LogClustering']\n",
    "\n",
    "#Load data\n",
    "label_file = 'data/HDFS/anomaly_label.csv'\n",
    "path = '../../../loglizer/HDFS_Data/'\n",
    "\n",
    "#dfs = []\n",
    "for train_ratio in [0.5, 0.75, 0.9]:\n",
    "    for reduction_train in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999]:\n",
    "        print('Training ratio', train_ratio, '- Reduction percentage', reduction_train)\n",
    "        #Load training data\n",
    "        data_train = np.load(path + 'train/' + str(train_ratio) + '/'+ str(reduction_train) +'/HDFS_train_' + str(train_ratio) + '_' + str(reduction_train) +  '.npz', allow_pickle=True)\n",
    "        x_tr, y_train = data_train['x_data'], data_train['y_data']\n",
    "\n",
    "        #Shuffle\n",
    "        indexes_train = shuffle(np.arange(len(x_tr)))\n",
    "        x_tr, y_train = x_tr[indexes_train], y_train[indexes_train]\n",
    "\n",
    "        for reduction_test in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999]:\n",
    "            print('Reduction percentage', reduction_test)\n",
    "\n",
    "            #Load test data\n",
    "            data_test = np.load(path + 'test/' + str(train_ratio) + '/'+ str(reduction_test) +'/HDFS_test_' + str(train_ratio) + '_' + str(reduction_test) + '.npz', allow_pickle=True)\n",
    "            x_te, y_test = data_test['x_data'], data_test['y_data']\n",
    "\n",
    "            #Shuffle\n",
    "            indexes_test = shuffle(np.arange(len(x_te)))\n",
    "            x_te, y_test = x_te[indexes_test], y_test[indexes_test]\n",
    "\n",
    "            for _model in supervised + unsupervised + semi_supervised:\n",
    "                print('-'*30,'CURRENTLY:', _model,'-'*30)\n",
    "                #If model is supervised, generate new samples by randomly sampling with replacement the current available samples to balance both classes.\n",
    "\n",
    "                if(_model in supervised):\n",
    "                    feature_extractor = preprocessing.FeatureExtractor()\n",
    "                    x_train = feature_extractor.fit_transform(x_tr, term_weighting='tf-idf')\n",
    "\n",
    "                    #Oversample anomalies\n",
    "                    sm = RandomOverSampler(random_state=2)\n",
    "                    x_train_sm, y_train_sm = sm.fit_sample(x_train, y_train)\n",
    "\n",
    "                    if(_model == 'SVM'):\n",
    "                        model = SVM()\n",
    "                    elif(_model == 'DecisionTree'):\n",
    "                        model = DecisionTree()\n",
    "                    else:\n",
    "                        model = LR()\n",
    "                    model.fit(x_train_sm, y_train_sm)\n",
    "\n",
    "                else:\n",
    "                    if _model == 'PCA':\n",
    "                        feature_extractor = preprocessing.FeatureExtractor()\n",
    "                        x_train = feature_extractor.fit_transform(x_tr, term_weighting='tf-idf', \n",
    "                                                                  normalization='zero-mean')\n",
    "                        model = PCA()\n",
    "                        model.fit(x_train)\n",
    "\n",
    "                    elif _model == 'InvariantsMiner':\n",
    "                        feature_extractor = preprocessing.FeatureExtractor()\n",
    "                        x_train = feature_extractor.fit_transform(x_tr)\n",
    "                        #model = InvariantsMiner(epsilon=0.5)\n",
    "                        model = InvariantsMiner(epsilon=0.001)\n",
    "                        model.fit(x_train)\n",
    "\n",
    "                    elif _model == 'LogClustering':\n",
    "                        feature_extractor = preprocessing.FeatureExtractor()\n",
    "                        x_train = feature_extractor.fit_transform(x_tr, term_weighting='tf-idf')\n",
    "                        model = LogClustering(max_dist=0.3, anomaly_threshold=0.3)\n",
    "                        model.fit(x_train[y_train == 0, :]) # Use only normal samples for training\n",
    "\n",
    "                    elif _model == 'IsolationForest':\n",
    "                        feature_extractor = preprocessing.FeatureExtractor()\n",
    "                        x_train = feature_extractor.fit_transform(x_tr)\n",
    "                        model = IsolationForest(random_state=2019, max_samples=0.9999, contamination=0.03, \n",
    "                                                n_jobs=4)\n",
    "                        model.fit(x_train)\n",
    "\n",
    "                x_test = feature_extractor.transform(x_te)\n",
    "\n",
    "                print('Train accuracy:')\n",
    "                if(_model in supervised):\n",
    "                    precision_train, recall_train, f1_train, mcc_train, roc_auc_train, tn_train, fp_train, fn_train, tp_train = model.evaluate(x_train_sm, y_train_sm)\n",
    "                else:\n",
    "                    precision_train, recall_train, f1_train, mcc_train, roc_auc_train, tn_train, fp_train, fn_train, tp_train = model.evaluate(x_train, y_train)                \n",
    "\n",
    "                print('Test accuracy:')\n",
    "                precision_test, recall_test, f1_test, mcc_test, roc_auc_test, tn_test, fp_test, fn_test, tp_test = model.evaluate(x_test, y_test)\n",
    "\n",
    "                results = [('model_name', _model), ('model', model), \n",
    "                           ('f1-train', f1_train), ('f1-test', f1_test),\n",
    "                           ('train_ratio', train_ratio),  ('reduction_train', reduction_train), \n",
    "                           ('reduction_test', reduction_test), \n",
    "                          ('precision_train', precision_train), ('precision_test', precision_test),\n",
    "                           ('recall_train', recall_train),('recall_test', recall_test), \n",
    "                           ('mcc_train', mcc_train), ('mcc_test', mcc_test),\n",
    "                          ('roc_auc_train', roc_auc_train), ('roc_auc_test', roc_auc_test),\n",
    "                           ('tn_train', tn_train),('tn_test', tn_test), \n",
    "                           ('fp_train', fp_train), ('fp_test', fp_test), \n",
    "                          ('fn_train', fn_train), ('fn_test', fn_test),\n",
    "                           ('tp_train', tp_train), ('tp_test', tp_test)]\n",
    "\n",
    "                dfs.append(pd.DataFrame({x[0]:x[1:] for x in results}))\n",
    "            \n",
    "end = time.time()\n",
    "\n",
    "print('It took:', end-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for model in supervised + unsupervised + semi_supervised:\n",
    "    results = pd.concat(dfs)\n",
    "    results[results.model_name == model].groupby(['reduction', 'train_ratio']).agg(np.mean)[['f1-test', 'f1-train']].plot()\n",
    "\n",
    "    train_ratios = [0.5, 0.75, 0.9]\n",
    "    reduction_percentages = [0.0, 0.25, 0.50, 0.75, 0.99, 0.999]\n",
    "    xticks = [str((x,y)) for x in reduction_percentages for y in train_ratios]\n",
    "    plt.ylim(0, 1)\n",
    "    for i in range(3, 16, 3):\n",
    "        _ = plt.axvline(x=i,color='r', linestyle='--')\n",
    "        _ = plt.xticks(range(len(xticks)), labels=xticks, rotation=45)\n",
    "        plt.title(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAE2CAYAAAByaKVZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsnXd4FNUaxt/ZTSUJKQRIIKQCgVAkEECaJHREBRvCRQQVUOyi2FCKBVGwIiqIqBQVBK4gTeDSpChNegkkJIQe0khI3/3uH18mu5tskt1kN7ObPb/nmWdmZ2Znzswm8845X5OICAKBQCAQVIVK6QYIBAKBwD4QgiEQCAQCkxCCIRAIBAKTEIIhEAgEApMQgiEQCAQCkxCCIRAIBAKTqFIwJElaJEnSDUmSTlSwXZIk6UtJks5LknRMkqSOlm+mQCAQCJTGlB7GjwAGVbJ9MIAWJdMEAN/UvFkCgUAgsDWqFAwi2gUgvZJdhgJYTMzfAHwkSQq0VAMFAoFAYBtYwobRFECK3udLJesEAoFAUIdwqs2TSZI0ATxsBV+VqlO4h0fZHSr+XNFy2XWmfMeUZaKqJ63WtP2I+NjyJJ/L1M9l1xmjqhQvFW2X18vtrGjZ1P3MpaLrsdb39H+36p5XpeJJkoz/LdgKfn5AWJjSrRBYiAuZF5CRlwG6QjeJqKESbbCEYFwG0Ezvc1DJunIQ0QIACwAgJiaGDh48aIHTCwTVQKMB8vKA3FyeKlqubFthIeDqypObm+nLFW1Xq7ldGg0LUHWXNRpg61ZgzhzgtdeA119X+m4Lasim85sweNlgTOs9DTPiZiQr1Q5LCMZaAM9JkvQrgK4AsojoqgWOKxBYD7Ua8PTkqS4yYACQkgK8+SbQti0wZIjSLRJUk9uFtzFx/US08m+FN3u+iRmYoVhbqhQMSZJ+ARALwF+SpEsApgFwBgAi+hbABgB3AzgPIBfA4yad+fjxirvLlQ03VLSt7JBURUM5pqxTqQAfH8DXl7v1Vc3r1+fvCAS2giQBixYB8fHAf/4D/P030Lq10q0SVINpO6YhKTMJu8bugquTq6JtkZRKbx7j708Hjb31VNYeU8bh5XlFy6asKy4GsrKA9HQgI4Pn+fkVt0tfYIyJiqXfYmVRU6t14+nVXXZ2BoKCWLzr6tu2I3PxIhATw3+f+/fzXGA3HL56GJ2/64xx0eMw/975AABJkg4RUYwS7VFOMOzNhpGXx+IhC4j+3Ng6/W0ajdKtNw1/fxaOslN4OBAcDLi4KN1CQXX46y+gb1+e1q3jlwWBzVOsLUbXhV1xJfsKTj97Gj5uLPZKCkateknZNe7uPDVpYt73iICCgup79VR0TNnQKRs7q7tcUMBj3RcuAImJPD98GPjvf4GiIt05VSqgaVPjYhIWBgQGimE5W6VXL+Crr4CnngLeeAOYPVvpFglM4Iu/v8Dhq4fx28O/lYqF0gjBsDaSxJ4wtky3buXXaTTAlSssIPpicuECe+BcuWI4ROjqCoSEGApHRXaisvPKtqlUgJMTvxXLc/3liubG1rm6Av36OeZY/oQJwNGj7DnVvj0werTSLRJUwoWMC5i6YyrubXkvHmz9oNLNKUUIhsA4ajXQrBlPd91VfntBAZCcrBMRWVSuXy8fk2CqDcnYPnKPqLhY5zIqL5edl11XEV27Ao8/DowYAXh71+w+2ROffw6cPAmMHw+0agV07qx0iwRGICI8s+EZqCQV5t09D5IlRydqiLBh6FNczLaHvDx+UIohFvtGqzUUk4wMYMUK4Icf+MHp5gY88ACLR58+jvF7p6ayUBQVAQcPco9QYFP8fPxnjFo9Cl8O+hLPd32+3HbHNHqHhdHBWbN0toHKJmdn82wAREBODpCWZjilp5dfpz9lZemO4evLQzU9evDUuTNQr57lb4Sg9iHih+UPPwC//AJkZrJRf8wYYOxYtsvUZY4eBbp356Gp7dttf8jUgUjLTUPrea0R7huOPU/sgVpV3kHBMQVDksjk/oVKVbmguLkBt28bPvz1DbZlqV8faNCg4snZmR8oe/YAp0/zd5ycgOhoFo/u3XlurgFcYHvk5wO//87isWULi0lsLAvHQw8BZdPX1BVWreLrGzuW4zVsaNjDkXlizRNYcmwJDk04hPaN2xvdxzEFo107Orh8OQ//WGLy8KhYAPz8DJednU1vaFoaBz3t2cPT/v26mIzQUJ149OjBEbXCZdF+SUkBFi8GfvwROH+e41IeeYSHrLp3r3sP1WnTgHffZdvGiy8q3RqHZ9uFbei7uC/e7PkmZvadWeF+jikYtmjDMIXCQuDIERaPvXt5frUkE4qXF3DnnbpeyJ138jqlKS4Gbt4Ebtzg6fp13Tw7Gxg8mCchdgwRsHs39zpWrODea8uW/Db+2GPsXlwX0GqBBx8E/vgD2LSJPcgEipBXlIf233KP4tjTx+Du7F7hvkIw7BkiIClJJx579nDaEyIeSmvfHmjXjntA9erpJnf3qpf1P5ftFeXmGj74jYmBPE9LMx4l7+zMrqY5ORzt/eSTwBNP8Hi+gMnJAVauZPHYtYt/0wEDuNcxdCjfP3smO5ttdVeucO+5eXOlW+SQTPnfFMzcPRP/e+x/6BPWp9J9hWDUNW7dMhzGOnfOMMtpde65k5POZnP7Nk/GqF8faNQIaNy46rm3N/c+1q0DFiwA/vyTjzF4MLteDhli3vBdXSchgYerfvqJh6/kGA/9HGRl09LrT1VtV6v5fjs58WTKckXb3NyA557j2JiqSExkp46AAGDfPv4bshbFxdyjSUtjDzU/P+udy044dv0YOi3ohEfbP4ofhv5Q5f5CMBwJIh7WMiWVtrHPubncWzEmAA0bsqBUl6QkNoB+/z2/cQYGco/jySdFXQV9NBpg2zZgxw5erqw2ijnr5fiR4mJ22jC2XNk2/eX0dGDKFOD99027pm3buOc0ZAhH+VvaxfjmTWDhQuDrr1lsAU41M2wYD/UNGOCQQ6IarQY9FvVAQkYCzjx7Bg3qNajyO0oKBohIkalTp04ksFGKiojWrCG65x4ilYpIkogGDCBauZKosFDp1glMoUULooceMu87X37J0vX