CreditCardFraudDetection (1).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Importing necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import make_scorer,f1_score,r2_score\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn.metrics import roc_auc_score, log_loss, accuracy_score,roc_curve\n",
    "from sklearn.metrics import confusion_matrix, recall_score, precision_score\n",
    "from sklearn.preprocessing import binarize, PolynomialFeatures\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.linear_model import LogisticRegression, RidgeClassifier\n",
    "from sklearn.calibration import calibration_curve\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "import warnings\n",
    "from sklearn.metrics import mean_squared_error\n",
    "warnings.filterwarnings('ignore')\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Creating functions which are being used repeateadly"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define a function to print accuracy metrics\n",
    "def print_accuracy_metrics(Input,Output):\n",
    "  print(\"Recall:\", recall_score(Input, Output))\n",
    "  print(\"Log Loss:\", log_loss(Input, Output))\n",
    "  print(\"Precision:\", precision_score(Input, Output))\n",
    "  print(\"Accurcay:\", accuracy_score(Input, Output))\n",
    "  print(\"AUC: \", roc_auc_score(Input, Output))\n",
    "  print(\"F1 Score:\", f1_score(Input, Output))\n",
    "  confusion_matrix_value = confusion_matrix(Input,Output)\n",
    "  print('Confusion matrix:\\n', confusion_matrix_value)\n",
    "  class_names=[0,1] # name  of classes\n",
    "  fig, ax = plt.subplots()\n",
    "  tick_marks = np.arange(len(class_names))\n",
    "  plt.xticks(tick_marks, class_names)\n",
    "  plt.yticks(tick_marks, class_names)\n",
    "  # create heatmap\n",
    "  sns.heatmap(pd.DataFrame( confusion_matrix_value), annot=True, cmap=\"YlGnBu\" ,fmt='g')\n",
    "  ax.xaxis.set_label_position(\"top\")\n",
    "  plt.tight_layout()\n",
    "  plt.title('Confusion matrix', y=1.1)\n",
    "  plt.ylabel('Actual label')\n",
    "  plt.xlabel('Predicted label')\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# defined a function to print cross validation score\n",
    "scoring = {'recall' : make_scorer(recall_score)}\n",
    "def cross_validation_metrics(log_reg, X, y):\n",
    " log_reg_score = cross_val_score(log_reg, X,y,cv=5,scoring='recall')\n",
    " print('Logistic Regression Cross Validation Score(Recall): ', round(log_reg_score.mean() * 100, 2).astype(str) + '%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function for plotting feature importance\n",
    "def feature_importance(model, X):\n",
    "  importances = model.feature_importances_\n",
    "  std = np.std([tree.feature_importances_ for tree in model.estimators_],\n",
    "               axis=0)\n",
    "  indices = np.argsort(importances)[::-1]\n",
    "\n",
    "  # Print the feature ranking\n",
    "  print(\"Feature ranking:\")\n",
    "\n",
    "  for f in range(X.shape[1]):\n",
    "      print(\"%d. %s (%f)\" % (f + 1, X.columns[indices[f]], importances[indices[f]]))\n",
    "\n",
    "  # Plot the feature importances of the forest\n",
    "  plt.figure()\n",
    "  plt.title(\"Feature importances\")\n",
    "  plt.bar(range(X.shape[1]), importances[indices],\n",
    "          color=\"r\", yerr=std[indices], align=\"center\")\n",
    "  plt.xticks(range(X.shape[1]), indices)\n",
    "  plt.xlim([-1, X.shape[1]])\n",
    "  plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to draw ROC curve\n",
    "def plot_auc_curve(model,):\n",
    "  auc = roc_auc_score(y, y_pred_prob)\n",
    "  fpr, tpr, thresholds = roc_curve(y, y_pred_prob)\n",
    "  \n",
    "  plt.plot(fpr, tpr)\n",
    "  plt.xlim([0.0, 1.0])\n",
    "  plt.ylim([0.0, 1.0])\n",
    "  plt.title('ROC Curve\\n AUC={auc}'.format(auc = auc))\n",
    "  plt.xlabel('False Positive Rate')\n",
    "  plt.ylabel('True Positive Rate')\n",
    "  plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Reading data as a pandas dataframe\n",
    "data = pd.read_csv('creditcard.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Data Exploration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>21749</td>\n",
       "      <td>58670.0</td>\n",
       "      <td>-0.854092</td>\n",
       "      <td>0.644458</td>\n",
       "      <td>1.805656</td>\n",
       "      <td>1.146369</td>\n",
       "      <td>-0.519127</td>\n",
       "      <td>1.844676</td>\n",
       "      <td>-0.935942</td>\n",
       "      <td>1.056104</td>\n",
       "      <td>...</td>\n",
       "      <td>0.193673</td>\n",
       "      <td>0.789467</td>\n",
       "      <td>0.218834</td>\n",
       "      <td>-0.577043</td>\n",
       "      <td>-0.727521</td>\n",
       "      <td>0.612977</td>\n",
       "      <td>-0.219109</td>\n",
       "      <td>-0.063157</td>\n",
       "      <td>11.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>105607</td>\n",
       "      <td>164361.0</td>\n",
       "      <td>-0.863534</td>\n",
       "      <td>0.291699</td>\n",
       "      <td>0.594479</td>\n",
       "      <td>-1.190707</td>\n",
       "      <td>0.117851</td>\n",
       "      <td>0.169880</td>\n",
       "      <td>0.065587</td>\n",
       "      <td>0.289947</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.223345</td>\n",
       "      <td>-0.333300</td>\n",
       "      <td>-0.455269</td>\n",
       "      <td>0.185385</td>\n",
       "      <td>0.432974</td>\n",
       "      <td>0.931127</td>\n",
       "      <td>-0.414413</td>\n",
       "      <td>-0.284978</td>\n",
       "      <td>25.42</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>187884</td>\n",
       "      <td>38767.0</td>\n",
       "      <td>-1.192107</td>\n",
       "      <td>-0.896044</td>\n",
       "      <td>1.204410</td>\n",
       "      <td>-1.593935</td>\n",
       "      <td>0.432699</td>\n",
       "      <td>-1.101769</td>\n",
       "      <td>-0.299815</td>\n",
       "      <td>0.222793</td>\n",
       "      <td>...</td>\n",
       "      <td>0.470749</td>\n",
       "      <td>0.932440</td>\n",
       "      <td>0.159099</td>\n",
       "      <td>0.215700</td>\n",
       "      <td>-0.169315</td>\n",
       "      <td>-0.320951</td>\n",
       "      <td>0.310243</td>\n",
       "      <td>0.210535</td>\n",
       "      <td>58.75</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>238501</td>\n",
       "      <td>571.0</td>\n",
       "      <td>-2.355336</td>\n",
       "      <td>2.316182</td>\n",
       "      <td>0.701735</td>\n",
       "      <td>0.174501</td>\n",
       "      <td>0.677346</td>\n",
       "      <td>1.029705</td>\n",
       "      <td>0.792868</td>\n",
       "      <td>-0.060581</td>\n",
       "      <td>...</td>\n",
       "      <td>0.008872</td>\n",
       "      <td>0.955806</td>\n",
       "      <td>0.047292</td>\n",
       "      <td>-0.650140</td>\n",
       "      <td>-0.282282</td>\n",
       "      <td>-0.286391</td>\n",
       "      <td>0.335493</td>\n",
       "      <td>0.223061</td>\n",
       "      <td>2.89</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>252649</td>\n",
       "      <td>51507.0</td>\n",
       "      <td>-1.302336</td>\n",
       "      <td>1.016359</td>\n",
       "      <td>1.007046</td>\n",
       "      <td>-0.127051</td>\n",
       "      <td>0.435740</td>\n",
       "      <td>-0.092143</td>\n",
       "      <td>0.709650</td>\n",
       "      <td>0.590142</td>\n",
       "      <td>...</td>\n",
       "      <td>0.150091</td>\n",
       "      <td>0.059446</td>\n",
       "      <td>-0.262177</td>\n",
       "      <td>-0.354871</td>\n",
       "      <td>0.680078</td>\n",
       "      <td>-0.402172</td>\n",
       "      <td>-0.111834</td>\n",
       "      <td>-0.044427</td>\n",
       "      <td>51.59</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       id      Time        V1        V2        V3        V4        V5  \\\n",
       "0   21749   58670.0 -0.854092  0.644458  1.805656  1.146369 -0.519127   \n",
       "1  105607  164361.0 -0.863534  0.291699  0.594479 -1.190707  0.117851   \n",
       "2  187884   38767.0 -1.192107 -0.896044  1.204410 -1.593935  0.432699   \n",
       "3  238501     571.0 -2.355336  2.316182  0.701735  0.174501  0.677346   \n",
       "4  252649   51507.0 -1.302336  1.016359  1.007046 -0.127051  0.435740   \n",
       "\n",
       "         V6        V7        V8  ...       V21       V22       V23       V24  \\\n",
       "0  1.844676 -0.935942  1.056104  ...  0.193673  0.789467  0.218834 -0.577043   \n",
       "1  0.169880  0.065587  0.289947  ... -0.223345 -0.333300 -0.455269  0.185385   \n",
       "2 -1.101769 -0.299815  0.222793  ...  0.470749  0.932440  0.159099  0.215700   \n",
       "3  1.029705  0.792868 -0.060581  ...  0.008872  0.955806  0.047292 -0.650140   \n",
       "4 -0.092143  0.709650  0.590142  ...  0.150091  0.059446 -0.262177 -0.354871   \n",
       "\n",
       "        V25       V26       V27       V28  Amount  Class  \n",
       "0 -0.727521  0.612977 -0.219109 -0.063157   11.50      0  \n",
       "1  0.432974  0.931127 -0.414413 -0.284978   25.42      0  \n",
       "2 -0.169315 -0.320951  0.310243  0.210535   58.75      0  \n",
       "3 -0.282282 -0.286391  0.335493  0.223061    2.89      0  \n",
       "4  0.680078 -0.402172 -0.111834 -0.044427   51.59      0  \n",
       "\n",
       "[5 rows x 32 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#### Exploring data set\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "      <td>244807.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>142394.376301</td>\n",
       "      <td>94817.409077</td>\n",
       "      <td>0.002267</td>\n",
       "      <td>-0.002347</td>\n",
       "      <td>0.001627</td>\n",
       "      <td>0.000226</td>\n",
       "      <td>0.000116</td>\n",
       "      <td>-0.000695</td>\n",
       "      <td>0.000235</td>\n",
       "      <td>0.000451</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.000663</td>\n",
       "      <td>0.000085</td>\n",
       "      <td>0.000678</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.000082</td>\n",
       "      <td>-0.000326</td>\n",
       "      <td>0.000042</td>\n",