225Zrx5EjRE8+SeTmxsfu25f/tg4dInrxRaIGDXh9kyZEb7xBdOaM5c5tB8z9Zy5hOmjp0aUmfwfAQVLouS0EQ1A5Fy8STZ9OFBTEfy6NGhG9/jrRuXNKt0xQGffcQ9S2rXnf0WqJnniCf+cVK6p/7qIifrno3ZuP5e5O9NRTRMePl9+3oIBo1Sqie+8lUqt5/27diBYsIMrMrH4b7ICLmRfJc6YnDVwykLRarcnfE4IhsH2Ki4nWrycaOlT3j92nD9GvvxLl5yvdOtMw45/S7nn1VSJXV/7dzCE/n6h7d6J69Yj+/de87968STRrFlGzZvz3ERpKNGcOUXq6ad+/epVo9myiqCid0IwaRbRlC5FGY15bbBytVkv3/nwvub/vTonpiWZ9VwiGwL64dInovfeIQkL4T8jfn+iVV5QfTtBqidLSiA4cIFq+nOjDD4nGj+dhkNBQIk9PfoA5At99x79NonkPIyLiB3dQEFFwMNH161Xvf/Qo0bhxumGnPn2Ifv/dfLGS0WqJ9u8nmjiRyMeHjxkcTPTOO0Tnz1fvmDbGypMrCdNBs/fMNvu7SgqGMHoLqo9Wy9HRCxYAa9eysbVDBzaWy0GSctCkvKz/2dvbfONqUREXBUpM5CkhQbecmGiY3gVgh4DwcJ5u3OBMu7Nm1f0613/9xUkjN24EBg0y//uHDgE9e7L31Nat5WuhFBfzb/7ll8DOnexsMXo08PzzHMBqKfLzgTVr2K1582Z2DrjrLnZrfughuyz6lZmfidbzWiPQMxD7x++Hk8q8HLDCS0pg/1y7xi6n27fr8nalp5d/gOujUukqE5YVFnk5O9tQEC5eZKGScXHR1eUIDwciInTLZasIFhdz4N0vvwAffAC89ZbVbofi3LjBnnM1ieL++Wdg1Cjg6aeBb77hdenp7O00bx7/FiEh7L77xBPWd5G9dAlYsoTF49w59hZ8+GH2srrrLruJxH963dP47vB32D9uPzo16WT294VgCOouxcW6CoRy8kdjy2U/6wtN48Y6ESg7NWliXi+luJgfMMuWAe+9B7z9tsUv2SYg4gf4iBG6h311eP114OOPOY3IlSvA0qXs6h0XB7zwAnDvvbXvDkvE8SI//AAsX84vFXaS3mT3xd3o9UMvTLpzEj4Z+Em1jiEEQyAoiyw07u6WH3bQaHhIY8kSYPp0fhjWRbp14/u3bVv1j6HRAPfdB2zYwMd69FEedmrXznLtrAm5uVyzPCSEh99smILiAnSY3wF5RXk4+cxJeLhUL7GlKNEqEJTFyYkDEa2BWs1vp2o1C4ZWy3M7GdIwmchItjHVBLUa+PVXtiPcfbftRWbXq8c529av556HDf+Gs3bPwpmbZ7DhPxuqLRZK4wAVYwQCI6jVHNH+xBOcsXXq1OqlbLFlIiN5GCk7u2bH8fLinoWtiYVMdDTbbOQkoDbI6dTTmLl7Jka2HYnBLQYr3ZxqI3oYAsdFpQK++47n77/PPY3337fpt1SziIzk+dmzPGxTV+nYkeeHD9tkjRotaTFh3QR4OHvg80GfK92cGiF6GALHRqUC5s8HJkwAZs4E3nyz7vQ0WrXi+dmzyrbD2txxB4v8v/8q3RKjLDy8ELsv7sYnAz5BI49GSjenRogehkCgUrEnkVoNfPQR9zQ++sj+exoREXxtdV0wPD25Xsnhw0q3pBxXs6/itS2vIS40DmM7jFW6OTVGCIbAIdBoNbhddBteLl6QjAmBSsWxBSoVMHs2ewfNmWPfouHqyrEodV0wALZj7NundCsMyMrPwoMrHkR+cT7m3zPf+N+dnSEEQ2AXEBFyi3KRkZ+BzPxMZOSVzEs+l64rKL8tIy8D2YVs+A30DET3Zt3RvVl39GjWA9GB0XBRl0QxSxIwdy6Lxqefck/j00/tWzQiIx1DMDp2ZG+utDQO+lSY9Lx0DFw6EEeuHcGvD/6KFg1aKN0ki6CYYKTlpmHNmTXwcfOBt5s3fNx84OPmg/qu9aGSLG9aISKk56Xj+u3ruJ5zvXR+LecaL+utz8zPhH89fwR4BiDAMwCBnoFGlxt7NtY9bKwMEaFAU4Ccwhy4qF0qflO2cYq1xcjMz0R6XnqlU1peGtLz0ksf/pn5mSjSFlV6bC8XL/i4+cDX3Rc+bj4I9QmFrxsv+7r5wt3ZHceuH8OelD1YdXoVAMDNyQ2dm3QuFZHuzbrD/4svWDQ+/5xF4/PPrS4aWuLodYv/7UdGcvS9Vmv5Ghe2RHQ0z48cAfr2VbQpqbdTMWDpAJxKPYXVw1fj3sh7FW2PJVEscE9qIhGeMr6tvmt9FhJXnZB4u3nDx9XH8LOeyOQU5pQ+8EtFQE8Ybty+YfSBo5bUaOTRqFQAGns0ho+bD9Ly0nAt5xqu5VzD1eyrSMtLM9rWBu4NdGLiFYgAjwDDz54B8HTxRHZBNrILs6ueV7KtWFts0G79h6Ovmy983X0NHpDy57L7eLt6Q62qPDqXiFCsLUZ+cX7plFecZ/DZ2HS78LbuwZ9fRgRy05BVUHGqEAkSfNx84OfuBz93P/i6+8LP3Q8+robXKf/u+uu83bzNyslzJfsK9qXsw56UPdibsheHrx4u/fto2aAlugd1w6SVV9Bu2RbQM89A+uqrGomGRqvB1ZyrSM5MRlJmkm7KSkJyZjKSs5IhQUKLBi0Q2SCSJ3/d3MfNp3onXrAAeOopLo5lQvU9IkJqbioS0hOQkJGA8+nnkZCRgPS8dHQL6oYBEQPQKbBTlX8/tU5aGuDvz1Hpkycr1oxrOdfQb3E/JGQk4PdHfsfA5gMtfg6HjPRuF92OFm9YXPr2mFWQVbpc9nNWfpbBevltrCKcVc5o5NGoVAAaezZGgEeAwWd57ufuZ9JbXaGmEDdu3ygVEFlMruVcw9WcqwbL+cX5Zt0LtaSGl6sXvFy8Kp6XLHu6eKJQU2gw7JKRn2H4OS+jyrdxWYy93bzLCYM8VXWfK0IlqUof+gaTm5F1epOPm49iD6K8ojwcvHIQe1P2Yu+lvdibshc3b9/Ex1uAyXuB9f2CcWTKOHQP6YkuTbuUC7zSaDW4kn3FQAySs3TicDHrYrnfpLFHY4T6hCLEJwSh3qHQkAZn087i7M2zSMxIhIY0pfs28mhkICQtG7REZINIhPuGw1ldSRndnTuB2FguvztgQGlbU26llIpCQnoCzmecL/2cU5hT+nUJEpp5N4OXixdOpp4EAPi6+aJPWB/0D++P/hH9Ee4bXsO7byFCQzm6/ZdfFDn95VuX0WdxH1y6dQnrRq5DXFicVc7jkIJR3dQgRIScwpxywuLh7FHaS/B181VsuIaIcKvgloGg5BTmVCgEni6ecHNys2h75fF+fQHRXy4d28/PQFZ+FpzVznBzcoOb2o3nZSZ3Z3ej641N9ZzrWW1YsTYhIpxLP4e9F/eg8fufY/DqY5jfCZg4BFCp1egQ0AGt/FuVikTKrRSDHiDA9pIQnxCE+oQi1DuU5yUCEewdjHrO9So8f6GmEIkZiTh78yzOpp1FfFp8qZik5qaW7uekckK4b3i5Xklzv+bIyM/ApbMHMaDXGPz2VC/80MsTCRkJuJBxwUC8XNQuCPcNR4RvBE9+PG/u1xyhPqFwdXIFANzMvYn/Jf4PmxM2Y0viFqTc4lKr4b7hGBA+AP0j+qNPWJ/q94Zqyv33A6dPA2fO1PqpkzOT0WdxH6TeTsWGURvQM7in1c4lBEMgsGWIuD72hx/i4sMD8d2ETthzeR8SMhIQVD+oVBBKxcEnFMHewXBzcrNKczLyMkrF42za2dLl8+nnUaApKNN2IHMWsDzaGd+ObYsIvwg0921eKgoRfhFo6tXU7J4dESE+LR5bErdgc8JmbE/ajpzCHKgkFTo36Vza+7gz6M5as/Phvfc4L9itW7Wa9jwhPQF9FvdBVn4W/nz0T3QN6mrV8wnBEAhsHSJOH/L++5xORI4QtxWIoElMwM1df+L23h1QnT6NMxOHw7f3QHQcNhFOPn6Qtm612umLNEX45/I/2JKwBZsTN2P/5f3QkhaeLp6IDY1lAQnvj1b+razX+1+3jrPn7t4N9OhhnXOU4ezNs+i7uC/yivOwZfQWdAzsaPVziuSDAoGtI0mcc0ql4rlWy3Uhaju1t8yVK8DBg8CBAzwdPAh1WhoaA1wjpKgIoT1uA6O6AlFt2VPKijirndEzuCd6BvfEjLgZyMzPxPYL20t7IOvi1wEAguoHoV94PzzW/jHLj/HrpwipBcE4eeMk+i7uCy1psWPMDrRrbCMZfK2IEAyBwFQkCZgxg0Vj+nQgMxMYOJDrdchTQAAX9rEkaWk6cZDnV67wNrUaaNMGGDqUq+PFxHDq8Tvu4IJTALvWLlkC3L5t+bZVgI+bD+5vfT/ub30/AOBCxgVsSdyCLYlbsObMGiw5ugSrhq/C0FZDLXfSwECusFgLKUKOXjuKfkv6wVnljO1jtqN1w9ZWP6ctYJJgSJI0CMAXANQAFhLRrDLbgwH8BMCnZJ83iGiDhdsKIiA1VVeZU54SE/l/QaPhFz+NxvRlY9vc3IAnnwTeeANo2tTSVyGwe6ZN47f4d94Bfv+9/HYPD0MBKSso+p/LjrVnZ3N5VH1xuHBBt71lSy5eJItDdDSn+C5LeDj/cwC6JITx8bp4hVomzDcMEzpNwIROE5BdkI1+S/ph+Mrh2DhqI/qE9bHMSSSJexlWThFy8MpBDFgyAB4uHtj22LY6E5RnClXaMCRJUgOIB9AfwCUABwCMJKJTevssAPAvEX0jSVIUgA1EFFrZcSuyYRQXAykphoKgLwxlMzUHBfH/Rv36/LKlUvFcnvQ/V7ZN/3NyMlenVKmA8eNZOIKCTLmdAoeiqIjfYK5f5+naNd1y2c83bxo/hr64ZGRwVLb8PxkSwsIgi0OnTlwH3RSee457FZmZwIkTQPv2/Ec9cqRlrr2GpOel464f7kJSZhK2jdmGLk27WObAU6ZwLEZODqdGsTD7UvZh0LJB8HP3w7bHtiHMN8zi56gKW7dhdAFwnogSAUCSpF8BDAVwSm8fAlC/ZNkbwJWqDpqXxzVZyopCUhKLhoxcsjkigsv2RkToprAw7g1YgxkzOHnp/Pls3xw3joWjWTPrnE9ghzg7czptU1JqlxWXsoJy7RoHno0cqROImhSQiohgb6GMDKBFC377tqEUIX7uftg8ejN6/dALg5cNxs6xO9G2UduaHzg6mh8gJ06wwFqQXcm7MOTnIQjwDMC2x7ahmbfjPQxM6WE8BGAQEY0r+TwaQFciek5vn0AAmwH4AvAA0I+IDhk51gQAE/hTp04A9zC8vQ2FQH9q2lQ5uyLAAvbhh8CiRdzjePJJzoAthENg06xZAwwbBuzfzwIUHg507apYUFtFJGYkoucijlnY/cTumgcBJibyg2PBAh4esBD/S/wf7v3lXoT4hGDbY9sQ6BVosWObi5I9DEv5BY4E8CMRBQG4G8ASSSofuUVEC4gohohiwsKAf/7hnnpGBg/brljBD+dx43iYNjhYWbEAOHh0/nzg/HkuA71wIf89TpwIXLyobNsEggoJL3nw6hu+baiHIRPuG44to7egQFOAfov74Up2lYMTlRMWxm+gFjR8bzy3EUN+HoLmfs2xc+xORcVCaUwRjMsA9N+ng0rW6fMkgBUAQET7ALgB8K/soH5+QJcunFjSHnLohYQA337LwvHkk1zds3lz4Omn2eYhENgUFQmGtnrpXqxJm0ZtsHHURqTmpmLAkgFIyzWet80kJImHpSxk+F5zZg2GLR+GNo3aYPuY7XZfAKmmmGLDOACghSRJYWChGAHgP2X2uQigL4AfJUlqDRaMVNRBgoO51s6bbwKzZnGPY9Ei7n289ZZJ+d2qRVYWZz04fZrtPy4uPITu4qKbqvNZ6R6cwErIxnR9T6ncXODyZZscT+3StAvWjliLwcsG4+6f78bW0Vvh5epVvYNFR/PbXXEx4FT9yIHfTv6G/6z+DzoFdsKmRzcpl/LEliCiKifwMFM8gAQAU0rWvQvgvpLlKAB7ABwFcATAgKqO2alTJ6oLXLxI9MwzRC4uRE5OROPHE124UP3jpaUR/fUX0fz5RC++SNS/P1HTpkTsOmP5yd2daMYMosJCi90Sga3QrRtRXBwvb9vGP/iWLcq2qQrWnFlD6hlqivsxjvKK8qp3kCVL+FpPnKh2O5YeXUqqGSrquagnZeVnVfs41gDAQTLhuW2NSbHUIO3bx9Aff1SeGsScpgUEWMZj6vJlrqHz+uuAr6/p37t0iXsc333Hvf6xY7nHEVaB111qKnDqlG46eZLn16/r9qlXD4iKMpxat2YX4sJCnoqKjC+buu3oUWDVKnbKWbyYjy+oI4weDfz1F3tuXLnCHiRffQU8+6zSLauUpceWYvR/R2No5FCsHL7SrLT1APifqW1b/oMePdrs8y/6dxHGrR2H2NBYrB25Fp4utZeXyhQcMpeUJMWQ7CVlCVQqtim0acNT27Y8b9mSh19MoaiIM0Hv3ct/Z4sXm9eGmzc5m/RXX/H/qVbLD/3CQp2dhkgXIKjfdjc3wN2dRxI8PFgUPDy47U5OPDk789zPj0cVgoJ086ZNq+92vnIl22Ju32angxdesK00SZagsJCvz8fHPmxmFmHaNM59lZfHfzz16/PY6ZdfKt2yKvlq/1d4fuPzGN1+NH4c9qN52Y+Li/lan36aKyaaSHJmMt7Z/g6WHFuCgRED8d9H/gt3Z/dqtN662HochlUICeG/56ow5Z9bq+WXqJMneVqzRvdAdnJiN3RZSOSpRQv+H9JnyhQWiz59OObp4Yc5l5k+RNybkO0J+lOqntXGzQ3w8uKsDnJbXF35Ye/jw3/Pnp48OTuziBQV8d96cTEv5+dz/JH8WZ6nprJNoyyNGpUXEnluTFQuXwZefZWTfJ44AUyYALz8Mt+/H35gDzFjFBXxw1eecnJ0y4WF1bOnVGVLyc/na87MND5Vti0zk5+ZAHu4PfAAe5x27VrHbTjh4fzHl5zMf/CRkYqk/q4Oz3V5Dpn5mXhn+zvwdvXGl4O/ND1poZMTp0Yx0fCdnpeOmX/NxNz9cyFBwmvdX8OMuBlWyzZszygmGP7+/LJjDfLz2SFEFpCTJ7ly46pVumEuZ2fufcgCkpcHzJ7ND825c3mI5skngU8+MRSIM2f4ASnj68tDRUOH8nCOPAUH81v61avcllat2AZpqbfb7Gx+4KekcPv05wkJwI4dlYtK06Y8HJWczBko7r+f13Xvzsk+mzfn542nJ9tK9UWhqPLaTNVCkoyLSUEBP/ALCir/vpMT/xY+PrqpaVPdckoKsHw5//6ff86/dePGwH338W/Xt6/1gkAVIyKC54mJOsHYvVvZNpnBlF5TkJGXgU///hS+7r54N+5d078cHQ0sW1Zpadq8ojzM3T8XH+7+EFn5WRjTYQzejX3XIQPyTMWh0pvn5fEDX19ITp7UeR4C/LBq1oxjLPQjzps2NRQEeWrUyHaHOCoTlePHeZuMi4tuGMzZmW0p2dl8fV27ctCxPFxmbPL01A2hyfYRU+0rFX0uKOAekb4IyJO3t+Fnd/eKf4fCQv6tEhPZ1nX8OLB1K6eB2rCBr9PDAxg8mHseQ4bwMe0e2W4xbx7wzDPclZw6lVXfWP4pG4SIMG7tOCw6sgifDvgUL3d72bQvLlzIgXvnz+uEswSNVoMlx5Zg6vapSLmVgrtb3I1ZfWfZTbZZJYekFLG0kw15SRUUEHXpQuThQfTRR0STJxM98gjR668T3XcfO1v8+qvSrbQsV68S+foSde9OlJ5OFBFBFBVFVFSk20ejIfr8cyI3NyI/P6Lly5Vrb0354gv+HZ99lucrV+q25ecTbdxI9NRTRAEBvN3Jib3T5s0jSklRrt01RqPhH/CVV/jz8uV8gUeOKNsuMynWFNNDKx4iTAd9f/h707508CBf62+/la7SarW0IX4Dtfu6HWE6qPOCzrT9wnbrNNqKQEEvKYcXjEmT+C6sWFF+W0EBUbs188wrAAAgAElEQVR2RIGB/GCtC2i1RPffT+TqSnT6NK9bvZrvwbx55fc/fZqoc2fePmIEu/3aExkZRA0aEPXtS1RcTNSsGVG/fsb31WiI9u3jl4WWLanU9bhzZ6IPPiA6eZLvn10RFcU/OBELBWCX6p9flE8Dlgwg1QwV/XbyNxO+kM/K/+abRER04PIBivsxjjAdFPFFBC0/sZy0dvdjMkIwFOL336n0zbMiDh0iUquJxoyptWZZlRUr+JpnzdKt02qJevfmB2tGRvnvFBURvfce//8FBhJt2FBrza0xb7zB13voEH9+913+fO5c1d89fZroww+JunbViUeLFtwL3bOHBcjmueceovbtefn2bb6IGTOUbVM1ySnIoe7fdyfnd53pz/N/Vv2FO+6g23160YiVIwjTQf4f+9OXf39JBcUF1m+sFVFSMBzKhqFPUhLbxcLD2TOqMpfUd95h78R163h82165eZMN9MHBwN9/GwbB/vsvp4OcNAmYM8f49w8fBh57jO0+EyawQ0Atlk42m5QUdmx48EFg6VJed+UKX/+kSZwF21SuXAHWrmW7x7ZtbGOpV49tNrLtRJLKL5u7zRiV/YtWtE2t5mDnwZte5FQEt27xSUJDuRrdsmUmX7stkZmfidgfY3Eu/Ry2jN6C7s26G90v9XYqEu6PRfjeUwh70x2Tur2CyT0mo75rfaP72xMOGYehpGAUFgK9erEB/PDhcjaxchQUsNdUejq7n5oT0GdLjBrFCR4PHeLyCGV54gl+sJ46xV5SxsjPZ7vpnDn87PnpJ76XliItjX8X2SPt1i12cqnO9PvvwLFjwCuvcM6yevU4vuaJJ4Bdu9gBoDqxK1lZwMaNnAhWo9E9tHX9kOp/rkg4zF2/ahW/3Czt8iXw4ovAjRvsuTBwoK6Cn51yPec6ev3QCzdu38DOsTtxR8Adpdtyi3Lx2b7P8NGej/D4Xzn4YiPh+plDaBxp/VrbtYUwetcyL71E5YyfVXHwoH0PTa1dy9c8bVrF+1y5wsZ/eci7MnbtIgoPJ5IkoldfJcozI4tDcTFRQgLR+vVEn3zC6VR69iTy99d/hLKdJSCAqFEj3ubnR+TjQ1S/PpGnJ1G9emzTldOyqFSG3zc2DRrERm6A6OefTW+zvfHII0RBQUTatX/wxe7bxxuef55vnp2O38skZSRR0KdB1Gh2I4q/GU9FmiL67tB3FDgnkDAdNOzXYZS0bhlf+x9/KN1ciwJhw6g9ZAPv88+b/90pU/i769ZZvl3WJCODqEkTorZt2ZBfGe+9x9e4fXvVx83OZu8igKhNG52dQOb2baJ//yX65ReiqVOJhg/n4XQ3N8OHuL8/C8b48Swg69ezoFTXRqDVEg0cyOJy9SpRbi5RTg4b9QG2Y4SHs92mrvLNN3ytyRtP8sKyZbzhq6/48+XLyjbQApxOPU3+H/tT8GfB1Pqr1oTpoG4Lu9Hu5N28Q3Y2v9HYqc2mIoRg1BKJiUTe3uz1kp9v/vfz8/mh26SJceOwrfLkk/z2feBA1fvm5rInUXS06Q/sjRv5njg5EY0ezW/xISGGoiBJ/JAeMoS9PL/7jpMspqbW6NKMsmULn3POHMP1Wi3Ro49yW558kvc5dcry57cFTp3i6/vh61xeeO893iDfnP/9T9kGWohDVw6R94feFDk3klafWl3e8ykykmjYMGUaZyWEYNQC+flEMTEsGImJ1T/OgQM8NDV2rOXaZk02b+Zf+fXXTf/OspKe/KJFpn8nLY1o1CgeLurQgWjkSH6xW76c6Ngx84asaoJGw2IXEmL8nDk53Btq0IAF7qWXaqddtY1Wy0N5o0cTq7n8B5uSwj/u118r2j5Lciv/FhVpioxvHDmSKDi4dhtkZYRg1ALPP89Xu3p1zY/11lt8rPXra34sa3LrFj84IyO552AqWi27kgYEcK/enli6lH+bpUsr3uf0abbV+PvzC4Q598aeePhh7i1qe/YkuusuXqnV8sW/+KKyjastPv6Y/yBu3lS6JRZDScGoYzlJjbNqFeeHeuklzplUU6ZO5fxTEyZwniNb5c03OcXJ999z6gxTkSTgs8+Aa9eAjz6yXvssTX4+p5SPjgZGjqx4v1atOHPEzZvs8bRiRe21sTaJjWXX4pxGEbr8N5LEvsY2WK7VKnQs8Y6yYMlWR6bOC0ZCArtRduliuYefqyvw44/8QJ00yTLHtDS7dnEKoRdeYLd7c+nWDRgxgt1n7aV2+VdfcVtnz646PfuIEbqyEDNnWr9tShAby/Nz2nBOHJafzyvsKGttjYmO5rmFSrY6OnVaMPLzgeHD+eGxfLnpdTFMISYGeO01TgO+caPljmsJcnM5025YGPDBB9U/zqxZPH/zTcu0y5qkp/O1DhrEmWdN4ZNPOIgvPp6DMusarVtz6MWBm+Hse5CUxBtateI0xXLO97qMnx/XUhA9DItQpwXj1Vf5xeKnnyqu7VATpk3jyOnx421raGrqVE7SuXAhZ2GtLiEh3IP6+Wfgn38s1z5rMHMmDy+Z04t0deXobYArJNa156ckAb17A3+e10tzDnAPg4j/SByB6GjRw7AQdVYwVqzgIZlJk7jmgTWQh6auXuVoYlvg77/Z/vDUU1wIqqa88QanBH/5ZfNK5tYmSUlsoxo71ngEe2XccQffp7Q0LtBWV9ixg19ievcG9lwL55UJCTyPjOS5I9kxzp3jPPaCGlEnBeP8eWDcOODOO3XDKtaic2cemlq0CNi0ybrnqoqCArbXNGliXp6kyvDy4qGefft4WM8WefttHnZ814z6OvrI31u8mIcY7Z39+4G4OH5hio0FbqARilw9dD2Mli157kiCQcQVwwQ1os4JRn4+l1Z1cuIHXNkyrNZg+nTd0JSxKne1xXvvcQ6mBQu4GJKlGDMG6NABeP112xu2OXyY8+i9/DKXoa0O3buz15uXF9cZsvfnilz6+NQp/rv095dwrV64TjA8PPhmCcO3wEzqnGC8/DKXY128mA2atYGrK7+ZXrmi3NDUv/9yb+qxx7hynCVRq3mY6+JFntsKRMDkyZxY8PXXq38cSQImTuQRCy8v4KGHlBX+mrBvH/d0VSruQKhUwF13Aafzw3VDUgAbvh2lhxEYyPV4heG7xtQpwfj1V07pPHkycM89tXvuLl34vN9/X/tDU0VFXB/d3996D/TYWC5f+uGH7E5sC/z5J6canzqVS7bWhEcf5Wy2XbsCFy6wl5mt2mwqY/p0/jsYM4b1gIh/uxN54dAmJOouKjJSt0NdR5KE4dtC1BnBiI/nIaHu3WvmSloTpk9nV8baHpr66CMeRvnmG/YitBazZ7Od5O23rXcOU9FoWKDDwy1jrPb25mC/7dv5d1y1Cvjii5oftzbZuxfYvJltah07Ajk57JDRuzeQgAio8vO4WDvAgnHrlu5zXadjRx6jk2NRBNWiTghGXh7bLVxduZdRG3YLY7i56YamXn21ds558iTbLoYPt0wUe2U0bw48/zwb+I8cse65qmLxYq5N8uGHlouveeop4PZtrncybBgL0t69ljl2bTBtGtCoEdth9B2h2rYFUj0d3FMqOhooLuY/GkG1sVvB0GjYG2rdOi6Kc+wYsGQJ0KyZsu3q2pXFYuFCHjKxJhoNe0V5ebFbaW3wzjvci5k0SbnRjNxcbkeXLvyiYCliYvhFdP58FsWQEBbi1FTLncNa7N4NbN3KvQsPD0M9UKmARneWCIZ+LAbgOIZvkSLEIti8YOTl8dvsL7/wG9Tw4exr7+EBtGgB3HsvDx9Mn255Y291mTGDbYrjxll3aOrzz9mFcu5cfrOsDXx8+Pq2b9cFvdU2X3zBmS5mz668tKm5SBL3Mo4f5+foypWcb2rUKBZnW2baNLbrTpzIn4OCOH+Y3IFoNSgUWkjI/LdEMJo1M9yhrhMWxuOOwo5RM5TKelg2W21qKtdHWLCAaNIkosGDiUJDuXaBXFNBpSKKiOC69pMnE33/PdHevUTp6TVI/Wgl9u3j9o4fb53jx8dzIaL77qv94mlFRUStWxM1b151QSZLc+MGkZcXX7c1uHWLj//YY/x54UKqslKh0uzcyW389FPD9e3bE919Ny8fOUKUjGZ0vsdo3Q533KHbwRGIjSXq0kXpVtQYOGJ686CgTjR+PFGvXuVLc7q7c02FESOUqalgKSZP5uvZtMmyx9Vo+L55eytXOG3DBuMPKWvzwgssxNYsfPT00yzGaWksxmPH8ouLpX9HSxEXx6noy6Zpf/hhfsEi4r+Zv5x6U3zjHrodhg/X7eAITJrEP2xRBbUz7ASHFAygEzVowKU5x43j0pwbNnBxI43GMjdWafLyiFq14rvcsCFRp05cL/vFF7ka3IoVRH//zbW0zblmucqmOQWOrIFcBrW2Sg2cO8dFjyZMsO55jhzh+/vZZ/z59m2idu246NLFi9Y9t7ns2MFt/fzz8tvefpvFVa4uuSX4CbqmDtTt8M47hjvUdZYs4Zt1/LjSLakRSgqGxOevfTp0iKEjRw4qcu7aJDkZWLqU6xJcvMhTcjK7POrj7MzDysHB5Sd5vacn501q25ZTlm/aZNkxfHM5eZJzMU2cWDtG90ceYSeH8+c5Fsua3Hkn259OneJ7HB/PRvE2bYCdOy3jmXXzJnDgAE8HD7K9LjiYje36U1BQxZ5/sbHctoSE8jVPli5lh5CTJznie8+QD9Bjw9u4HH8bTVvU46ySo0ax51CbNjW/IFvn1Cm+zsWL+cbYKZIkHSKiGCXO7aTESQFO3eEIhIQAU6YYriPih5EsIGWn7dvZqKvVGn7P15fvmyRx+g8lxQLQFZH65ht25Wzd2nrn+ucfTig5dar1xQLg2I7HH+e6Ir17c/qlRYvYK+u119jhwBxu3QIOHTIUCDnbuCSx05K3N7BhQ/nASJWK84OFhHDWZVlIMjNZvGbPNl4gS99TKioKaNozHNgA/Lv6Apq+3sZwB0cQjMhIvlGHD9u1YCiJYj2MmJgYOniw7vcwqktxMcdzlBWTlBT+Wx8+XOkWMqmp7K3Wowewfr11zkEl0cpnznDvwsvLOufRJzeXH9J3380v4jIvv8xisWJFxS69smefLAwHDhgGVYeGctJKeerY0TD3V34+/87JycanlJTyXlsNGxr2SsLC2IMwNJRjVd54A9Ds2w91966Y238tnt98L+dCqV+fc8PbQ9ETS9CtGwds7dihdEuqjUP2MASV4+SkG5KyZRo25MjvyZM5ynjAAMuf448/+E3/669rRywAThMyZgz3nr74gq8T4Kj6f/7h+Jf27TnS/MQJnTAcOMCfi4t5/4AAFoX//IeHtGJidMeqCDc3FuEWLYxv12hYsP7zH84d1rKlTkxOnOBeSl4e90ACAnSes+oWHItx69+S4D0vL1ZFR3GtBTiAb9ky7r5XVZZRUA6TehiSJA0C8AUANYCFRFQuabgkScMBTAdAAI4S0X8qO6boYdQdCgp4RMPNjd+sLTncWFwMtGvHb+fHj9duFL885P3xxyyIMikp3CvQarknImeb8PVlQejcWTdv2tTyQ4dEQK9ePKR1/jzf97LbIyI4sPHaNaCwsCRinQgF7t6YXzAWD176Ek2bgouB5OZyIRVHYOFCzt1z7hynLrBDlOxhVCmxkiSpAcwDMBhAFICRkiRFldmnBYA3AfQgojYAXrJCWwU2iqsrP1RPnuT/R0uyaBEPRc2aVfspX6Ki+ME8f76hPalZM37D79iRDf4//8zPn7Q07mV98AGnaQkKso6daetWYM8e4K23yosFwOds04YFT84xKG/QhEYgHInYubNknZy1VqGh6VpHTnUuIr6rhSnvgl0AnCeiRACQJOlXAEMBnNLbZzyAeUSUAQBEdMPSDRXYNvffz8bhiRP5Qebvz1PDhoZzY8teXsYfrDk5HMHcowcwdGjtXxPAxu9Rozgrbr9+uvVxcTzVNkR8T5o144y6FREVxeI1ejTXO795k++3e1Q4Wpw7hU938pAWIiN57Co1tfbSBShJ27bcBT582LJ5ZRwEUwSjKYAUvc+XAHQts09LAJAkaQ942Go6EZVL8i1J0gQAEwAg2NYH5wVmIUmcvuX77zkB6s2b/AxKTmbvoNRUHhoxhrOzcSG5dImHVFavVs4j7MEHgRde4LT5+oKhFJs3c82Lb77hnl1FREXx/ZbTvp89y/dVighHGNZj53YtAJWhp5QjCIarK4uG6GFUC0uNNjsBaAEgFkAQgF2SJLUjokz9nYhoAYAFANswLHRugY0QGFhx6nMi7jGkprKYyIKiP5eXjxzheUYGv91361a716GPqyu71372GacKrw2X3oqQexfBwWx0r4yokkFjeSgtPp57aoiIgIu2ANnnruLq1aYI1E9C2KuX1dpuU0RHc0APkfK+6XaGKYJxGYB+DtigknX6XALwDxEVAbggSVI8WEAOWKSVArtHknjoycuLPYtMobiYq/0pzYQJwJw5bE8pG1NTm2zaxB5a8+dXHTjYqhXP09K4B1dqxyi5+RFIwM6dTTHi4WBWRUfylOrYUVeHoGlTpVtjV5jiV3YAQAtJksIkSXIBMAJA2Tylv4N7F5AkyR88RJVowXYKHBA5SFFpWrQA+vblYEmlstbKvYuQEGDs2Kr39/LinsiZM+wxVVYw2rglciiCWs1+uY4kGKLGd7WpUjCIqBjAcwD+BHAawAoiOilJ0ruSJN1XstufANIkSToFYDuAyUSUZq1GCwS1zVNPceCktWucVMSGDRzj8fbbpqcliYoy4ikVEgKoVLgrSM9TymAHB+COO/hNRAiG2ZgUuUJEG4ioJRFFENEHJeumEtHakmUioklEFEVE7YjoV2s2WiCobYYO5XoT335b++cm4novoaEcTGgqUVHA6dPcgTh/viSY0NkZCA7GHZ4JOHOmJA1JZCQXVqrIK6Gu4enJN0UYvs1GhDoKBCbg4sKG5vXrOXCvNlm3jiPJ337bvFiU1q054rtBA6CoSJe7CuHhCC7mEeOdO8GCodHoyrc6Ah07ih5GNRCCIRCYyPjx/LZv6eDEypB7F+HhnAbEHGRPKTkDRumoU0QEPG4kwsurRDBkC7kjDUtFR7Py37ypdEvsCiEYAoGJhIUBAweyYMi5oqzN2rX8Imxu7wLQZQ++fZvn+oZv6cYN9L8zmw3f+rEYjoKo8V0thGAIBGbw9NPsjblunfXPJfcuIiKql43b15fjRpKTAT+/8p5SQ6Iu4PRp4EZ+fcMshY6ASBFSLYRgCARmMGQIu+7Pn2/9c61Zw0GM77xT/YSORj2lIiIAAD2blLFjOJJg+Pmxx5iwY5iFEAyBwAycnIBx49i99sIF651Hq+XeRYsWHO1eXWTBMAi10Ave8/SEbljqzJkattrOEIZvsxGCIRCYybhx7Mb/3XfWO8fvvwNHj9asdwGwYOTk8IjTtWtc+Q++voCPD9TJiejRQ8/wLWcpdBSioznN8K1bSrfEbhCCIRCYSVAQcM89HPm9ZAmXk7Akcu+iZUtg5MiaHUv2lJIN5vHxJRsiIoDERMTGclr6rAAHNnwfPapsO+wIIRgCQTWYMQPw8WFX14AAdrndu9cyZSVWr+ZiUVOn1rwYlSwYckyewbBUQgJ69+aP+9IdUDBqw/C9ciVX36oj9UaEYAgE1aBDBx7N2LmTU6D/8gtng23VimtoXy6bntNEtFoWo1atgBEjat5OOVX8zZscj2EgGElJiInWoF49YMOpUI5OdCQ7RmAgh+9by46xZw8XHZkzB9i92zrnqGWEYAgE1USSgLvu4sSn165xNtvGjbmAVHAwMHgwV+aTS7iawsqVXJd76lTLZeqNimKhCA0t4ylVVATnG5fRsyewfZeaS5Y6Ug9DknhYyho9jMuX+U0iJIRD7T/5xPLnUAAhGAKBBfD05LoZu3Zxz+Ott9g28MgjQJMmwHPPcSGpykYmNBruXbRuDQwfbrm2VeYpJQ9LnTgBFIS1cizBAHhY6uRJ81S9KvLzuQTl7dvsG/3ssxyBWQfurRAMgcDCNG8OvPceu91u3gwMGsSVCGNiOFHqp58CN4wUMf7tN36wT5tm2TogUVFcjKpZMzZ6a7XQCUaJ4RsAklwjOZ9UUZHlTm7rdOzISn38uGWOR8TRnQcOAEuX8s1/9lke7vvsM8ucQ0GEYAgEVkKtBvr3B37+mav1ffMN4O4OvPIKB/8NG8YvoEVFut5FVBTw0EOWbYecIqRePU5GeOkSWD2cnIDERMTEcLsO3IrknCfWDDCxNSxt+J47F/jpJ3ZzkwvRN2rEaYZ/+sn4m4IdIQRDIKgFfHz4xfOff3gE5OWXeXnYMHbTffBBtjdbuncBlC/XevYsWCxCQoCEBLi4sMF+0wW9cq2OQlgYFz63hOF7+3Zg0iT+Ud95x3DbpEk8VPX11zU/j4IIwRAIapmoKODjjzlZ6h9/AD17coGkO+6wfO8CYLdfHx8gK4s/G9gxEjk9SGwssD7BAV1rJYl7GTXtYSQlAQ8/zBHzixfrUgTLREYC990HzJtn+cCdWkQIhkCgEE5OHAC4ahVw/TobzMs+ZyyBJLFIJSezcd7AU6pEMHr3BjLhi3zvRo4lGADbMY4dq34K4txc7lVoNByi7+VlfL9XX2X/5sWLq99WhRGCIRDYAL6+QP361ju+fhLC0mjv8HAgLQ3IykLnzmzHuOzhYEkIAe5h5OdXbyiOiCtrHTvGwTgtWlS8b8+eQJcu7PWgVHH4GiIEQyBwAKKigNRUNluUc61NTISrK9CtG3A030GTEALVs2PMng0sX87RmoMGVb6vJHEv49w5Hou0Q4RgCAQOgGz4rl8fuHiRvaXkNOf6dox96ZE8bJKerkg7FSEykrtX5grGpk3AG29wsM1rr5n2nfvvZ0P7nDnmt9MGEIIhEDgA+uVaifglF2FhvLKklndsLHAaDliuVa1mjwNzDN/nznFmyPbtOchGkkz7npMT8NJLnDZk377qtVdBhGAIBA5AUBAbvGUHnbNnwe6kDRqU9jC6dAGSXBzQUwrQpQiRfY8rIzubjdxqNRu5PTzMO9cTT7Dbmh2mCxGCIRA4ALKn1LVr/NmYp5SrKxDQLQxFkrPjCUZ0NAtByb2oEK2WUxSfPcuh+aGh5p/L0xOYOJHTEpf07uwFIRgCgYMgJyEMCiqf5lymV5wTzlMECo8Lw7dR3nuPexWffgrExVX/fM8/z8NTdpYuRAiGQOAgREVxipLw8DKCkZxcGoMQGwucQSsUHHOwHkabNvwAr8yOsWYNp/wYM4Yf+DUhMBB49FFOcZyWVrNj1SJCMAQCB0E2fPv7s2AQgYekNBoOOwfQtStwXh0J98vnqx/IZo+4ugJt21bcwzh1ih/wnTsD335rupG7Ml55hd3Vvvmm5seqJYRgCAQOgpyE0MWFy1hfvw6DNOcA4OYGaJpHwklbxOkurIBWy9lzba4IXXQ0C0bZhmVkcCJBDw+2O7i5WeZ8bdpw0ZS5cy2bXt2KCMEQCByEkBAONzAo16oXvCfTsAd7SuUcsvyw1KVL/JLu58cxIW3bcnqUZ5/lGLjffgP27+ekrrUuKB07cgyKfrlEjYar5iUncw6XoCDLnvPVV/lily2z7HGtRA0rBgsEAntBrebSr/KQ+dmzQO+eTbnLoScYkfdFAouAi5vPIOqRIRY7//79/KJ++zanck9P5+dwUhKHJWRmGu7v7s4iFxpqfB4QYOHcW/qGb1kYpkzhAL358zmlr6WJi+OezSefcAUuayQTsyBCMAQCByIqCvjrLx5ViY8Hq0hoqIGnVKcBDZAKf9w6aLkexvLlwNixbOvdupVHY8qSlcUCIotIUpJu+eBBfvnXx8VFJyDvvAP06lXDRrZvz7aJf//lzLLLlwMffQQ89RQwYUIND14BcrqQUaOAjRuBIZYTaGsgBEMgcCCionj0Q3axBWCQ5hzgN/tz9SPhklhzwSDi3sSMGZx7b/VqoGFD4/t6e/Mzu31749tzcgwFRZ5v2gR8/rkFBMPTk9OEHD4MHDnCb/w9egBfflnDA1fBww9zipE5c4RgCAQC20H2lGrcuEzw3t9/G+xXFB6JpkfWIyuLH+TVIS+PexUrVvD822/ZGam6eHpyz6Rs7+Txx7lktlZrgRGd6Ghg2zaO5PbzA1au5K6MNXF25nQhr7zCXamYGOuerwbY9oCZQCCwKLJgeHhwp6KwENzDyMw0SDjo3TkSAbiOf/7MNH6gKrhyhWts/PYbF4tatKhmYlEZcXHcdIuU5e7Ykd3Hrl0D/vtfNpTUBuPGsReAjacLEYIhEDgQ4eH8wqzVsgNQYiKMeko168eeUmfXmj8sdfgw56U6dYqDoidPtkzYQkXIAdfbtlngYL16cWMXLGB3rtqifn22lfz2m9XcmS2BEAyBwIFwcuJheoNyrWXSnAOA6x2ctTZtr3mCsWoV2yrUamDvXrYdW5tmzYDmzbmkdo3p2pV7W489ZoGDmckLL7BYffFF7Z/bREwSDEmSBkmSdFaSpPOSJL1RyX4PSpJEkiTZ7iCcQODgREXpQg3OnkW5NOcAgPBwaFROcE06i1u3qj4mEfDBB1yT/I472IW2IuO1NYiL4xK3FilkZ83Sh5URFASMGAF89x0HC9ogVQqGJElqAPMADAYQBWCkJElRRvbzAvAigH8s3UiBQGA55PreDRuWCIanJ9CokWGmVmdn5DcJRws6iz17Kj9efj4wejTw9tvsHbp9OxvVa5O4OO41mVPSwiZ55RUOVFmwQOmWGMWUHkYXAOeJKJGICgH8CmCokf3eA/ARAPuIcRcIHJSoKO4RGGSt1UtzLuPaPhKtcQY7d1Z8rOvX+WG9bBnw/vvAkiWWy5xhDrGxPLfIsJSSdOgA9OvHw1JySL4NYYpgNAWQovf5Usm6UiRJ6gigGRGtr+xAkiRNkCTpoCRJB1NTU81urEAgqDmyp5S3d8VpzgHAqbR4xK8AACAASURBVE0rNJfOY9d24+M8x46xcfvoUfY+nTLFusbtyggM5Ch2uxcMgAP5rl4FfvlF6ZaUo8ZGb0mSVAA+BfBKVfsS0QIiiiGimIYVRe8IBAKr0rw5G6XVar3y3eHhnLFW/602MhKuVIDUg8nIzjY8xtq1QPfubDPYvRt48MFavQSj9OnDdoyiIqVbUkMGDOAkW3Pm2FyGRlME4zKAZnqfg0rWyXgBaAtghyRJSQDuBLBWGL4FAtvExQVo0YKHyoGSFCEREexre/GibsdIdq1trj2LvXt5FREnCRw2jLPf7t+vS8GkNHFxfE0HDyrdkhoipws5cQLYvFnp1hhgimAcANBCkqQwSZJcAIwAsFbeSERZRORPRKFEFArgbwD3EZG9/2wCQZ0lKqokvTnKZK3VH5YqEYw2qjPYsQMoKOBy1K+9xtksdu4EmjSp1WZXSp2xYwDAyJF8c+fMUbolBlQpGERUDOA5AH8COA1gBRGdlCTpXUmSasHLWiAQWJqoKO5MqNUVpzmHvz/g64seDc9i/Xqgf3/gxx+BadN4eL1ePSVaXjH+/kC7dnVEMFxcOC5j61bOa2UjmGTDIKINRNSSiCKI6IOSdVOJaK2RfWNF70IgsG2iotj+UOopFRjI7k36giFJQKtWuMP1LI4f5+GnX37hKqW2moU7Lo5TpRcUKN0SC/DUU+zybEPpQmz0ZxcIBNakbLlWqFQcwFfGUwqRkWiWdxZxcTwENWJErTfVLOLiOOnh/v1Kt8QC+Phwjqlff+XKUzaAEAyBwAFp2ZI1wsUFOH++JEK6TJpzAEBkJJxTr2Lb77fQtasiTTWL3r25Y2SRvFK2wIsvsqeBtVOsm4gQDIHAAXF3Z30oLOThm+Rk6IL39F05SwzfuoAN28bXlzOU1wk7BsDVoR5+mCv+mZKjxcoIwRAIHJSoKF1G81LDd3a2YWk7OxMMgIel9u3joak6wSuvsFgsXKh0S4RgCASOSlSUbmi8Qk+piAg9Vyr7IC6Oe0779indEgsRE8NjbZ9/rnhUohAMgcBBiYri54+XV8VpzuHqysZwOxKMXr1Y4+rMsBTAgXwpKVwvQ0GEYAgEDorsKRUQUKIHoaG8woinlD0JRv36QKdOdcjwDQB3383JsmbPVrQZQjAEAgelFddIgodHSXqQevU4HsOIpxTi4zl1iJ3Qpw+71ubkKN0SC6FSsS1D4SA+J0XPLhAIFMPDgzsVRFxQKScH8DSS5hytWnHRC1dXLtmnVvO8JssNGnClpQ4deLJwAY24OGDWLA7iGzjQoodWjkcf5ZTAN24o1gQhGAKBA9O6NXD6NC/HxwMdw8PLj+UMHw6kpnJmP40GKC7WTZV9NratoICPc+aMYfrugACdeMiTnFa3GvToATg7sx2jzgiGmxsXSe/eXbEmCMEQCByYqCidPpw9WyIYS5Zwj0KuhOTtDbz1luVPnp7OxTSOHNFNW7eyuAA8RNa+PQdWyCLStq1JSaw8PLhWR50yfANAt26Knl4IhkDgwERF6fIunT0LoHkEj1ElJ+tiMKyFnx+PHcXF6dYVFACnThmKyLJlwDff8HaVittVtjfSqFG5w8fFATNnculWb2/rXoqjIARDIHBgZE+pRo1KBGOAXppzawuGMVxduUcRHa1bRwQkJRmKyJ49hkNaX38NTJxocKi4OC4b+9dfwD331E7z6zrCS0ogcGBat+Z5ablWY8F7SiNJHAty//3AjBnAmjXcA0pL4/G0jh25BnaZ6nTdurH+1LlhKQURgiEQODDe3kDTpmxbjo8HqFFjthHYkmBUhDykNX48q93x4wab3d1ZNIRgWA4hGAKBgxMVBeTmsvPS5SsS9zLKBu/ZMg88wLaNFSvKbYqL4xEsOWeWoGYIwRAIHJyoKJ1rf+mwlD30MGQaNeJIveXLyw1LxcXxql27FGpbHUMIhkDg4ERFsRctoJdTqmyac1tn+HAu7FEmErpLFx6aEsNSlkEIhkDg4MieUq6uJSlCwsN5jOr6dUXbZRYPPMCGmDLDUq6uQM+edSyvlIIIwRAIHBzZU6pBAxv2lKqKBg2Afv0qHJY6cYKD1QU1QwiGQODgNGjAqZxcXStJc24PDB8OXLgAHDpksFqOC9yxo/abVNcQgiEQCEprYyQlAfmNQzj2wZ48pQCO03B25l6GHp06AZ6ewo5hCYRgCAQCtG7NrqdEwPlLbhycYW89DF9foH9/tmPoDUs5O3NRJSEYNUcIhkAgKI3FAMp4Stkbw4cDFy9yMQw94uI4Qe7Vqwq1q44gBEMgEJR6SgF6hm97G5ICgKFDAReXcsNSffrwXPQyaoYQDIFAUCoY9evrCcbVq7puh73g48MFMH77zaBCYIcOvEkIRs0QgiEQCNCoEadm8vQs4ymVlKRks6rH8OHApUvA33+XrlKrgbvuEoJRU4RgCAQCSBL3MohYMChML825vXHffewjXGZYKi6OLyclRaF21QGEYAgEAgAsGFlZQGYmcLO+HQbvydSvDwweXG5YSo7HEL2M6iMEQyAQADD0lDpz0x/w8rJPwQB4WOrqVS60VEK7dhykKASj+gjBEAgEAAw9peLP2WGac33uvZdrkusNS6lUQGws55Wyp7yKtoQQDIFAAEAnGE5OdprmXB9PT2DIEGDlSkCjKV0dF8dhGhcuKNg2O0YIhkAgAAA0acLD/6WutRER/GTVswPYFY88whl3//qrdJWwY9QMIRgCgQCAzlNKrdbrYeTn22949N13c7lZvWGp1q050aIQjOphkmBIkjRIkqSzkiSdlyTpDSPbJ0mSdEqSpGOSJP1PkqQQyzdVIBBYG9nwnZAAFAfbsacUAHh4APfcA6xaBRQXA2BRjI1lwRB2DPOpUjAkSVIDmAdgMIAoACMlSYoqs9u/AGKIqD2AlQA+tnRDBQKB9Wndmmt7FxcDKS52muZcn0ce4UIYO3eWroqLA65cAc6dU7BddoopPYwuAM4TUSIRFQL4FcBQ/R2IaDsRyTkE/gYQZNlmCgSC2kDfU+pkdjC7FtmrpxTA8RgeHgbDUnJeKVGFz3xMEYymAPRjIy+VrKuIJwFsNLZBkqQJkiQdlCTpYKoofyUQ2Bz6gnEm0QVo1sy+exju7hz5vXo1F/wA0Lw5Z28XdgzzsajRW5KkRwHEAJhtbDsRLSCiGCKKadiwoSVPLRAILEBwMNuJ3dzsPM25Po88AqSllXYpJImHpXbsEHYMczFFMC4DaKb3OahknQGSJPUDMAXAfURUYJnmCQSC2kSlYjtGqWDYc/CezMCBHLW+YkXpqrg44MYN4NQpBdtlh5giGAcAtJAkKUySJBcAIwCs1d9BkqRoAPPBYnHD8s0UCAS1hVyutVQwbtwAcnKUblb1cXPjOhn//S9QWAhAxGNUlyoFg4iKATwH4E8ApwGsIKKTkiS9K0nSfSW7zQbgCeA3SZKOSJK0toLDCQQCGycqij2lbtwAbgeUeErZe2j0I48AGRnA1q0AgLAwICRECIa5OJmyExFtALChzLqpesv9LNwugUCgEPqG7yRVONoAPCzVrp1STao5/fsD3t48LHX33QDYW2rNGg5kV4kQZpMQt0kgEBigLxgncu08eE/G1RUYNgz4/XeggE2scXFAejpw7JjCbbMjhGAIBAIDwsL4+SpJwPHLflzb1N4FA+BhqawsYPNmAMKOUR2EYAgEAgPUaqBVKw5hqDOeUgDQty/g61vqLRUUxDEZQjBMRwiGQCAoh365VrtOc66PiwvwwANsuMjPB8C9jF27DDKgCypBCIZAIChHVBSQlwfExwMUHgEkJdnGU5WIh5XOnOGuwc8/A3PnApfLhYYZZ/hwIDsb2LQJAAtGVhbw779WbHMdwiQvKYFA4Fi0bs3zggIgzTsc/oWFQMuWgL8/4OdX9eTry5Ozs2kn1Go5Gvvq1aqnvLzy3//uO+DAATa+VEZcHNdpXbECGDas1I6xbRsQE2P6/XFUhGAIBIJy6HtKHWv+APq8dJqLEaWn84P93DlezsysPL+Gl5dxQQEMReDatdIU5AZ4ewOBgUBAANC1Ky+XnU6eBB5+GHj3XeCDDyq/MGdn4MEHgWXLgLw8BAS4o3Vr7qy89pr598nREIIhEAjK0bw5l2otLgZOXPNHn88+M76jRsNjOunpPGVk6JaNTceP85xI98Bv08a4EAQEcGKrqmjdGnj8cWDWLI7o7tKl8v2HDwcWLAA2bAAefBBxccBPP3F0u6kdIkdFCIZAICiHszOPQMXHlxi+K0KtNuw1KMVnn3EU95gxwOHD7OJVEb17Aw0b8rBUiWB8/TVw8CDQrVvtNdkeEUZvgUBgFLlca3y80i0xAW9vYNEiNoa/807l+zo5AQ89BKxbB9y+jdhYXi3ca6tGCIZAIDBKVBTn6jtzRumWVMzFi2y6OHAAQL9+wMSJwKefArt3V/7F4cO5Fu369fD356wnQjCqRgiGQCAwihyLcekSJyO0FYiAvXv5mR8eDkybBgwaVBIq8vHHHKo+dmzlje7Vi20kJUF8ffqwxhSIwgyVIgRDIBAYRd9TyhbqXxcWcthF165Ajx7Ali3ApEncMyAC7r0XuKX1BH74gdXjjTcqPphazcNS69cD2dmIi+NYvn/+qb3rsUeEYAgEAqO0bKnL4lqp4dvK3LwJzJzJHYdRo9gpa948ICWFOxSxscDKlWxrGTkS0PS4C3jxReCrryov3D18OKvEunW46y7OnSWGpSpHCIZAIDCKqytXaAWUEYwTJ4Dx47ms+JQpQNu27Al7+jTwzDOAp6du3z59OOB7w4aSeIqZM1nxHn8cuHXL+Al69ACaNAFWrICvLxAdLQSjKoRgCASCCmnThp2KakswtFoeJerfnw3RS5cCjz3G4vHnn8DgwRXXrnj6aeC559jm/f3P7hxccekS8Morxr+gUnHA38aNwK1biIsD9u0zHkguYIRgCASCComK4ti806ete56cHB5matUKuOcerrU9cyYPO82fz8JlCp99xmIzcSLwV9GdwOTJwMKFLArGGD6cLd1r16JPH7aT7NtnueuqawjBEAgEFaKftbayDCDVJTmZn+lBQdw78PVlw3ZSEvDmm5y6yhycnIDly9ne8cADwIWxM1htxo3jKPSy3Hknj3mtWIFevdgWXpnZw9ERkd4CgaBCZE+p3FzOuOHlBXh4VDzVq1f5dnn6+2/giy+A1avZ2PzQQ8BLL/Hzu6b4+gJ//MHeVPc97Iq/v/4JHn26siF88WLDneVhqblz4aXJREyMj7BjVIJE1nhtMIGYmBg6ePCgIucWCASmkZvLxuU2bYCmTTm0wdhUVGT+sX19gQkTgGef5Zd8U5HTV8lpq27dAjp3BurXN9xv61aOzxg8GFjTYRpU77/LJVqHDjXc8Z9/WKl+/BFvnhmDOXOAVavY6O/kxGlSnJwMlyual12WJPPvS1VIknSIiBTJrSsEQyAQVEp4OPcufv214n2KiioWE2NTYCA/twsKdA9+U+dZWeWHxyIiuFchp2WXmTePh7remFSID7d1Ba5c4ey2+mNdRDyG1aYNdr2+Hr17W+7eqdUsuEFBPDVrplvWX1dW7CpDCIZAILBZ7rmHjc9Hj1a+n1bLD/Tr13XTtWvGl2/eZANzRcg5DX19y8/LrisoYFHIzWVRu/tuw2M98wzwzTfAf989jmHvdQLuv58NHfq89hpbzK9fx9lUP2RnswgWF5s/L7vu/+2debgcRdXGf+cmudlXAtEECJGwxQBBQEQWUZYAQVYJm2EHQdlFdjEgyOanRCBACCSAsgtJkF1AxQAKsoOICHwCH4tIgIQly73n++PU5NbtzNI9Uz0zCfU+Tz/T1V11+q06PV3LqTr10Ue2v9Prr9ukrXfeWbLC69u3eEXihwcMsB5LIyuMaMOIiIgoi1GjbFX1Aw+Urwjefbf4lhbdusGQIXYMHWrrHZZfvnSFMGiQtcqzDOdstJH1WHbYwRbz/fCHHeknTbJFfXuctTYvHTCR4VNOtT0xxo/vEDB+PFxwAcyYwRoHHlhbgVXAggW2BUihAvGP11+3DtBbby1ZqfTqZRVHIxF7GBEREWXx61/DhAmdr3XrBiusYO6YCpVBqfOBA/MZy0/i449tnd7NN9vajcsvhx497N7771ulMu+DRbw2bBO6v/Ev+zIPGWIRVG0TkNVWW7x9ayOxcKFVxMlK5fXX4ZZbYg8jIiKiSTF+vI2x9+3bURnUqxLIgt69baRp9GhzSPjSS3DbbcZ30CCzcXzta13Z49Pp3DZvPeR737MINs7T0ct4773s83kDo1s3G5IqNhmgkeUe12FERESURWsr7LijbYc9apR9fJutsihABE4/3XxLPfOMzZ564gm7t+aa5pz29pfXYvrIs2HmTOs+FTB+vE3Buu22xpBfChArjIiIiGUOu+0Gs2dbBbLppou9mLPNNmbbPvj5Y3htxU3gyCPNIg0wZowNSRUiRyyBOCQVERGxTGLMGNtYadddYY89zB/VxIlWRzz/fBe2mjKdF1vXpevBB5vXwsKw1DnnmBGkMFRV6E6lDSfv9esH665r1v6RI0s7w1oKEI3eERERyzTmzzffUtOmWeVxzTU2zLbNNrDOQ5cwqe0IuOIKcx/yr3/ZxhqffGKJVTumKxXOk+FK9z78sGP6WO/eVnmMGWMVyJgxZnQpWOdTIK7DiIiIiMgRqnDhhXD88eYFd+ZMm7q70YbtTHtzazZpfYyW556F4cPDP3zBAvOm+OST8NRTHb9z59r9Ll1sxWGhAllvPatUBg0qKi5WGBERERF1wN13w557Wg/jtttsttf4jf6XRz9Zm56bbUiXB+6rz5BRezu8+mrnCuTJJ20legHDh3fuiay3Hqy0EtLSEiuMiIiIiHrgH/+wUafXXoPLLjM3JbeNm8oUPYT2iy6m5YgfNI7cu+8uWYm89FLH8NagQcj778cKIyIiIqJemDPHDOH33WdecocNVUadMI6tuv2R1heeNuN0s+Djj22OsKtA5IorYoURERERUU8sWmQ2jUmTYOxYWL33m5xx62gWrbYWy5/zQxuaKhxdupQPl4vTrZuteuzTx4zeNQ55RV9SEREREXVG165mCB892hwUvjpiGP1WuZgz/znBNujIAyJWafTt21GJFM6LhBf16svC1j7Mb+3LZ93saCRSVRgisi0wCegCTFXVcxP3uwPXAOsD/wX2UNXXwlKNiIiICI+DD4Y11rAptxct2Id3ttmMXgs/5MM57cz7qJ2PP2rj47ntLJjfTgvtdKGNFto7Ha0tbfTu1U7vnu307tFO7x5t9OzeTo/Wdnq2zKd1wTxaF8yldcE8eiyYS/eFc+n53ly6vz2Pnovm0qvtTXq1zaV3+1x6tc+jDx8D9oHuCvRsaAl1oGKFISJdgEuArYE3gMdEZJaqvuBFOwiYo6ojRWRP4DxgjzwIR0RERITGZpvZIr8dd4Sp966cXUA7MM8dAdBCG735mL7MpQ/z6MvcxQfsVDF9XkjTw/gq8LKqvgIgIjdgjP0KYydgoju/BbhYREQbZSCJiIiIyIhVVoGHH4Zrr7XFfoVd87p2NXOEHy53PXkNbBZtYa+MtrbOv8WutbV1YdGifixa1G+Je7cf3bgySlNhDANe98JvABuViqOqi0TkQ2A54D0/kogcChzqgvNF5LlqSJfA4OTzmkTW501eM3MLLa+ZuYWW18zcQstrZm4AawSUlQl1NXqr6hRgCoCIPB7S0h9SXjNza3Z5zcwttLxm5hZaXjNzCy2vmbkV5IWSlRVp5ne9Cfhe2Vd014rGEZGuQH/M+B0RERERsYwgTYXxGLCaiIwQkVZgT2BWIs4sYD93/h3ggWi/iIiIiFi2UHFIytkkjgDuwabVXqWqz4vImcDjqjoLuBK4VkReBt7HKpVKmFID77zlNTO3ZpfXzNxCy2tmbqHlNTO30PKamVse8lKjYSu9IyIiIiKWLiy9O3lERERERNQVscKIiIiIiEiFpabCEGnWbecNzc6vWdHs5dbs/CIi6omlpsLIY9aViKwpIt28cNUfh9D8RGQdERkWUqaT28f9Vp3XkNxy0mvT8stLr052TboNza2Z8+rSNu1/rIjMXUSkm4j0cOGqvt21cmt6b7UiMh74OnAO8KmqfhRApgBrAr8EnhCRdlU9rZqPg4jsDmwSip/jNhI4BvjM6fUooK3Wj5dzIrm3iEwG/gHMqUJGEG4ish+wBTAZeENV38rKpYjMYGUnIvsC3wzFL0+9Ovk16TaUXp2sps6rk9G0/7EiMluBvYCvAP1F5DxVfVNEWlS1PYOcscA+NZVbs8+SEpGhwNF0OG6coqrP1yizp6p+KiLLYY4gpwFvASdm+TA4x4xDgOMACcFPRHqr6sfuJRHgBuA/wNXAX1V1YQ2yh2KV727Ay8BTqvrbDOlFVVVEemLu1m4E3q2Gm8vf4dgiz/WBs1T1sdSZKS4zWNk5GUcAfULwy1OvTn7Vug2pVyevafPq0gfNb0huReStrKr/duerYksWDgW2d8sbUlcarje1CbBL1dxUtSkPYDtggBf+MnAs8AiwXg1yfwEcVuT61VjF0cuFpYKcfkA3L7xurfyAy4HT3HkX7/oZwGXAWmm4FZH7XaC/F14NmADcBByQUsb3sFbYasBA7/qZWbgBewPDvXB/4BDgbWCTGvQapOywBsBgLzyoVn556TWEbkPpdWnIax75DcmtiMwLsW+KJK4fD7wCrOrCLRXkrI6toeviwmtVy60pexgiMhXoqar7+DWocztyGDAO+L6qvppR7iXAIFXdq8T924B5qjqhgpyLgRUwh4uvq+qF7norVvtn5ue49Sv1bBGZ5O4fkFamS3cZMExVv11oWbnr/bAhoV2AK1X1z2VkXAEsDzwMrAc8A9yiqv/Mwk1ELnfxlih/EdkfOAn4jqpmckoZquxEZArmNPMz4BVV/XGt/PLSq0tbk25D6dWT17R5dXGD5jcktxJce6nqPiXun4x51dhWVf9TRs5kYATwFDYCcpqqzheRAcDmmblVU/PleWAtlPu8cCteDYq1+E7F1Yykb/kcB7R74U2BbwArJ+LNAHYpI+dnwEzgC8AY4N/A6d79gVn5YZWgz20bp8gxdO5l3QEcnaEsJ2FuWgrh5bCJDi0uPAQ4AThQS7RUMFvPPV54T+BV4Me4Fk4abk6vD3nhPlijwI9zFHAW0D2DXr8fouyczmZhQygrAB8Av6qFX156DaHbUHpdGvKaR35DcisiczywABjhwuPcsWMi3mnAbmXkHAPcDvQA1nbnNwPd3f0VsnJrqllSIrIaVsB3ufAEzOh4qYgcCaCq7wPPYx971OU0Ba4BZorIBSJyGLbJ00HADBHZ2ot3CzbcVAqfAP+jqm+r6lPAb4FjReQ0x2dOVn6qehlwn4jcIiKHYHuLjMWGz3bxZkT8NGVeEZGBwMrAPSLS4vJ8OebG5VwRaVXVd4A/AweJyOpafCz038A8sY2xAO4EngW+iHV1CyjJTUSGAG3AkyLS3el1KnCDmIuZAmZjf7jUxkdVnQzcH6Ds3sdam6qq72KTGHYXkbO8OH/Owi8PvUIw3das16UorxA4v47b8oG4+TJbgLuxIbJTROR84BRgA2CyiPzIi/4mZggvhW7Anar6mao+C/wV+75eJSJd3XuemhvQlD2Msdif9S7gSWB7rKXyV2zr10K8S4EtU8i7BDjYq/0fwMajV3LXDgH+DvRx4VHAiXj2iYS8nzpey2NdvSux2n8WsGYV/Lp65w9iMxeGufBOWLd5sAuvjLVyi3Lz5GztfsdgfmdmAX/DjLfbums/8OLvB4wtIatgnL7aHQ8D+2KtoNl0jIuW5Ya1cE7EnFk+gxkGNwBeA47w4l0I7Jqi3K6nc8/uD7WUncvjg47TGCd/nHvvNq+CX3C9htRtKL0uDXkNnd+E3NG1cku8W2e682HAuZhxekV3bS1sGHykC7dgQ+C9S8jbFqsod8G8jN+JfV8vpvO3qiK3xXHTRKrHgdfFd5m6BFjFu3YysJcXXglnoC4j82LMOPQ/dHTDBgOrJ+LdhNk2CuHuFeRe5Qr/XuAkd+3XwDpp+SXy6xsH10jEm4XrmqbkNg2b+VF4qdbBus2reXFOAI71wsNxFabPzfvtj/1p9wbGuWs9gCsSz16CWyKfawOnF7i5a98FjvLCfX0uJfI4CXjOlbkvq9ayOxu41qU7wV07D9jMi9OnHL+89BpCtyH12ux5zSO/Zbh+OSu3IjIuxRpPV3jXvoA3QcRdm4F9/AtDXkWHkbw8j8cauNcDP3LXbgB2SMutk9wsBZP3QWdbxfKJe3dihuQl4paQNRn4DWbcehTYOPmSu/B1mAfeNPz8ltRgYAUvfDeuRZSSX4vPx//DeXGuB67OUH6XAFcAF2BTBXu7670T8W7FphBX5FYsH1hr7dbkHy2lXvsn7t2Oq3RTyprs8jbclc/2JeKlLjs6z3brSefZLvcA47O+wyH1Gkq3eem1GfOaR34TaQs9EqmGWyLulVivpy/WAz/Ju+dXyjdgBuqs3PoAPbz7vyNhD0md72oShTpw3f9imfXCA91LU7GgvDRbAtO88CnYENfAhNxJwLXFlFOGn19pCDAAq/VvTsltsns5fg58s8Rz+wHnA78uxa2I3BOw8UqwFsg1uG6n98dZDjPYX1lMbjluWPdX3HGan98i/McCmyVe0mJ6vSmjXr+K2Y8K4SOAF4EvJsru3HJlV4Kfr9cWrDU6s9F6DaHbUHpdGvKaR369+9uQaKAUyXdZbkVkrorXYAW2xhqxfo9sADYUPq3Mc9NwG4hV4Del/c8twbfahLUerlDvAl5yheG3zv3u7TjgfC9c0ZKfeE4Xp5QrgbW96z2A9UvJrcDP/7isTeeeT7mX42dY7f4VzOD+Fm6WQ5HnD0+bZ5dHP2/dsY9xsqu9FnBeMblZuCVkGC8YNwAAGBJJREFUJnnfhvUG/4INBR5Uoty+Bpxco167Yq3RnQtl766PCMBvKJ2HQMuVQS56DaHbUHpdGvKaR369+zcB92PfhBnAt+jo+XRJwy1l/r+E9U62Sdz7cpl3Oi23YYn3PdN/TrVBFQZmGLrLy8QpWGsgOW3sC1mU6uJIifNpwK2V0mTk98Us/IAfAft54bHY7JydvGs/xlsgluSWIv/d3O9Q4I+UbgEmX7o03E4DtihTbqt65TYAm7o4Gc+o7e6tU45LGh151ycCM1PqNS2/MYl3p+F6rVa3IfS6tOQ1j/y6+6PpPCX3BMyovSOevYMlbaOZP8guXaHnPCTFO52W22rl5KQ9GjWtdhGwqoispKpvYuOV/wd8XUS+BCAih2MzpBZDU0z7UlcahXORxU62Dgd6ishO5dJk5LddRn6fYP6JCvHvwQxwp4rIGmKuRmap6uwy3BbDy5vPYaG7/j7Wih5diF6Baxput6vqHypwW01E1lLVD7A57fe5a1s4zt/B7ErluCyB5LMKeVfVicByInJ0pTQZ+K2beI/qqlc/f4n8VKPbUHpdGvKall+W/IJNX+1fmH6vqudjM792xhqUiMjO2ASOctw6wXHxw4V3+mKsx/D1ZJoiXNNym1BBTjpUU8uEOLB5xqfiDMfYfOgbcYtmyFADUtzW0KmFiM1JvgxbqVtXfp7MrpjrkGmJ67/EGx5LK5+E0bHI/a2xP9Aq9eKGtXCm4RZEYgstzwN+XIVei9ka/NZzwbh3HDAhpcxg/PLSa0jd5vDONW1ec8pvwfZxDOaSw58JNRn4bcZ3ZQlbQ4l4FwFH1pNbKv6hBWYouK2wYZ5jgaHu2tcxW4P/gajUXUxlC3HhFevNL8kFq7iewhauFeamPwTsnYFbWaNj4nxcPbh5L+9IzG/Qr+hYqToKW+DYvxjHEvJS2RpcePl688tDr6F1mwO3ps1rHvwSsjfEFuUdg7OzYLOapuN9DyrIqGhrSPveheaW9mjYSm9V/T02NjkEuFBENsXGPj9Q1c+8eOWGZNYHWlV1O6wb+ikwVkR2dGnbXLyVnI+XN1y4oi/4EPwS8trc6sqFmIL7AeeLyCPAq6p6XRo5IvIzbHHRJGzB4XUispuqamE1rc9JVe9w6UrqOgS3wjNV9WXsY/8e8GsR+TbWuvs/Vf0wGb9EHlfFXvTtsV7GY8D6InKES7vIxfuK4/0fFy6p15D8EnKD6NXxD6rbkNyaPa955Dch+zGsclsFOFpEjsemubb434NSEJHRWINkS/e9ehizoW0pIt0L3yrMzuanS/OtqolbJoSsfTLUiH7LYQRwJNZy/3mxOGXkrIv1LAqrtpfHhifOBb7krh0O7F9vfqXu47UksIU5X/HupTH+pjXqbV5vbglZ3YEDnC7OzqjXVbEVrgXPoX2x1aqTcMZKzPHavjW8g1Xxy7PsatVtaG7NnNd6vcdFZA/H3P78nM6z/Cq9NwOx9WD+CMgR2CLgwrdqZ+CMenPL9IxQglJkplKBtlajVALZGkLxcx+2SrKKLSJK+0f7ATA9cW1bzIXFGtjUvHXz5EYJVwQpuGfRa9W2hjz45a3XWnQbmpv/rjdbXuvIL4utrdIsutB2kGDcsh65DUmJyEEisrmIbAiLZyy1JOIsDqvqAndNNMMuUsCfsNp7HxEZqrYB0iRgHRHpoYUSTnTt8uAnIpdiLST1rknh2YVZEcXSZ8jz5cAaIjLNS3s35g+nj6q2qerTyTyH4ibmDG3zcgTLDJFUzKPH+VbMD87xIjJCzenk1cAYEelfRq/B+dVJr1CFbkNzc8++SERmichoEenurrc0Oq958hORvURkS3GzIN33IDmLqar32tPNbGyfjHEisra7diIwV9zWq/XmlhW5VBgichXm2Gt7TLmFsefCvhYrFMLJP7z/4qeBVmFryIOfiFwP9FXV/V24MEVOnYJHYI7YUo1LlnhGF7Xx+82B9URkqogMdrc3wFpmS3ANxU3M7/9IVb2rxP1hTu4S5ZYWBc5aha0hD3710KtLm1m3obmJyBnYx/p72Bj7NGA3ERngyqxhec2TnyvHPYDdgR+IyEnuuQUb6FAR6Vfrx1ersDXUi1tqlOp6VHtg488P4bqNmJ3hbeAYF+6FOdo6NsCzMtsa8uCH+ZV6EbjAhQ/E/ODMomPWwhbA4zifVjXkuav77YbNurgKm0Z4TZ7csG1yX8btbQBsjDldW6/AC/PddUHg9ymVrSEPfq7sXqqHXrPqNg9u2Myxfb3wrdj+Cd9y4W81Iq958sM8Ncz03rUjgKeBn3jvzSnAT0jpwTbFM1PZGhrBrSL34AJtBe2VOEOOuzYK88S4twtvg02BzTzlK1moRe6XtTXkxQ+rsO7GFs08hH2wfo5NoytMy90X64p3qZSPcnklo1EvBDdsj+IL3e9EJ2c6NoNpfxdndWxMdkSlvLn4wWwNofnh5uy7srszpF5r1W1e3LAG11nYvs8DsQ/yJGxLgEIDa0I985o3P2wR6at0bHe6oXtHpgLbuWvb4vkwy/Ce1mRryJNbtUc+Qs2/zx2JazsCk935cOyD3DOFrIOwruuGpQq3RGGXM4oG4Qfsg7UUCn/gL2EtvFFenGuBQ7xwRTfCBDBi5sEN2BVrDT6FtXBbsNktM7EpjAMx//xp8nh+4aXPksd68MMWeL6G89/jym5mrXoNodvQ3LDe28ZYo2kAVuHMxPa1ONLFuQbYqIq81myczosfnd1mnIztZXEQ1sA4CDNOn1zgg1VUgyrI3AtzfOo3RpNrwdLMhAzOLdQRVlhnRd8D3OuFR2Bj0r2ShVJG3lWYfeJcbEpa0ueP71489croEPywWv5+bPjq1sILS2IPDPdyZ3GPfSlLzh6RQv6SL2A9uNF56G9LOm++sjo2K61Qbl1TyLuMEn693P1hGfUajB/WgrsPa6Vu712vSa8hdBuam3tP7samov8Z2NNd74e3pwW20OxrGfM6zfGchbn1KOxHU6j40r7Hwfl53G7Hhi37Yw2sY4B9XJz1sf10KlZ6Lv712GK8y7CFpicl7g/F9guvO7eQR+0ClqxB/b0FZrpC/AHmtyeLK+sgtoY8+GHjhoWxxeWwGn5fFy78+bs6+VMy5Pl6yrvlHoHb7S15LzS3IuXWpRgvzC5wUYY8BrE15MEPG9K60Z3vh9kIBiTiZNZrCN2G5oYZl/0G0wHAfOB7BQ7Y3iAPpv1feLLOwLkNB07CWsZ7ezpP8x7nwi/B7WRsGu8+JIafsQ/2eSllBrE15MEt9FHTLCkRmQj8QkT2FJGhsNhpWKs73wlr4bYDf1PVg1y6NDMY/ov9KVZ0sp7GjFrHiMjeqvoJ1iMYVGpKWo78/oXtzoaq/hd4A3PDjqqqiHTD7CAvquqhTmbZsnazRNbD3DEjIgdiq1RneVPwhmP7/G6s7s3Jg1uJcmsTbwWuiAwQkd8An6rqkSnLDcfnd9gK14nY0NSPgCkisr/a7JkzgN5u1kuxssqL36OquoeTcTW27es+hbRe2b2QVq8uzmCsVViLbkNzU+y9L+BpbAbPCSLybcdhMHB/xv8t2NDf7Y7rudi2orvQsf/0iAp5zZOfz+0c7H3cBefoT0R6i8hk4C1VPTFlfrti0/hXVdX5mDub2cBKIrKde6efwHoYC+vMLSyqrWmwFs+92OrEGVhLNs0e1lkWz1Rta8ibn/9MrCK7wQuPrFJmEMN5LdyylBuwbZV6rdrWkAe/ZDnSMYPncLxegbtW1BFiGdk1Gahz5nYPNuS7L/ZB2wpzV/HDav8XLm4Q43Qe/CpwK7hUXz2NXALbGkJyy+uoLpENB02jw/A20hXY+XT2g38gnp0hg/yabA158yvyvK8A17vzG4FTMqTNxXBeDbcM5XYQ1Tnqq8nWkDe/Is/rhrVsj6syfZ7G86q50Xn47jTg+8ABLjwOmFSFzGDG6dD8MnD7aqn3tYjMaQSwNeTBLc+jK1VAVT8RkTew7uHRqvqyiFyHfYA3AWaLSB+sVnw3jUy3mKfNyW8XkW6qulBVx4rITBGZgRn8tgfeVhuSQq0LmCs/n1sJfIqtWp0JvKeqP0uZ56lYBfgSMEREzlPVv4jInoX8OfQD/MVq8/LglqHcump6R4K+XrUQVtX7E0MIPwHe9fS6KG9+ybITESnEdfcWisipwHdEZIW077JLPxmzw92D9YSfV9VXRGSvNLoNzS2hhzbv/3VWIuphwJNp8+lkT8WGju/Feo8Xq+rxItIP22TsHy7q8tCxn0Wp9zgkv4zcOg3hlXlvCgsIdxeRk7HKYxLm4sNfgDcRGyJcUC9uuSNL7YLrRrrz1bHa8HBcSwEz/ryC13JMUxtiBTsJ624O9a77ayr2dc86s5TcPPiV4ea3lr8AfEbn/aYr+Zep2TgdilsD9Or3IAdghumpzaZXF14POCrj/6QmA3Vobin10A/b3rTkHvclZNdsnM6LXwhuJeQWW0B4Cx0LCHtjM9rKTSrJbdJBnkdWo/ddwFsi0gub6fIENixwqIj0VdVnsVXF3f1E6kqgGETkQsyo8yD2wnxfRLZ06RbXzKp6japeqqqnu3QtReQG5VeBm3rp38a6jz/0uFVaql+TcTowt3rrdfHz1Xa+u1ZVD/b4NYVeXfhJVf2VS5fW6Fu1gTo0twx6+AjTw4QCn3L69WlQg3E6Z341cSuDfwKri8gmIjLQXXsTOM31jD4GLiz2v60Dt3yRpXbBDDE3A9934RbM+PgL4FnMqPe7DPKC2hpC8svA7RDMn88SraIUz6jKOB2aWwP1msrW0CC9HkCKjZkSsmsyUIfmllEPAyvpocxzqjJO14NftdyK3MvDDpLLpIM8j0w9DFWdgw2dnCwiE1S1XVVnqOpx2Nzj81R1B0hXG6qN5RbGpAeoOZy7Dht338TJSW1rCMkvA7cWVZ3rpUvtBExVP/WCH2CtDkTkRmB8Iq7f2grKrYF6TWVraJBeW9VtzJQWSf7aYYeZCqwtIsd593zHmO15cMuohzml8lEK0uEZdixmX+yDrXn5PTAX2yApyama9zgzv1q5JWRNxRwAbgxMwWbfHY8NlR6mqhe5qKlsDSG51R3V1DJYTfgCbuaCd70w9l5p/D6XMfMQ/PLiRuVVvGthrYyZlB7rbdpya3Z+eXJL6tZPQ4e/pB0w31ZL9JRDc6tzXosuRMNmD51ZT34huBWJG8TWkAe3RhxVLdxTqwkPBU4XkaNEZPHYu/utVBsGHzMPyC+P8fyJJBaauet+a3gO1t19WTvGLJP6aeZya3Z+uXArpltV1YJutWOm05vAE1q8pxyaWz3zujBhW+sntv3qB9phb0z2+uryH6uSWxJKjbaGHLnVHVWv9FbVP2OGu0XAN0Vk/wzJ27C5y/u7P/kN2GKmFYGHReROoF3dBip15heUW1ojpqYzTjdzuTU7v+Dc0urWhcsZqENzq3desxqn6/kfq8lwrqp/s6jyRxHZFxtevAnb5XN1F+d1ddN/kzLz5NYQhOqqkMKxViL+upjLgAmJ698AvuGFgyxQycIvFDdyMJw3c7k1O7+Q3DLoNq2BOmi5NSivqY3TDfiPZTacU+MCwjy5Nepo7MNrHDNfGrhhrkuupsPx2nDMV9IJLtwHNx66LJRbs/MLyS20bkOXWzPnNSS/kNwIbGvIo9waeTSeAGyKbRJyFDCu0XxCcCNn42+zl1uz86uFW966DV1uzZzXWvjlwY1ACwjrUW6NOnLZ0zsLtLYx81xRA7dcjb81cqsLmplfjdzyNuwHLbdmzmuN/Oq5mDOrrSH3cmsUpNk4im1o/lGjeRRDWm5iqz+nAA+q6mQ3G2JHbOfArbGx23Z1awfqya1RaGZ+WbjVW7ehy62Z85qFX0hu7sN+CfBzVX1eREYCu2MLRmeq6mwX7yBs4685LizFPvKNKLe6odFdnGX1oM7G33hE3X7e8hqSG+HtUU1bbjWVeaMJLMsHTWz8jUfU7bKQ11q40cQLTZv1qMq9eUQ6qOrvReRQ4FoR6Qv8S1XvUPe2aLMs94/IjM+Tbps5rzVyuwtYNWFrWB+zNVyhqs+KSE0LTZu13KpFw43eyzq0iY2/EbXh86TbZs5rDdyaeaFpU6LpjN7LOprZ+BtRGz5Pum3mvGY07K+L7TF/iqpe613/BoCq/tGFixq48+TWjIgVRkRExOcaIrIVtinSBao6zbsuqqol3PR8LhErjIiIiM89RGRTbH/1X+JsDQ2m1JSIFUZEREQEICKrYeskvgQ8p6rTG8uo+RArjIiIiIgElnZbQ16IFUZERERERCrEabUREREREakQK4yIiIiIiFSIFUZERERERCrECiMiIiIiIhVihRGxBETkNREZXGucpQ3V5klEVhGRvb3wBiLyq0CcTqky3VQRGRWCQwn5Y0Rkey+8o4iclNfzIpoDscKIWKYhIvVwsLkKsLjCUNXHVfWoQLKLVhhiKPn/VdWDVfWFWh5coezGAIsrDFWdparn1vK8iOZHrDCWEbhW7osiMl1EXhKR34jIViIyW0T+KSJfFZFBIjJDRJ4RkUdFZB2XdjkRuVdEnheRqYB4cr8rIn8VkadE5HIR6ZKSy99F5Aon814R6enuHSIij4nI0yLyW+cpFMf7UsfrFRHZQkSucnKme7K3EZFHROQJEblZRPoUef4fRORCEXkcOFpElnfPeswdm5TLt+P/nCfveBGZ6M5HisjvHf8nRGRV4FxgM1dGxzruv3PxS5X5RJe/P7j8LlHBiMi5QE8n9zeO1z9E5BrgOWAlV2aPuzyckSiDDdz5PBE523F+VESGlNHddBG5TET+Apzv3ptHRORJEXlYRNYQkVbgTGAPx20PEdlfRC72yu8Bl+f7RWTlSu9MxFKCRvtXj0eYA2vlLsJ8+LcAfwOuwj6COwEzgIuAn7j43wKecue/Ak535+MABQYDa2Gb3Xdz9yYD+7rz14DBFbiMceGbgO+68+W8eGcBR7rz6Zi30ALfjxJ5GeM4/Qno7dKcWOCdeP4fgMle+DpgU3e+MvD3CvleBVvpW0h/PDDRnf8F2MWd9wB6AVsAv/PiLw6XKfOJwMOY6+zBwH8L5ZzIy7xEubYDX/OuDXK/XVy+1/HKYAN3rsC33fn5wGll3qPpmDO+Li7cD+jqzrcCfuvO9wcu9tItDmPvzH7u/EBgRqP/H/EIc8T9MJYtvKq2XzAi8jxwv6qqiDyLfWyGA7sBqOoDroXdD9s6cld3/Q4RmePkbYntD/CYiAD0BN7NwOUpd/4393yA0SJyFjAA28XsHi/N7R7fdxJ5WQVzOz0KmO34tAKPlHj+jd75VsAolwagn+uZlMp3UYjtaTBMVW9zaT5z18sl25TiZQ5wh6rOB+aLyLvAEOCNcsKA/1XVR73weLE9F7oCX8TK55lEmgVYJQCmi60rPONmVW1z5/2Bq8XcZijQrUJagI1x5Yr5Zzo/RZqIpQCxwli2MN87b/fC7ZiuF2aUJ8DVqnpyjVzasMoGrAW7s6o+LbY3wBZF0vjcC+GuTs59qrpXiud/7J23YK3yz/wIZT70i+g8XNsjxfOqQbKM0vwfF+dLREZgvZ8NVXWOG7orxnWhqhZcOqR5jl92P8X2pt5FRFbBei4Rn1NEG8bnCw8B+wCIyBbAe2r+cv6EM9qKyHbAQBf/fuA7IrKCuzdIRIbXyKEv8JaIdCtwyYBHgU1EZKTj01tEVnfn54jILiXS3QscWQiIyBh3Wirf7wAruN5Ad2AHAFWdC7whIju7NN2dDWauy1cxlCrzknDj/sNccKErq2Loh33cP3R2ie3Kya0S/YE33fn+3vVyeX4Y2NOd74OVQcQygFhhfL4wEVhfRJ7BDLX7uetnAJu7oZ9dgX8DqM2yOQ2416W5Dxv2qAU/xuwAs4EXsyRU1f9gH63rHZ9HgDXd7bWBt0skPQrYwBlhXwAOc9dL5XshZtT9K5Znn+cE4Cj3/IeBL2BDQG3OqHxs4tkTKV7mRSE282kk8L67NAV4RkR+U6Q8ngaedPyuw8o0NM4HzhGRJ+ncM3kQG+Z7SkT2SKQ5EjjA5XkCcHQOvCIagOh8MGKZgIjco6pjG82jVojIaOBAVT2u0VwiIpKIFUZERERERCpEo3dE1RCR5TA7RxJbqup/680nIj1E5FRg98Tlm1X17EbwiVg6EHsYERERERGpEI3eERERERGpECuMiIiIiIhUiBVGREREREQqxAojIiIiIiIVYoUREREREZEK/w9KXIMYhTrQugAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f1_scores = results[['f1-test', 'model_name', 'train_ratio', 'reduction']].groupby(['model_name', 'reduction','train_ratio']).agg(np.mean)\n",
    "\n",
    "xticks = [str((x,y)) for x in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999] for y in [0.5, 0.75, 0.9]]\n",
    "colors = ['b']*len(unsupervised) + ['g']*len(semi_supervised) + ['r']*len(supervised)\n",
    "fig, ax = plt.subplots()\n",
    "for model, color in zip(unsupervised + semi_supervised + supervised, colors):\n",
    "    f1_scores.loc[[model]].plot(ax=ax, legend=False, color=color)\n",
    "    \n",
    "_ = plt.xticks(range(len(xticks)),xticks, rotation=45)\n",
    "_ = plt.ylim(0, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_lip(model, training_ratio, lipschitz):\n",
    "    test_f1 = results[results.model_name.str.contains(model)].groupby(['reduction', 'train_ratio']).agg(np.mean)['f1-test'].loc[:,training_ratio].values\n",
    "    final_results = []\n",
    "    for i in range(len(test_f1)):\n",
    "        elem_i = test_f1[i]\n",
    "        for j in range(i+1, len(test_f1)):\n",
    "            elem_j = test_f1[j]\n",
    "            if(lipschitz):\n",
    "                lipschitz_cte_estimator = (abs(elem_j-elem_i)/abs(j-i))\n",
    "                final_results.append((j-i, lipschitz_cte_estimator))\n",
    "            else:\n",
    "                final_results.append(((j-i), (elem_j-elem_i)**2))\n",
    "    final_results.sort(key=lambda tup: tup[0])  # sorts in place\n",
    "    return final_results, np.mean([x[1] for x in final_results])\n",
    "\n",
    "def plot_distance_performance(model, training_ratio, lipschitz, ax, marker):\n",
    "    l = compute_lip(model, training_ratio, lipschitz)[0]\n",
    "    d = defaultdict(list)\n",
    "    distance_performance = []\n",
    "    for key, value in l:\n",
    "        d[key].append(value)\n",
    "\n",
    "    for key in d:\n",
    "        distance_performance.append((key, np.mean(d[key])))\n",
    "\n",
    "    ax.scatter(*zip(*distance_performance), label=model, marker=marker, s=50, alpha=0.8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x2160 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2, ax3) = plt.subplots(3, sharex=True, figsize=(16, 30))\n",
    "\n",
    "lipschitz = True\n",
    "training_ratios = [0.5, 0.75, 0.9]\n",
    "lipschitz_per_model = []\n",
    "\n",
    "for ax, training_ratio in zip([ax1, ax2, ax3], training_ratios):\n",
    "    for model in unsupervised + supervised + semi_supervised:\n",
    "        lipschitz_per_model.append((model, training_ratio, compute_lip(model, training_ratio, lipschitz)[1]))\n",
    "        \n",
    "        if(model in unsupervised_models):\n",
    "            plot_distance_performance(model, training_ratio, lipschitz, ax, 'x')\n",
    "        elif(model in supervised_models):\n",
    "            plot_distance_performance(model, training_ratio,lipschitz, ax, 's')\n",
    "        else:\n",
    "            plot_distance_performance(model, training_ratio, lipschitz, ax, '+')\n",
    "\n",
    "    if(lipschitz):\n",
    "        ax.set_title('Training ratio:'+ str(training_ratio) + ' - Percentage distance vs Lipschitz constant')\n",
    "    else:\n",
    "        ax.set_title('Training ratio:'+ str(training_ratio) + '- Percentage distance vs Difference Performance Squared')\n",
    "    legend1 = ax.legend(loc=1)\n",
    "    legend_elements = [Line2D([], [], marker='x', color='black', linestyle='None', label='Unsupervised'),\n",
    "                       Line2D([], [], marker='+', color='black', linestyle='None', label='Semi-Supervised'),\n",
    "                       Line2D([], [], marker='s',color='black', linestyle='None', label='Supervised')]\n",
    "    \n",
    "    legend2 = ax.legend(handles=legend_elements, loc=9) \n",
    "    ax.add_artist(legend1)\n",
    "    ax.add_artist(legend2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Overall Ranking**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0\n",
       "LogClustering      0.024873\n",
       "DecisionTree       0.093811\n",
       "PCA                0.131013\n",
       "InvariantsMiner    0.146302\n",
       "IsolationForest    0.174642\n",
       "LR                 0.190742\n",
       "SVM                0.200110\n",
       "Name: 2, dtype: float64"
      ]
     },
     "execution_count": 259,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(lipschitz_per_model).groupby(0).agg(np.mean)[2].sort_values(ascending=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Observations:** \n",
    "- Train scores are fixed by train ratio.\n",
    "- Supervised methods \"fall\" harder than other methods.\n",
    "- LogClustering score goes up."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>f1-test</th>\n",
       "      <th>fn_test</th>\n",
       "      <th>fp_test</th>\n",
       "      <th>tn_test</th>\n",
       "      <th>tp_test</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>reduction</th>\n",
       "      <th>train_ratio</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">0.000</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.844338</td>\n",
       "      <td>2268</td>\n",
       "      <td>0</td>\n",
       "      <td>279112</td>\n",
       "      <td>6151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.838781</td>\n",
       "      <td>1169</td>\n",
       "      <td>0</td>\n",
       "      <td>139556</td>\n",
       "      <td>3041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.846970</td>\n",
       "      <td>447</td>\n",
       "      <td>0</td>\n",
       "      <td>55823</td>\n",
       "      <td>1237</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">0.250</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.838377</td>\n",
       "      <td>1757</td>\n",
       "      <td>0</td>\n",
       "      <td>279112</td>\n",
       "      <td>4557</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.842954</td>\n",
       "      <td>857</td>\n",
       "      <td>0</td>\n",
       "      <td>139556</td>\n",
       "      <td>2300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.837017</td>\n",
       "      <td>354</td>\n",
       "      <td>0</td>\n",
       "      <td>55823</td>\n",
       "      <td>909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">0.500</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.839857</td>\n",
       "      <td>1162</td>\n",
       "      <td>0</td>\n",
       "      <td>279112</td>\n",
       "      <td>3047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.846259</td>\n",
       "      <td>561</td>\n",
       "      <td>0</td>\n",
       "      <td>139556</td>\n",
       "      <td>1544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.852079</td>\n",
       "      <td>217</td>\n",
       "      <td>0</td>\n",
       "      <td>55823</td>\n",
       "      <td>625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">0.750</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.831436</td>\n",
       "      <td>607</td>\n",
       "      <td>0</td>\n",
       "      <td>279112</td>\n",
       "      <td>1497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.842684</td>\n",
       "      <td>286</td>\n",
       "      <td>0</td>\n",
       "      <td>139556</td>\n",
       "      <td>766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.828929</td>\n",
       "      <td>123</td>\n",
       "      <td>0</td>\n",
       "      <td>55823</td>\n",
       "      <td>298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">0.990</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.841379</td>\n",
       "      <td>23</td>\n",
       "      <td>0</td>\n",
       "      <td>279112</td>\n",
       "      <td>61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.816901</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>139556</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.720000</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>55823</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">0.999</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.769231</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>279112</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.857143</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>139556</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>55823</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        f1-test  fn_test  fp_test  tn_test  tp_test\n",
       "reduction train_ratio                                              \n",
       "0.000     0.50         0.844338     2268        0   279112     6151\n",
       "          0.75         0.838781     1169        0   139556     3041\n",
       "          0.90         0.846970      447        0    55823     1237\n",
       "0.250     0.50         0.838377     1757        0   279112     4557\n",
       "          0.75         0.842954      857        0   139556     2300\n",
       "          0.90         0.837017      354        0    55823      909\n",
       "0.500     0.50         0.839857     1162        0   279112     3047\n",
       "          0.75         0.846259      561        0   139556     1544\n",
       "          0.90         0.852079      217        0    55823      625\n",
       "0.750     0.50         0.831436      607        0   279112     1497\n",
       "          0.75         0.842684      286        0   139556      766\n",
       "          0.90         0.828929      123        0    55823      298\n",
       "0.990     0.50         0.841379       23        0   279112       61\n",
       "          0.75         0.816901       13        0   139556       29\n",
       "          0.90         0.720000        7        0    55823        9\n",
       "0.999     0.50         0.769231        3        0   279112        5\n",
       "          0.75         0.857143        1        0   139556        3\n",
       "          0.90         1.000000        0        0    55823        1"
      ]
     },
     "execution_count": 273,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns_test = list(filter(lambda x : 'test' in x and len(x)<8, results.columns))\n",
    "results[results.model_name == 'LogClustering'].groupby(['reduction', 'train_ratio']).agg(np.mean)[columns_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 717,
   "metadata": {},
   "outputs": [],
   "source": [
    "results = pd.concat(dfs)\n",
    "results_grouped = results[results.reduction_train == 0.0].groupby(['model_name', 'reduction_train', 'reduction_test', 'train_ratio']).agg(np.mean)[['f1-test', 'f1-train']]\n",
    "results_grouped.loc['DecisionTree']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 718,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pick(results, train_ratio=None, reduction_train=None, reduction_test=None, model=None):\n",
    "    tmp = results\n",
    "    if(train_ratio != None):\n",
    "        tmp = tmp[(tmp.index.get_level_values('train_ratio') == train_ratio)]\n",
    "    if(reduction_train != None):\n",
    "        tmp = tmp[(tmp.index.get_level_values('reduction_train') == reduction_train)]\n",
    "    if(reduction_test != None):\n",
    "        tmp = tmp[(tmp.index.get_level_values('reduction_test') == reduction_test)]\n",
    "    if(model != None):\n",
    "        tmp = tmp.loc[model]\n",
    "    return tmp\n",
    "\n",
    "def compute_lip(model, results, lipschitz, train_ratio, reduction_train=None, reduction_test=None):\n",
    "    #test_f1 = results[results.model.str.contains(model)].groupby(['reduction_train']).agg(np.mean)['f_score'].values\n",
    "    test_f1 = pick(results, train_ratio=train_ratio, reduction_train=reduction_train,  reduction_test=reduction_test, model=model).values\n",
    "    final_results = []\n",
    "    for i in range(len(test_f1)):\n",
    "        elem_i = test_f1[i]\n",
    "        for j in range(i+1, len(test_f1)):\n",
    "            elem_j = test_f1[j]\n",
    "            if(lipschitz):\n",
    "                lipschitz_cte_estimator = (abs(elem_j-elem_i)/abs(j-i))\n",
    "                final_results.append((j-i, lipschitz_cte_estimator))\n",
    "            else:\n",
    "                final_results.append(((j-i), (elem_j-elem_i)**2))\n",
    "    final_results.sort(key=lambda tup: tup[0])  # sorts in place\n",
    "    return final_results, np.mean([x[1] for x in final_results])\n",
    "    \n",
    "def plot_distance_performance(model,l, ax, marker):\n",
    "    d = defaultdict(list)\n",
    "    distance_performance = []\n",
    "    for key, value in l:\n",
    "        d[key].append(value)\n",
    "\n",
    "    for key in d:\n",
    "        distance_performance.append((key, np.mean(d[key])))\n",
    "\n",
    "    ax.scatter(*zip(*distance_performance), label=model, marker=marker, s=50, alpha=0.8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 719,
   "metadata": {},
   "outputs": [],
   "source": [
    "reduction_train= 0\n",
    "lipschitz = True\n",
    "training_ratios = [0.5, 0.75, 0.9]\n",
    "lipschitz_per_model = []\n",
    "\n",
    "for training_ratio in training_ratios:\n",
    "    for model in set(results.model_name):\n",
    "        l, mean_lip = compute_lip(model, results_grouped, lipschitz, training_ratio, reduction_train)\n",
    "        lipschitz_per_model.append((model, training_ratio, mean_lip))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 720,
   "metadata": {},
   "outputs": [],
   "source": [
    "deeplog = pd.read_csv('deeplog.csv', names=['train_ratio', 'reduction_test', 'f1-test'], header=0)\n",
    "deeplog['model_name'] = 'DeepLog'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 721,
   "metadata": {},
   "outputs": [],
   "source": [
    "results_all = pd.concat([results, deeplog]).fillna(0)\n",