       "      <td>0.000369</td>\n",
       "      <td>88.221638</td>\n",
       "      <td>0.001724</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>82214.109334</td>\n",
       "      <td>47544.660749</td>\n",
       "      <td>1.958778</td>\n",
       "      <td>1.655048</td>\n",
       "      <td>1.511011</td>\n",
       "      <td>1.414850</td>\n",
       "      <td>1.378017</td>\n",
       "      <td>1.332263</td>\n",
       "      <td>1.229381</td>\n",
       "      <td>1.174459</td>\n",
       "      <td>...</td>\n",
       "      <td>0.725261</td>\n",
       "      <td>0.725420</td>\n",
       "      <td>0.622169</td>\n",
       "      <td>0.606201</td>\n",
       "      <td>0.520588</td>\n",
       "      <td>0.482409</td>\n",
       "      <td>0.404819</td>\n",
       "      <td>0.330222</td>\n",
       "      <td>250.663749</td>\n",
       "      <td>0.041483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-56.407510</td>\n",
       "      <td>-72.715728</td>\n",
       "      <td>-48.325589</td>\n",
       "      <td>-5.683171</td>\n",
       "      <td>-113.743307</td>\n",
       "      <td>-26.160506</td>\n",
       "      <td>-37.060311</td>\n",
       "      <td>-50.943369</td>\n",
       "      <td>...</td>\n",
       "      <td>-22.757540</td>\n",
       "      <td>-10.933144</td>\n",
       "      <td>-36.666000</td>\n",
       "      <td>-2.836627</td>\n",
       "      <td>-8.696627</td>\n",
       "      <td>-2.604551</td>\n",
       "      <td>-22.565679</td>\n",
       "      <td>-15.430084</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>71174.500000</td>\n",
       "      <td>54134.500000</td>\n",
       "      <td>-0.918719</td>\n",
       "      <td>-0.597789</td>\n",
       "      <td>-0.888899</td>\n",
       "      <td>-0.846298</td>\n",
       "      <td>-0.690880</td>\n",
       "      <td>-0.767694</td>\n",
       "      <td>-0.553464</td>\n",
       "      <td>-0.207827</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.228840</td>\n",
       "      <td>-0.543300</td>\n",
       "      <td>-0.161045</td>\n",
       "      <td>-0.354847</td>\n",
       "      <td>-0.316905</td>\n",
       "      <td>-0.327190</td>\n",
       "      <td>-0.070876</td>\n",
       "      <td>-0.052801</td>\n",
       "      <td>5.610000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>142272.000000</td>\n",
       "      <td>84747.000000</td>\n",
       "      <td>0.019709</td>\n",
       "      <td>0.065045</td>\n",
       "      <td>0.180200</td>\n",
       "      <td>-0.020248</td>\n",
       "      <td>-0.055758</td>\n",
       "      <td>-0.273607</td>\n",
       "      <td>0.039812</td>\n",
       "      <td>0.022329</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.029477</td>\n",
       "      <td>0.006637</td>\n",
       "      <td>-0.010750</td>\n",
       "      <td>0.041093</td>\n",
       "      <td>0.015969</td>\n",
       "      <td>-0.052227</td>\n",
       "      <td>0.001480</td>\n",
       "      <td>0.011233</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>213591.500000</td>\n",
       "      <td>139393.000000</td>\n",
       "      <td>1.315616</td>\n",
       "      <td>0.802233</td>\n",
       "      <td>1.027866</td>\n",
       "      <td>0.741737</td>\n",
       "      <td>0.610432</td>\n",
       "      <td>0.397342</td>\n",
       "      <td>0.569658</td>\n",
       "      <td>0.326222</td>\n",
       "      <td>...</td>\n",
       "      <td>0.186693</td>\n",
       "      <td>0.530219</td>\n",
       "      <td>0.147784</td>\n",
       "      <td>0.439776</td>\n",
       "      <td>0.350486</td>\n",
       "      <td>0.240038</td>\n",
       "      <td>0.091125</td>\n",
       "      <td>0.078232</td>\n",
       "      <td>77.050000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>284807.000000</td>\n",
       "      <td>172792.000000</td>\n",
       "      <td>2.454930</td>\n",
       "      <td>19.167239</td>\n",
       "      <td>9.382558</td>\n",
       "      <td>16.875344</td>\n",
       "      <td>34.801666</td>\n",
       "      <td>73.301626</td>\n",
       "      <td>120.589494</td>\n",
       "      <td>20.007208</td>\n",
       "      <td>...</td>\n",
       "      <td>27.202839</td>\n",
       "      <td>8.272233</td>\n",
       "      <td>22.528412</td>\n",
       "      <td>4.584549</td>\n",
       "      <td>7.519589</td>\n",
       "      <td>3.517346</td>\n",
       "      <td>31.612198</td>\n",
       "      <td>33.847808</td>\n",
       "      <td>25691.160000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  id           Time             V1             V2  \\\n",
       "count  244807.000000  244807.000000  244807.000000  244807.000000   \n",
       "mean   142394.376301   94817.409077       0.002267      -0.002347   \n",
       "std     82214.109334   47544.660749       1.958778       1.655048   \n",
       "min         1.000000       0.000000     -56.407510     -72.715728   \n",
       "25%     71174.500000   54134.500000      -0.918719      -0.597789   \n",
       "50%    142272.000000   84747.000000       0.019709       0.065045   \n",
       "75%    213591.500000  139393.000000       1.315616       0.802233   \n",
       "max    284807.000000  172792.000000       2.454930      19.167239   \n",
       "\n",
       "                  V3             V4             V5             V6  \\\n",
       "count  244807.000000  244807.000000  244807.000000  244807.000000   \n",
       "mean        0.001627       0.000226       0.000116      -0.000695   \n",
       "std         1.511011       1.414850       1.378017       1.332263   \n",
       "min       -48.325589      -5.683171    -113.743307     -26.160506   \n",
       "25%        -0.888899      -0.846298      -0.690880      -0.767694   \n",
       "50%         0.180200      -0.020248      -0.055758      -0.273607   \n",
       "75%         1.027866       0.741737       0.610432       0.397342   \n",
       "max         9.382558      16.875344      34.801666      73.301626   \n",
       "\n",
       "                  V7             V8  ...            V21            V22  \\\n",
       "count  244807.000000  244807.000000  ...  244807.000000  244807.000000   \n",
       "mean        0.000235       0.000451  ...      -0.000663       0.000085   \n",
       "std         1.229381       1.174459  ...       0.725261       0.725420   \n",
       "min       -37.060311     -50.943369  ...     -22.757540     -10.933144   \n",
       "25%        -0.553464      -0.207827  ...      -0.228840      -0.543300   \n",
       "50%         0.039812       0.022329  ...      -0.029477       0.006637   \n",
       "75%         0.569658       0.326222  ...       0.186693       0.530219   \n",
       "max       120.589494      20.007208  ...      27.202839       8.272233   \n",
       "\n",
       "                 V23            V24            V25            V26  \\\n",
       "count  244807.000000  244807.000000  244807.000000  244807.000000   \n",
       "mean        0.000678       0.000009       0.000082      -0.000326   \n",
       "std         0.622169       0.606201       0.520588       0.482409   \n",
       "min       -36.666000      -2.836627      -8.696627      -2.604551   \n",
       "25%        -0.161045      -0.354847      -0.316905      -0.327190   \n",
       "50%        -0.010750       0.041093       0.015969      -0.052227   \n",
       "75%         0.147784       0.439776       0.350486       0.240038   \n",
       "max        22.528412       4.584549       7.519589       3.517346   \n",
       "\n",
       "                 V27            V28         Amount          Class  \n",
       "count  244807.000000  244807.000000  244807.000000  244807.000000  \n",
       "mean        0.000042       0.000369      88.221638       0.001724  \n",
       "std         0.404819       0.330222     250.663749       0.041483  \n",
       "min       -22.565679     -15.430084       0.000000       0.000000  \n",
       "25%        -0.070876      -0.052801       5.610000       0.000000  \n",
       "50%         0.001480       0.011233      22.000000       0.000000  \n",
       "75%         0.091125       0.078232      77.050000       0.000000  \n",
       "max        31.612198      33.847808   25691.160000       1.000000  \n",
       "\n",
       "[8 rows x 32 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Observations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(244807, 32)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['id', 'Time', 'V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9',\n",
       "       'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19',\n",
       "       'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount',\n",
       "       'Class'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#### Checking for null values in dataset\n",
    "data.isnull().sum().max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "244807"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#### There are no null values in dataset \n",
    "####  Checking for unique values of ids\n",
    "data.id.nunique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Data is pretty clean and there are no duplicate ids are present now let's check distribution of each feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x1440 with 36 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the histograms of each \n",
    "data.hist(bins=50, figsize=(30,20))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### We can observe that all the features in dataset are scaled except amount and time.\n",
    "##### So, in next step I am going to scale Amount column in dataset and delete time column."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['normal_amount'] = StandardScaler().fit_transform(data['Amount'].values.reshape(-1,1))\n",
    "data = data.drop(['Amount','Time'], axis=1)\n",
    "X = data.loc[:,data.columns != 'Class']\n",
    "y = data.loc[:,data.columns == 'Class']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, (ax1, ax2) = plt.subplots(2, 1, sharex=True, figsize=(12,4))\n",
    "\n",
    "bins = 30\n",
    "\n",
    "ax1.hist(data.normal_amount[data.Class == 1], bins = bins)\n",
    "ax1.set_title('Fraud')\n",