    "results_all_grouped = results_all[results_all.reduction_train == 0.0].groupby(['model_name', 'reduction_train', 'reduction_test', 'train_ratio']).agg(np.mean)[['f1-test', 'f1-train']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 722,
   "metadata": {},
   "outputs": [],
   "source": [
    "reduction_train= 0\n",
    "lipschitz = True\n",
    "training_ratios = [0.5, 0.75, 0.9]\n",
    "lipschitz_per_model = []\n",
    "\n",
    "for training_ratio in training_ratios:\n",
    "    for model in set(results_all.model_name):\n",
    "        l, mean_lip = compute_lip(model, results_all_grouped, lipschitz, training_ratio, reduction_train)\n",
    "        lipschitz_per_model.append((model, training_ratio, mean_lip))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Average constant"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 735,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Save dataframe to csv\n",
    "results_all_grouped.to_csv('Robustness-HDFS.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 732,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f6f7814d6a0>"
      ]
     },
     "execution_count": 732,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib.patches import Patch\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "robustness_df = pd.DataFrame(lipschitz_per_model, columns=['Model', 'train_ratio', 'lipschitz'])\n",
    "robustness_df_grouped = robustness_df.groupby(['Model']).agg(np.mean)['lipschitz'].sort_values()\n",
    "random_models = set([x for x in results_all['model_name'] if('random' in x)])\n",
    "color_code = ['r' if x in supervised_models else 'g' if (x in unsupervised_models) else 'b' if(x in random_models) else 'orange' for x  in robustness_df_grouped.index]\n",
    "robustness_df_grouped.plot.bar(color=color_code)\n",
    "\n",
    "\n",
    "legend_elements = [Patch(facecolor='r',label='Supervised'),\n",
    "                   Patch(facecolor='g',label='Unsupervised'),\n",
    "                   Patch(facecolor='b',label='Semi-supervised')]\n",
    "\n",
    "legend = ax.legend(handles=legend_elements, loc=2)\n",
    "ax.add_artist(legend)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Average constant by train ratio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 729,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f6f69cb9198>"
      ]
     },
     "execution_count": 729,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lip_df = pd.DataFrame(lipschitz_per_model, columns=['model', 'train_ratio', 'Lipschitz_constant'])\n",
    "lip_df.pivot(index='model', columns='train_ratio', values='Lipschitz_constant').sort_values(by=0.5).plot(kind='bar', stacked=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 733,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for train_ratio in [0.5, 0.75, 0.9]:\n",
    "    fig, ax = plt.subplots()\n",
    "    sorted_lip = lip_df[lip_df.train_ratio == train_ratio].set_index('model')['Lipschitz_constant'].sort_values()\n",
    "    color_code = ['r' if x in supervised_models else 'g' if (x in unsupervised_models) else 'b' if(x in random_models) else 'orange' for x in sorted_lip.index]\n",
    "    sorted_lip.plot.bar(color=color_code)\n",
    "    plt.title('Training Ratio ' + str(train_ratio))\n",
    "    \n",
    "    legend_elements = [Patch(facecolor='r',label='Supervised'),\n",
    "                       Patch(facecolor='g',label='Unsupervised'),\n",
    "                       Patch(facecolor='b',label='Semi-supervised')]\n",
    "\n",
    "    legend = ax.legend(handles=legend_elements, loc=2)\n",
    "    ax.add_artist(legend)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 533,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 7))\n",
    "xticks = [str((x,y)) for x in [0, 0.25, 0.5, 0.75, 0.99, 0.999] for y in [0.5, 0.75, 0.9]]\n",
    "for model in set(results_all.model_name):\n",
    "    results_all_grouped.loc[model]['f1-test'].plot()\n",
    "    plt.xticks(range(len(xticks)), xticks, rotation=47)\n",
    "plt.legend(set(results_all.model_name))\n",
    "_ = plt.xlabel('(Reduction_Train, Traing_Ratio)')\n",
    "for i in range(3, 16, 3):\n",
    "    plt.axvline(i, color='red', linestyle='--')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 652,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xticks = [str(x) for x in [0, 0.25, 0.5, 0.75, 0.99, 0.999]]\n",
    "for train_ratio in [0.50, 0.75, 0.9]:\n",
    "    plt.figure(figsize=(10,5))\n",
    "    for model in set(results_all.model_name):\n",
    "        results_all_grouped.loc[model,:,:,train_ratio]['f1-test'].plot()\n",
    "        plt.xticks(range(len(xticks)), xticks)\n",
    "        plt.title('Training Ratio '+ str(train_ratio))\n",
    "        _, _ = plt.xlabel('Reduction Test'), plt.ylabel('F-Score')\n",
    "        \n",
    "    plt.legend(set(results_all.model_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 655,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>f1-test</th>\n",
       "      <th>lipschitz</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model_name</th>\n",
       "      <th>train_ratio</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>LogClustering</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.853924</td>\n",
       "      <td>0.008868</td>\n",
       "      <td>-0.017923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DecisionTree</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.840997</td>\n",
       "      <td>0.047963</td>\n",
       "      <td>-0.101729</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LogClustering</th>\n",
       "      <th>0.90</th>\n",
       "      <td>0.827703</td>\n",
       "      <td>0.025527</td>\n",
       "      <td>-0.052678</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DecisionTree</th>\n",
       "      <th>0.90</th>\n",
       "      <td>0.827392</td>\n",
       "      <td>0.054537</td>\n",
       "      <td>-0.116771</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LogClustering</th>\n",
       "      <th>0.75</th>\n",
       "      <td>0.799575</td>\n",
       "      <td>0.016275</td>\n",
       "      <td>-0.033227</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            f1-test  lipschitz     score\n",
       "model_name    train_ratio                               \n",
       "LogClustering 0.50         0.853924   0.008868 -0.017923\n",
       "DecisionTree  0.50         0.840997   0.047963 -0.101729\n",
       "LogClustering 0.90         0.827703   0.025527 -0.052678\n",
       "DecisionTree  0.90         0.827392   0.054537 -0.116771\n",
       "LogClustering 0.75         0.799575   0.016275 -0.033227"
      ]
     },
     "execution_count": 655,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_lipschitz = results_all.groupby(['model_name', 'train_ratio']).agg(np.mean)[['f1-test']]\n",
    "f1_lipschitz['lipschitz'] = robustness_df.groupby(['Model', 'train_ratio']).agg(np.mean)['lipschitz']\n",
    "f1_lipschitz.sort_values(['f1-test', 'lipschitz'], ascending=[False, True]).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 653,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>f1-test</th>\n",
       "      <th>lipschitz</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model_name</th>\n",
       "      <th>train_ratio</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">SVM</th>\n",
       "      <th>0.90</th>\n",
       "      <td>0.615792</td>\n",
       "      <td>0.100536</td>\n",
       "      <td>-0.240306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.50</th>\n",
       "      <td>0.594799</td>\n",
       "      <td>0.099903</td>\n",
       "      <td>-0.240139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.610968</td>\n",
       "      <td>0.099884</td>\n",
       "      <td>-0.238809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">LR</th>\n",
       "      <th>0.75</th>\n",
       "      <td>0.485184</td>\n",
       "      <td>0.095171</td>\n",
       "      <td>-0.236790</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.560911</td>\n",
       "      <td>0.094774</td>\n",
       "      <td>-0.228086</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         f1-test  lipschitz     score\n",
       "model_name train_ratio                               \n",
       "SVM        0.90         0.615792   0.100536 -0.240306\n",
       "           0.50         0.594799   0.099903 -0.240139\n",
       "           0.75         0.610968   0.099884 -0.238809\n",
       "LR         0.75         0.485184   0.095171 -0.236790\n",
       "           0.90         0.560911   0.094774 -0.228086"
      ]
     },
     "execution_count": 653,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_lipschitz['score'] = stats.hmean(f1_lipschitz[['f1-test', 'lipschitz']], axis=1)\n",
    "f1_lipschitz.sort_values('score', ascending=True).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 654,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>f1-test</th>\n",
       "      <th>lipschitz</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model_name</th>\n",
       "      <th>train_ratio</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">LogClustering</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.853924</td>\n",
       "      <td>0.008868</td>\n",
       "      <td>-0.017923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>0.799575</td>\n",
       "      <td>0.016275</td>\n",
       "      <td>-0.033227</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.827703</td>\n",
       "      <td>0.025527</td>\n",
       "      <td>-0.052678</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DeepLog</th>\n",
       "      <th>0.90</th>\n",
       "      <td>0.666577</td>\n",
       "      <td>0.039245</td>\n",
       "      <td>-0.083399</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DecisionTree</th>\n",
       "      <th>0.50</th>\n",
       "      <td>0.840997</td>\n",
       "      <td>0.047963</td>\n",
       "      <td>-0.101729</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            f1-test  lipschitz     score\n",
       "model_name    train_ratio                               \n",
       "LogClustering 0.50         0.853924   0.008868 -0.017923\n",
       "              0.75         0.799575   0.016275 -0.033227\n",
       "              0.90         0.827703   0.025527 -0.052678\n",