    "\n",
    "ax2.hist(data.normal_amount[data.Class == 0], bins = bins)\n",
    "ax2.set_title('Normal')\n",
    "\n",
    "plt.xlabel('Amount ($)')\n",
    "plt.ylabel('Number of Transactions')\n",
    "plt.yscale('log')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### We can observe from the above figure that the fraud transactions amount is very less.\n",
    "##### In next step I am going to visualize number of fraud transactions and number of Non-fraud transactions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Frequency')"
      ]
     },
     "execution_count": 16,
     "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": [
    "# Now lets check the class distributions\n",
    "sns.countplot(\"Class\",data=data)\n",
    "plt.title(\"Fraud class histogram\")\n",
    "plt.xlabel(\"Class\")\n",
    "plt.ylabel(\"Frequency\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### As you can observe from the plot, we have so many 0 (non-fraud) compared to 1 (fraud). \n",
    "##### This kind of imbalance in the target variable is known as class imbalance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of normal transactions:  0.9982761930827141\n",
      "Percentage of fraud transactions:  0.0017238069172858619\n",
      "Total number of transactions in data:  244807\n"
     ]
    }
   ],
   "source": [
    "# Showing ratio\n",
    "print(\"Percentage of normal transactions: \", len(data[data.Class == 0])/len(data))\n",
    "print(\"Percentage of fraud transactions: \", len(data[data.Class == 1])/len(data))\n",
    "print(\"Total number of transactions in data: \", len(data))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Most of the transactions were Non-Fraud (99.83%) of the time,\n",
    "##### while Fraud transactions occurs 0.17% of the time in the dataframe."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Splitting original dataset into test and train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X,y,test_size = 0.3, random_state = 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1] Logistic regression on imbalanced dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression()\n",
    "lr.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  60.16%\n",
      "Recall: 0.6481481481481481\n",
      "Log Loss: 0.021162677654140833\n",
      "Precision: 0.9090909090909091\n",
      "Accurcay: 0.9993872799313753\n",
      "AUC:  0.8240263478860329\n",
      "F1 Score: 0.7567567567567568\n",
      "Confusion matrix:\n",
      " [[73328     7]\n",
      " [   38    70]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Accuracy metrics for \n",
    "y_pred = lr.predict(X_test)\n",
    "cross_validation_metrics(lr,X_train,y_train)\n",
    "print_accuracy_metrics(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Observatios\n",
    "By observing the accuracy we can conclude that algorithm is performing extremely well . But it’s not true. As most of the labels 0, even random guess gives you 99% accuracy. So we need a better measure to understand the performance of the model. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Recall\n",
    "Recall is a measure which measures the ability of model to predict right for a given label. In our case, we want to test the model how accurately it can recall fraud cases as we are interested in that.\n",
    "As you can observe from the results, the recall for 1.0 is only 0.6016 compared to 99% for 0. So our model is not doing a good job of recognising frauds. So this shows that how imbalanced data is effecting accuracy of model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2]  Using Class Weight (Logistic regression)\n",
    "Scikit-learn logistic regression has a option named class_weight when specified does class imbalance handling implicitly. So trying to predict using this technique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight='balanced', dual=False,\n",
       "          fit_intercept=True, intercept_scaling=1, max_iter=100,\n",
       "          multi_class='warn', n_jobs=None, penalty='l2', random_state=None,\n",
       "          solver='warn', tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_balanced = LogisticRegression(class_weight = 'balanced')\n",
    "lr_balanced.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  89.86%\n",
      "Recall: 0.9115646258503401\n",
      "Log Loss: 0.7882701074435376\n",
      "Precision: 0.06470304200869145\n",
      "Accurcay: 0.9771777676345634\n",
      "AUC:  0.9444277359227315\n",
      "F1 Score: 0.12082957619477007\n",
      "Confusion matrix:\n",
      " [[83359  1937]\n",
      " [   13   134]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_balanced_pred = lr_balanced.predict(X_test)\n",
    "cross_validation_metrics(lr_balanced,X_train,y_train)\n",
    "print_accuracy_metrics(y_test,y_balanced_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_balanced_pred_prob = lr_balanced.predict_proba(X_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prob: [0.11795312 0.10853555 0.16233188 0.04109433 0.06952336 0.8054823\n",
      " 0.33638875 0.01525269 0.01663033 0.45291477 0.03846819 0.02975751\n",