       "DeepLog       0.90         0.666577   0.039245 -0.083399\n",
       "DecisionTree  0.50         0.840997   0.047963 -0.101729"
      ]
     },
     "execution_count": 654,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from scipy import stats\n",
    "f1_lipschitz['score'] = 2*(f1_lipschitz['f1-test']*(-f1_lipschitz['lipschitz']))/(f1_lipschitz['f1-test']-f1_lipschitz['lipschitz'])\n",
    "f1_lipschitz.sort_values('score', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start = time.time()\n",
    "\n",
    "supervised = ['LR', 'SVM', 'DecisionTree']\n",
    "unsupervised = ['InvariantsMiner', 'PCA', 'IsolationForest']\n",
    "semi_supervised = ['LogClustering']\n",
    "\n",
    "#Load data\n",
    "label_file = 'data/HDFS/anomaly_label.csv'\n",
    "path = '../../../loglizer/HDFS_Data/'\n",
    "\n",
    "#dfs = []\n",
    "for train_ratio in [0.5, 0.75, 0.9]:\n",
    "    for reduction_train in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999]:\n",
    "        print('Training ratio', train_ratio, '- Reduction percentage', reduction_train)\n",
    "        #Load training data\n",
    "        data_train = np.load(path + 'train/' + str(train_ratio) + '/'+ str(reduction_train) +'/HDFS_train_' + str(train_ratio) + '_' + str(reduction_train) +  '.npz', allow_pickle=True)\n",
    "        x_tr, y_train = data_train['x_data'], data_train['y_data']\n",
    "\n",
    "\n",
    "        for reduction_test in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999]:\n",
    "            print('Reduction percentage', reduction_test)\n",
    "\n",
    "            #Load test data\n",
    "            data_test = np.load(path + 'test/' + str(train_ratio) + '/'+ str(reduction_test) +'/HDFS_test_' + str(train_ratio) + '_' + str(reduction_test) + '.npz', allow_pickle=True)\n",
    "            x_te, y_test = data_test['x_data'], data_test['y_data']\n",
    "\n",
    "            print('Train accuracy:')\n",
    "            anomaly_perc = sum(y_train)/len(y_train)\n",
    "            y_pred_train = np.random.choice([0, 1], size=(len(y_train),), p=[1-anomaly_perc, anomaly_perc])\n",
    "            y_pred_test = np.random.choice([0, 1], size=(len(y_test),), p=[1-anomaly_perc, anomaly_perc])\n",
    "            \n",
    "            precision_train, recall_train, f1_train, _ = precision_recall_fscore_support(y_train, y_pred_train, average='binary')\n",
    "            mcc_train = matthews_corrcoef(y_train, y_pred_train)\n",
    "            roc_auc_train = roc_auc_score(y_train, y_pred_train)\n",
    "            tn_train, fp_train, fn_train, tp_train = confusion_matrix(y_train, y_pred_train).ravel()\n",
    "            \n",
    "            print('Test accuracy:')\n",
    "            precision_test, recall_test, f1_test, _ = precision_recall_fscore_support(y_test, y_pred_test, average='binary')\n",
    "            mcc_test = matthews_corrcoef(y_test, y_pred_test)\n",
    "            roc_auc_test = roc_auc_score(y_test, y_pred_test)\n",
    "            tn_test, fp_test, fn_test, tp_test = confusion_matrix(y_test, y_pred_test).ravel()\n",
    "\n",
    "            \n",
    "            results = [('model_name', 'random_anomaly_perc'), ('model', 'random_anomaly_perc'), \n",
    "                       ('f1-train', f1_train), ('f1-test', f1_test),\n",
    "                       ('train_ratio', train_ratio),  ('reduction_train', reduction_train), \n",
    "                       ('reduction_test', reduction_test), \n",
    "                       ('precision_train', precision_train), ('precision_test', precision_test),\n",
    "                       ('recall_train', recall_train),('recall_test', recall_test), \n",
    "                       ('mcc_train', mcc_train), ('mcc_test', mcc_test),\n",
    "                       ('roc_auc_train', roc_auc_train), ('roc_auc_test', roc_auc_test),\n",
    "                       ('tn_train', tn_train),('tn_test', tn_test), \n",
    "                       ('fp_train', fp_train), ('fp_test', fp_test), \n",
    "                       ('fn_train', fn_train), ('fn_test', fn_test),\n",
    "                       ('tp_train', tp_train), ('tp_test', tp_test)]\n",
    "\n",
    "            dfs.append(pd.DataFrame({x[0]:x[1:] for x in results}))\n",
    "\n",
    "end = time.time()\n",
    "\n",
    "print('It took:', end-start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Baselines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 710,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training ratio 0.5 - Reduction percentage 0.0\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.5 - Reduction percentage 0.25\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.5 - Reduction percentage 0.5\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.5 - Reduction percentage 0.75\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.5 - Reduction percentage 0.99\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.5 - Reduction percentage 0.999\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.75 - Reduction percentage 0.0\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.75 - Reduction percentage 0.25\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.75 - Reduction percentage 0.5\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.75 - Reduction percentage 0.75\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.75 - Reduction percentage 0.99\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.75 - Reduction percentage 0.999\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.9 - Reduction percentage 0.0\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.9 - Reduction percentage 0.25\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.9 - Reduction percentage 0.5\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.9 - Reduction percentage 0.75\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.9 - Reduction percentage 0.99\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Training ratio 0.9 - Reduction percentage 0.999\n",
      "Reduction percentage 0.0\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.25\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.5\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.75\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.99\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy:\n",
      "Test accuracy:\n",
      "Reduction percentage 0.999\n",
      "Train accuracy:\n",
      "Test accuracy:\n",
      "It took: 635.617330789566\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "\n",
    "#Load data\n",
    "label_file = 'data/HDFS/anomaly_label.csv'\n",
    "path = '../../../loglizer/HDFS_Data/'\n",
    "\n",
    "#dfs = []\n",
    "for train_ratio in [0.5, 0.75, 0.9]:\n",
    "    for reduction_train in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999]:\n",
    "        print('Training ratio', train_ratio, '- Reduction percentage', reduction_train)\n",
    "        #Load training data\n",
    "        data_train = np.load(path + 'train/' + str(train_ratio) + '/'+ str(reduction_train) +'/HDFS_train_' + str(train_ratio) + '_' + str(reduction_train) +  '.npz', allow_pickle=True)\n",
    "        x_tr, y_train = data_train['x_data'], data_train['y_data']]\n",
    "\n",
    "        for reduction_test in [0.0, 0.25, 0.5, 0.75, 0.99, 0.999]:\n",
    "            print('Reduction percentage', reduction_test)\n",
    "\n",
    "            #Load test data\n",
    "            data_test = np.load(path + 'test/' + str(train_ratio) + '/'+ str(reduction_test) +'/HDFS_test_' + str(train_ratio) + '_' + str(reduction_test) + '.npz', allow_pickle=True)\n",
    "            x_te, y_test = data_test['x_data'], data_test['y_data']\n",
    "\n",
    "            print('Train accuracy:')\n",
    "            anomaly_perc = sum(y_train)/len(y_train)\n",
    "            y_pred_train = np.random.choice([0, 1], size=(len(y_train),), p=[1-anomaly_perc, anomaly_perc])\n",
    "            y_pred_test = np.random.choice([0, 1], size=(len(y_test),), p=[1-anomaly_perc, anomaly_perc])\n",
    "            \n",
    "            precision_train, recall_train, f1_train, _ = precision_recall_fscore_support(y_train, y_pred_train, average='binary')\n",
    "            mcc_train = matthews_corrcoef(y_train, y_pred_train)\n",
    "            roc_auc_train = roc_auc_score(y_train, y_pred_train)\n",
    "            tn_train, fp_train, fn_train, tp_train = confusion_matrix(y_train, y_pred_train).ravel()\n",
    "            \n",
    "            print('Test accuracy:')\n",
    "            precision_test, recall_test, f1_test, _ = precision_recall_fscore_support(y_test, y_pred_test, average='binary')\n",
    "            mcc_test = matthews_corrcoef(y_test, y_pred_test)\n",
    "            roc_auc_test = roc_auc_score(y_test, y_pred_test)\n",
    "            tn_test, fp_test, fn_test, tp_test = confusion_matrix(y_test, y_pred_test).ravel()\n",
    "\n",
    "            \n",
    "            results = [('model_name', 'random_anomaly_perc'), ('model', 'random_anomaly_perc'), \n",
    "                       ('f1-train', f1_train), ('f1-test', f1_test),\n",
    "                       ('train_ratio', train_ratio),  ('reduction_train', reduction_train), \n",
    "                       ('reduction_test', reduction_test), \n",
    "                       ('precision_train', precision_train), ('precision_test', precision_test),\n",
    "                       ('recall_train', recall_train),('recall_test', recall_test), \n",
    "                       ('mcc_train', mcc_train), ('mcc_test', mcc_test),\n",
    "                       ('roc_auc_train', roc_auc_train), ('roc_auc_test', roc_auc_test),\n",
    "                       ('tn_train', tn_train),('tn_test', tn_test), \n",
    "                       ('fp_train', fp_train), ('fp_test', fp_test), \n",
    "                       ('fn_train', fn_train), ('fn_test', fn_test),\n",
    "                       ('tp_train', tp_train), ('tp_test', tp_test)]\n",
    "\n",
    "            dfs.append(pd.DataFrame({x[0]:x[1:] for x in results}))\n",
    "\n",
    "end = time.time()\n",
    "\n",
    "print('It took:', end-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 714,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "with open('dfs_with_random.pkl', 'wb') as _:\n",
    "    pickle.dump(dfs, _)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}