      " 0.00137332 0.00313013 0.00775784 0.05660248 0.01766819 0.00898342\n",
      " 0.09351198 0.05154421]\n"
     ]
    }
   ],
   "source": [
    "print('Prob:', y_balanced_pred_prob[0:20])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prob: [0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "print('Prob:', y_balanced_pred[0:20])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Undersampling of the dataset\n",
    "Undersampling is one of the techniques used for handling class imbalance.\n",
    "In this technique, we under sample majority class to match the minority class.\n",
    "So in our example, we take random sample of non-fraud class to match number of fraud samples. \n",
    "This makes sure that the training data has equal amount of fraud and non-fraud samples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "number_records_fraud = len(data[data.Class == 1])\n",
    "fraud_indices = np.array(data[data.Class == 1].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "normal_indices = data[data.Class == 0].index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "random_normal_indices = np.random.choice(normal_indices, number_records_fraud, replace = False)\n",
    "random_normal_indices = np.array(random_normal_indices)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "under_sample_indices = np.concatenate([fraud_indices,random_normal_indices])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "under_sample = data.iloc[under_sample_indices,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(844, 30)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "under_sample.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So there are total 844 observations in our undersample dataframe.\n",
    "##### Visualising Undersampled Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Frequency')"
      ]
     },
     "execution_count": 29,
     "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": [
    "# Now lets check the class distributions\n",
    "sns.countplot(\"Class\",data=under_sample)\n",
    "plt.title(\"Fraud class histogram\")\n",
    "plt.xlabel(\"Class\")\n",
    "plt.ylabel(\"Frequency\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### In the above plot, you can observe that classes are distributed evenly now."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we try to correlate class and features on imbalanced dataset then it will be of no use because we will not see true correlations of features with result. While now I am going to see the features and their correlations w.r.t class on undersampled dataframe."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cea74d0>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA90AAAOWCAYAAADyUfEpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3X285nVZL/rPNcPAIA8aoKiIiqa11QzahFGBElIKx9pqGhD7VLsaTzv3U1EZeTDb57W3FO2aylONAu4KGNMeNiVmT4Y9iDTttFICFUFR0XSQhxkeZ13njxk685rgvtfcs+7vrLV4v1+v9WKt+7duPteatWbNutb3+n2/1d0BAAAAlt6a/V0AAAAArFaabgAAAJgTTTcAAADMiaYbAAAA5kTTDQAAAHOi6QYAAIA50XQDAADAnGi6AQAAYE403QAAADAnB4wIeeALN/WInCR5yjPPHBWVT/75xmFZL3rpm4ZlvfddPzwsK0nu+9mfG5a1/sdeNyzrim/5n8Oyzr74mcOycsTjh0W98vuvHpaVJO98/VcMy/qjCz83LOv077p3WNatvzMu62mv/+phWVl/8LisJFkz7nfiP3X+R4ZlPaZrWNYp994/LOverB2WlYxdMXkg4z5nC+Oi8vg19w3Letozbh+WdcDBC8OykuRHP3rUsKxz7xnStvyzMz739oFfkSvDyJ5uVuuOesay/LxZ6QYAAIA50XQDAADAnGi6AQAAYE7G3hwBAADAyrOwY39XsGJZ6QYAAIA50XQDAADAnBgvBwAAYLIeeyTdamKlGwAAAOZE0w0AAABzYrwcAACAyRaMl8/KSjcAAADMiaYbAAAA5mTmpruqzphyfUNVbamqLW/9tStnjQEAAGA/615Y9i/L1b7c031Jkqc+0sXu3pRkU5I88IWbeh9yAAAAYEWa2HRX1VWPdCnJkUtfDgAAAKwe01a6T0lyXpK793i8kpw0l4oAAABYXuxePrNpTfe1SbZ39zV7XqiqG+ZTEgAAAKwO0zZSuynJ/Q93obtPXfpyAAAAYPWY1nTfmOTiqrq5qi6qquNHFAUAAACrwcTx8u7emGRjVT0tydlJLquq9UmuTLK5u28cUCMAAAD70zI+kmu5W9Q53d19S3df1N0nJDk3ycuTXD/XygAAAGCFW1TTXVXrquplVXV5kndn59j5K+daGQAAAKxw087pPiPJOUnOSnJdks1JNnT3tgG1AQAAsBws7NjfFaxY044MuyDJFUnO7+6tA+oBAACAVWPaRmqnjSoEAAAAVptpK90AAAA82tm9fGbV3XMPOfqxXzn/kF1u/fjVo6Jy+LHjBgG2Xvrdw7Ke8QPvGJaVJGc/7vnDst7y+Q8My3rfkeM+rnPv+8ywrLse2D4s62P/46xhWUly4uuuGZZ15cFPGJb12oGfs+/Jk4ZlvXnhlmFZOwb/oLFuzbjfif/eVwyLyv13j/u4PvfZw4ZlbX9w3bCsJDn26C8Ny/rEbV82LOspR9w5LOufbj90WNY/HLB+WNZTHhh7z+2LNozL+5lfG/v37CdvubyGBq4A99+8ZVhPN6sDn37isvy8LWr3cgAAAGDvGS8HAABgsgXj5bOy0g0AAABzoukGAACAOTFeDgAAwERt9/KZWekGAACAOdF0AwAAwJwYLwcAAGAyu5fPzEo3AAAAzImmGwAAAOZE0w0AAABz4p5uAAAAJnNk2MymrnRX1eFV9cyHefz58ykJAAAAVoeJTXdVvTrJPyb5rar6cFV97W6X3zbluRuqaktVbbnn/i/te6UAAACwwkwbL78gyb/u7s9W1UlJfr2qLuju305Sk57Y3ZuSbEqSox/7lb0k1QIAADDewo79XcGKNa3pPqC7P5sk3X1dVZ2W5Per6ilJNNIAAAAwwbR7uu/c/X7uXQ34i5J8W5LnzrEuAAAAWPGmrXTfnuTJST7+0APdfVdVvSTJq+dZGAAAAMuE3ctnNm2l+w+T/HRV3VxVF1XV8UnS3Q909+XzLw8AAABWrolNd3dv7O6Tk7wwydYkl1XV9VV1YVU9a0iFAAAAsEJNGy9PknT3LUkuSnJRVZ2Q5NIkb0iydo61AQAAsBwsGC+f1bTx8iRJVa2rqpdV1eVJ3p3kxiSvnGtlAAAAsMJNXOmuqjOSnJPkrCTXJdmcZEN3bxtQGwAAAKxo08bLL0hyRZLzu3vrgHoAAABYbuxePrOJTXd3nzaqEAAAAFhtFnVPNwAAALD3FrV7+b765J9vHBGTJDn82HGL83d+6r3Dso798rOGZd38J/99WFaS7HjXO4ZlvelVlw7L+sVvuWRY1t++/TXDsuqIJw/Les7pPzEsK0k+/I4fHJb1++eO+/7xnp8/cVjWp9/4V8Oyzv7ZVw3LGu7+e4dF/cR/+t/Dsh4z8Hf93/zgA8OytmbdsKwk2X7bkcOyDl8z7s/xI7cfMSzr8H5wWNapB4+7Q/OQJ943LCtJvvWycQcZ/eSDhw/LgqU2pOkGAABgBXNk2MyMlwMAAMCcaLoBAABgToyXAwAAMFH3jv1dwoplpRsAAADmRNMNAAAAc2K8HAAAgMna7uWzstINAAAAc6LpBgAAgDkxXg4AAMBkC8bLZ2WlGwAAAOZE0w0AAABzMnW8vKqemCTdfVtVPT7JKUlu6O4Pz7s4AAAAlgG7l89s4kp3Vb0myfuTXFtVP5Dk95P8H0l+u6q+d8pzN1TVlqra8tZ3Xr1kBQMAAMBKMW2l+7VJnpvk4CS3JPnyXSveX5bkvUkueaQndvemJJuS5L6/e08vTbkAAACwckxruh/s7u1JtlfVx7v7tiTp7turSiMNAADwaLCwY39XsGJN20htR1Wt2/X6WQ89WFXrF/FcAAAAeFSb1jj/XZIXJEl337rb40cm+eF5FQUAAACrwbSm+0NJfqaqbq6qi6rq+CTp7k939x/PvzwAAABYuSbe093dG5NsrKqnJTk7yWW7RsuvTLK5u28cUCMAAAD7kyPDZrao+7K7+5buvqi7T0hybpKXJ7l+rpUBAADACreopruq1lXVy6rq8iTvTnJjklfOtTIAAABY4SaOl1fVGUnOyc6dy69LsjnJhu7eNqA2AAAAloMF4+WzmnZO9wVJrkhyfndvHVAPAAAArBrTNlI7bVQhAAAAsNpMW+kGAADg0c7u5TOr7p57yMnHnDb/kF3+9OJTR0XluB/4zWFZn/rYu4ZlnXH8hmFZSfKktYcOy9qy/VPDst7zxCOHZZ10y83DsqpqWNanfv/1w7KS5KhvuXBY1vuPft6wrNO+eMOwrI2HnjQs6/X3/f2wrAPWjP0d9WEHHDws60+/6cBhWQ9+YcewrOs/9PhhWU943Nitbm6/c9zXx2cW1g/LOqzHfX088wm3D8v69D8dPizrc3XQsKwkOetXvnpY1v9+zXXDspLkG25757gfeFaIe99/5bCeblbrTz5nWX7eFrV7OQAAALD3NN0AAABMtrCw/F8WoapeUlU3VNXHqup1D3P9qVX13qr626r6u6o6c1//6DTdAAAArHpVtTbJm5O8NMlzkpxTVc/Z491en+Q3u/uEJGcn+X/3NVfTDQAAwKPBSUk+1t03dff9STYn+bY93qeTPLQZw2OTfGZfQ+1eDgAAwGSLHN/en6pqQ5Ldd4Xe1N2bdnv7mCS776x8a5IX7PG/+ckkf1hV/yHJIUlevK91aboBAABY8XY12JsmvMvD7W6+567s5yR5W3f/bFWdnOTXq+p53bOfmWa8HAAAgEeDW5Mcu9vbT8m/HB//3iS/mSTd/f4k65MctS+hmm4AAAAeDf46ybOq6riqOjA7N0q7ao/3+WSS05Okqv5Vdjbd/7QvocbLAQAAmKh7x/4uYZ9194NV9dok70myNsml3f3hqvqpJFu6+6okP5zkLVX1X7Jz9Py7u3vPEfS9oukGAADgUaG7r05y9R6PXbjb6x9J8g1LmWm8HAAAAObESjcAAACTrYAjw5arvV7prqr/No9CAAAAYLWZuNJdVb+w50NJ/m1VHZok3f0fJzz3nw8mP+6xz87Rhzx5H0sFAACAlWXaePkrkvxZkj/M/3+Q+NlJ/mba/3j3g8lPPua0fdrtDQAAgP2ojZfPatp4+XOSfCHJS5L8cXf/zyR3dff/3PU6AAAA8AgmrnR3951J/nNV/eskv1FV74odzwEAAGBRpt3T/UtJrujuv6qqb0ry75P8xZDKAAAAWB7sXj6zaavWH03ys1V1c5I3JfnL7j5v7lUBAADAKjCx6e7ujd19cpIXJtma5LKqur6qLqyqZw+pEAAAAFaoRd2f3d23dPdF3X1CknOTvDzJ9XOtDAAAgOWhF5b/yzK1qKa7qtZV1cuq6vIk705yY5JXzrUyAAAAWOGmbaR2RpJzkpyV5Lokm5Ns6O5tA2oDAACAFW1i053kgiRXJDm/u7cOqAcAAABWjWnndJ82qhAAAACWKUeGzWxR93QDAAAAe2/aePmSeO+7fnhETJLkuBf96LCsm//kvw/LOuP4DcOy/uiDm4ZlJcld3/s9w7Ie8/oLhmX9zLf/9rCsT/3yC4Zl5ehjhkW94juvHJaVJP/05lcNy3r76z8zLOvm1zx3WNZN77hjWNZHfuk7hmXlkMPHZQ32g6/5s2FZz+r1w7K+MfcOy/r7O44YlpUka7qHZd29dtz6zLod4z6uL249ZFjWcU8bd4fml68b92eYJK/7TzuGZX3tmnHfP5LkG4amsdoNaboBAABYwZbxkVzLnfFyAAAAmBNNNwAAAMyJ8XIAAAAms3v5zKx0AwAAwJxougEAAGBOjJcDAAAwmfHymVnpBgAAgDnRdAMAAMCcGC8HAABgsjZePisr3QAAADAnmm4AAACYk70aL6+q45KckOQj3f2P8ykJAACAZcXu5TObuNJdVb+72+vfluRPk7wsyf+qqu+e8twNVbWlqrZc8lt/sBS1AgAAwIoybaX7abu9/mNJvqm7P1FVRyX5kyRve6QndvemJJuS5N4P/n7vY50AAACw4ky7p3v3ZvmA7v5EknT3F5KYLwAAAIAJpq10P7+q7kxSSdZX1RO7+7aqOjDJ2vmXBwAAwH7nyLCZTVvp/tUkL+3uw7p7XXfftuvxxyR5zXxLAwAAgJVtWtN9Y5KLq+rmqrqoqo5Pku7+Une/f/7lAQAAwMo1cby8uzcm2VhVT0tydpLLqmp9kiuTbO7uGwfUCAAAwP7kyLCZTVvpTpJ09y3dfVF3n5Dk3CQvT3L9XCsDAACAFW5RTXdVrauql1XV5UnenZ1j56+ca2UAAACwwk0cL6+qM5Kck+SsJNcl2ZxkQ3dvG1AbAAAAy4Hdy2c27ciwC5JckeT87t46oB4AAABYNaZtpHbaqEIAAABgtZm20g0AAMCjnd3LZzak6b7vZ39uREyS5OzHPX9Y1o53vWNY1pPWHjos667v/Z5hWUly2CWXDcva/iMbhmW9vw8clrXjbz40LKsO/eiwrDWpYVlJsnDb54ZlPW7HuH+41jzpqGFZt9x797CsZ3xgy7CsJMlCj4v60ritU5678IRhWUfsGBaVtTXu83XowsAPLMnxX3HbsKy/v+HoYVmfX7tuWNa6Bw8alvWxTz1pWNZROx4clpUkdx78wLCs058x7t9oWGqL2r0cAB7VBjbcAMDqYrwcAACAyYyXz8xKNwAAAMyJphsAAADmRNMNAAAAc+KebgAAACZrm4rOyko3AAAAzImmGwAAAObEeDkAAACTOTJsZla6AQAAYE403QAAADAnxssBAACYzHj5zCY23VX11CSf7+57q6qSfHeSr0nykSRv6e4H518iAAAArEzTxsuv3u193pTkrCQfSPK1STZNemJVbaiqLVW15W0f/fQ+FwoAAAArzbTx8jXdvX3X6y9O8rXdvZDkN6rqQ5Oe2N2bsqsxv+Pfnu4kdQAAgJWqjZfPatpK96eq6pt2vX5zkmOTpKqOnGdRAAAAsBpMW+n+viS/VlU/meSOJB+sqr9N8mVJfmjOtQEAAMCKNq3p/rEkr09ye5JnJXlbkluT/PWuMXMAAABWO7uXz2zaePlHk1ycnRuqfX2Sj3f3BzTcAAAAMN3Epru7N3b3yUlemGRrksuq6vqqurCqnj2kQgAAAFihpq10J0m6+5buvqi7T0hybpKXJ7l+rpUBAACwPHQv/5dlalFNd1Wtq6qXVdXlSd6d5MYkr5xrZQAAALDCTdxIrarOSHJOkrOSXJdkc5IN3b1tQG0AAACwok3bvfyCJFckOb+7tw6oBwAAAFaNiU13d582qhAAAACWKUeGzWxR93QDAAAAe2/aePmSWP9jrxsRkyR5y29tGJb1plddOixryy+ePyzrBX+V/P3VFwzL2/4j4z5nj/mZTcOyzn/uuK/7tWe9eFjWmqOPG5b1W69OTnrxG4blrfmalw7LWpu/HJaVgw4aFvX8J39iWNYX/yh5ws+8elhe1g75J3NnVI9bTfj4H/3hsKzb14z7MzzunnHrCuuzI59eO+7v2Xs/dsywrCOzY1jW0xfuHZa1UMOicvyX3zYuLMkD96wdlnXX5w8ZlnXPHeuGZcFSG/evHyvGyIYbJhnZcLPyrNaGm5VnZMMNk4xsuHkUMl4+M+PlAAAAMCeabgAAAJgT83IAAABMNnC/kdXGSjcAAADMiaYbAAAA5sR4OQAAABP1Qu/vElYsK90AAAAwJ5puAAAAmBPj5QAAAEy2YPfyWVnpBgAAgDnRdAMAAMCcaLoBAABgTibe011V35rkD7v73kH1AAAAsNy0e7pnNW2l++1Jbq2qX6+qM6tq7WL/x1W1oaq2VNWWt77jXftWJQAAAKxA05ruf0zyrCTvS/LDST5TVb9SVS+c9j/u7k3dfWJ3n/h9rzprCUoFAACAlWXakWHd3bcneUuSt1TVE5O8Osmbquop3X3s3CsEAABg/1ro/V3BirVXG6l1923d/QvdfXKSb5xTTQAAALAqTGu611fV1z/che6+ZQ71AAAAwKoxbbz8V5P8bFU9KTs3Vbuyuz84/7IAAABYNhbsXj6riSvd3b1x1yj5C5NsTXJZVV1fVRdW1bOHVAgAAAAr1KLu6e7uW7r7ou4+Icm5SV6e5Pq5VgYAAAAr3LTx8iRJVa1L8pIkZyc5Pck1Sd44x7oAAABYLoyXz2xi011VZyQ5J8lZSa5LsjnJhu7eNqA2AAAAWNGmrXRfkOSKJOd399YB9QAAAMCqMbHp7u7TRhUCAADAMtW9vytYsRa1kRoAAACw96oH/MbismPOG/Zrka9ec9eoqPxpHTYs6xWP/fywrMvvfPywrCR5f98+LOv8+w4dlnXKh980LOs3vvrCYVnrBv6S84XHfHZcWJLfve1Jw7Je/cxbh2X95sefMizrmw/54rCsP9h+5LCsu2vsb/fXdw3LOvepnx6W9YmPHzEs68Y1Bw/L+pp1dwzLSpJfybiP7fsX7huW9YRjxv0M9/lPj/sZ7snPGPf1cf31Y3+Ge9z6cV8fd9530LCsJDn1tneM+0a8Qmz/+dcs+6Xux/znX12Wnzcr3QAAADAnizoyDAAAgEcxR4bNzEo3AAAAzImmGwAAAObEeDkAAACTLSz7fdSWLSvdAAAAMCeabgAAAJgT4+UAAABM1nYvn5WVbgAAAJgTTTcAAADMifFyAAAAJrN7+cysdAMAAMCcTG26q+rUqvqKXa9/Y1WdX1Vnzb80AAAAWNkmjpdX1c8nOSnJAVX1niSnJ3l3kv9SVS/q7h+Z8NwNSTYkyf/52JPyokOetXRVAwAAMEwv2L18VtPu6T4jyfOSHJzk00mO6e7tVfWmJH+b5BGb7u7elGRTklx2zHluAAAAAOBRZ9p4eXd3J3no1xoPNc8Li3guAAAAPKpNW+l+V1X9RZKDkrw1yW9W1bVJXpjkffMuDgAAgGXA7uUzm9Z0H5rkx5Pc390fqKpnJnl5djbg75x3cQAAALCSTWu6b0zy00meVFVvT3Jld188/7IAAABg5Zt4X3Z3b+zuk7NznHxrksuq6vqqurCqnj2kQgAAAFihFrUZWnff0t0XdfcJSc7NzhHz6+daGQAAAMtDLyz/l2VqUU13Va2rqpdV1eXZeU73jUleOdfKAAAAYIWbeE93VZ2R5JwkZyW5LsnmJBu6e9uA2gAAAGBFm7aR2gVJrkhyfndvHVAPAAAAy40jw2Y2senu7tNGFQIAAACrzaLu6QYAAAD23rTx8iVx9sXPHBGTJDnhP109LOtv3/6aYVnHfttPD8v61C+/YFhWkuz4mw8Ny1p71ouHZf3GV184LOu8D/3UsKwdn/yHYVnP/5ZxH1eSfOgd3zks691n//GwrO/75ecNy/r4D//FsKwNb/m6YVk5cP24rCTZduewqNe99sFhWYcdNO53/a/OHcOy3rvw2GFZSXLWfTuGZd1d64ZlfeGmxw/LuvWAcR9XfWLcSO4THzt226Uf3D7u+8d/WzOkbWGSheW7O/hyZ6UbAAAA5kTTDQAAAHNiTgMAAIDJ7F4+MyvdAAAAMCeabgAAAJgT4+UAAABM1nYvn5WVbgAAAJgTTTcAAADMiaYbAAAA5sQ93QAAAEzmyLCZWekGAACAOdF0AwAAwJxMHS+vqkOTvCTJsUkeTPLRJH/Ybc94AACAR4Ne0P7NauJKd1W9Osl7s7Ppfm2Sk5L82yQfrKqvmvLcDVW1paq2XPInf7NU9QIAAMCKMW2l+/VJvq67t1fVUUku7+5vqarnJ/nVJF//SE/s7k1JNiXJPVe+wV33AAAAPOpMa7oryT27Xt+W5AlJ0t1/V1WHz7MwAAAAlgm7l89sWtP9riR/UFXXJHlpknckSVUdkZ0NOQAAAPAIpjXdhyX5iyT3JXljd//xrse/lORr5lkYAAAArHTTmu4bk5yd5MlJNlfVF7r7g7t2Lr9v7tUBAACw/xkvn9nE3cu7e2N3n5zk1CRbk1xWVddX1YVV9ewhFQIAAMAKNbHpfkh339LdF3X3CUnOTfLyJNfPtTIAAABY4aaNlydJqmpddp7VfXaS05Nck+SNc6wLAACA5aIX9ncFK9bEpruqzkhyTpKzklyXZHOSDd29bUBtAAAAsKJNW+m+IMkVSc7v7q0D6gEAAIBVY2LT3d2njSoEAACAZcru5TNb1EZqAAAAwN5b1EZq++yIxw+JSZK7Htg+LKuOePK4rKphWTn6mHFZSerQjw7LWnP0ccOy1vX7hmXt+OQ/DMta+9TnDcvaet+dw7KSsV8fa4clJXncUcOiDn7MA8Oy6uinD8sa7sCDh0V9KeM+Z4fmoGFZd9934LCsNWN+mvpnCwN/JOiMCzuoxm3SdPSDDw7LWn/IuKwDDtgxLCtJ7lq4b1jWjj5sWBYsNSvdAAAAMCeDfzcLAADAStPu6Z6ZlW4AAACYE003AAAAzInxcgAAACYzXj4zK90AAAAwJ5puAAAAmBPj5QAAAEy2sLC/K1ixrHQDAADAnGi6AQAAYE403QAAAEy20Mv/ZRGq6iVVdUNVfayqXjfh/b69qrqqTtzXPzpNNwAAAKteVa1N8uYkL03ynCTnVNVzHub9DkvyH5N8YClyJzbdVXVAVb2mqv6gqv6uqj5UVe+uqv+rqtYtRQEAAAAwwElJPtbdN3X3/Uk2J/m2h3m//5rkp5PcuxSh01a6fz3J8Ul+MsmZSc5K8sYkX53kNyY9sao2VNWWqtpyydV/uQSlAgAAsF/s79HxRbzs3oPuetmwx0dxTJJP7fb2rbse+2dVdUKSY7v795fqj27akWFf091fscdjtya5tqpunPTE7t6UZFOS3POeX1rcgD0AAADMYPce9BHUwz3tny9WrUnyc0m+eynrmrbSfXtVvWpX+D8XUlXfkeT2pSwEAAAA5ujWJMfu9vZTknxmt7cPS/K8JH9WVTcn+bokV+3rZmrTmu6zk3x7ks9V1Y27VrdvS/KKXdcAAABgJfjrJM+qquOq6sDs7Gmveuhid9/R3Ud199O7++lJrk3yrd29ZV9CJ46Xd/fNSb4jSarqyCTV3V/Yl0AAAABWlu6Vf8dwdz9YVa9N8p4ka5Nc2t0frqqfSrKlu6+a/H+YzbR7uncv8Iu7v11VZ3T3Hy19SQAAALD0uvvqJFfv8diFj/C+L1qKzH05p/uSpSgAAAAAVquJK91V9UjL65XkyKUvBwAAgGVnYeWPl+8v08bLT0lyXpK793i8svNgcQAAAOARTGu6r02yvbuv2fNCVd0wn5IAAABgdZjWdN+U5P6Hu9Ddpy59OQAAACw7xstnNm0jtRuTXFxVN1fVRVV1/IiiAAAAYDWY2HR398buPjnJC5NsTXJZVV1fVRdW1bOHVAgAAAArVO3tIedVdUKSS5M8v7vXLuY5Zz71zGGzCL/1xueNispzf2TcMeUf+Z0fGpb1iu+8clhWkqxJDcv61P23D8v6/WMPHpZ1+s13DMvaet+dw7I+e9MfDMtKkoOffMqwrOuOPnFY1kmf2zIs63eOGHfn0XdtG/dxHbR23bCsJDnyoMOHZb3/O44alnXHB7YPy7r3rnGfs9vveMywrCS5f8eifvxaErfWQcOynrFm3NfHkUdsG5ZVa8aN5H7iti8blpUkJ7/5+cOy3vuD/zAsK0nO/NzmcT+grhB3fM+Ll/18+WMv++Nl+Xlb1DndVbWuql5WVZcneXd2jp2/cq6VAQAAwAo37ZzuM5Kck+SsJNcl2ZxkQ3eP+/UgAAAArFDTdi+/IMkVSc7v7q0D6gEAAGC5sXv5zCY23d192qhCAAAAYLVZ1D3dAAAAwN7TdAMAAMCcTLunGwAAgEe7hf1dwMplpRsAAADmRNMNAAAAc2K8HAAAgInakWEzs9INAAAAc6LpBgAAgDkxXg4AAMBkxstnNvNKd1VtmnJ9Q1Vtqaotn7z7k7PGAAAAwIo1caW7qo54pEtJzpz03O7elGRTkpz51DP9WgQAAIBHnWnj5f+U5JbsbLIf0rvefsK8igIAAGAZWdjfBaxc05rum5Kc3t3/Yj68qj41n5IAAABgdZh2T/fPJ/myR7j200tcCwAAAKwqE1e6u/vNE6794tKXAwAAwHLTdi+f2b7sXn7GUhYCAAAAq83MTXeSS5asCgAAAFiFph0ZdtUjXUrwY/rbAAAgAElEQVRy5NKXAwAAwLJj9/KZTdu9/JQk5yW5e4/HK8lJc6kIAAAAVolpTfe1SbZ39zV7XqiqG+ZTEgAAAKwO0+7pvinJ/Q93obtPXfpyAAAAYPWYttJ9Y5KLq+pJSd6e5Mru/uD8ywIAAGC5cGTY7Kad070xycaqelqSs5NcVlXrk1yZZHN337iYkHe+/iv2udDFOvF1/2ISfm4+/I4fHJZ11LdcOCzrn978qmFZSbJw2+eGZa35mpcOy3rr9107LOtD7/jOYVlrjj5uWNbBTz5lWFaS3POZPx+W9XvPe/2wrG1/+2vDsq4/c+OwrM/+2VuHZSVJHnzYwa+56Ds+Pyzru859x7Csr+8nD8t68gPjfji8r2pYVpIc3Q8My3ramnuGZT2wsC+H6uydv/riE4ZlPWHHg8OyDl877vtUkpz4g783LOvStccMy4Kltqjvbt19S3df1N0nJDk3ycuTXD/XygBguRjYcAMAq8u08fIkSVWtS/KS7FztPj3JNUneOMe6AAAAWC4cGTazaed0n5HknCRnJbkuyeYkG7p724DaAAAAYEWbttJ9QZIrkpzf3VsH1AMAAACrxrSN1E4bVQgAAADLUxsvn9m4bSIBAADgUUbTDQAAAHOyqN3LAQAAeBQzXj4zK90AAAAwJ5puAAAAmBPj5QAAAExk9/LZWekGAACAOdF0AwAAwJxMbLqram1Vvaaq/mtVfcMe114/5bkbqmpLVW259H1/vxS1AgAAwIoybaX7V5O8MMkXk/xCVf2P3a69YtITu3tTd5/Y3Sf+u1O/ah/LBAAAYL9ZWAEvy9S0pvuk7j63u38+yQuSHFpVv11VByWp+ZcHAAAAK9e0pvvAh17p7ge7e0OSDyb50ySHzrMwAAAAWOmmHRm2pape0t1/8NAD3f1TVfWZJL8839IAAABYDhwZNruJK93dfd7uDfduj7+1u9fNrywAAABY+WY+MqyqzljKQgAAAGC1mTZePsklSZ66VIUAAACwPBkvn93EpruqrnqkS0mOXPpyAAAAYPWYttJ9SpLzkty9x+OV5KS5VAQAAACrxLSm+9ok27v7mj0vVNUN8ykJAACA5cR4+eymbaR2U5L7H+5Cd5+69OUAAADA6jGt6b4xycVVdXNVXVRVx48oCgAAAFaDaed0b+zuk5O8MMnWJJdV1fVVdWFVPXtIhQAAAOxfXcv/ZZmq7t67J1SdkOTSJM/v7rWLec7/euK5exeyD45bf9eoqNxw3+HDsr5y/Z3Dsv7m/scOy0qSx+0Yd4PI2gz7UsxJX/nZYVl/+Y9PHpa1qL/0S+TYddsGpiU3P3DIsKyX/cP/Myzrfc/98WFZB9aOYVmfXLN+WNaO5fvv+D570ZPGfa+66VNHDMu6f+B3qyPW3TssK0m+8MC4r/0nHHjPsKxtD6wblnXgmnE/e+wY2AjcUuO+NpLk8IE/w60b+DNcknzz5zav4u/8s/nci1409pMwg6P/7M+W5edt2nh5kqSq1lXVy6rq8iTvzs6x81fOtTIAAABY4aad031GknOSnJXkuiSbk2zo7rHLTwAAAOw3di+f3bQjwy5IckWS87t764B6AAAAYNWY2HR392mjCgEAAIDVZlH3dAMAAAB7b9p4OQAAAI9yvbAsNwZfEax0AwAAwJxougEAAGBOjJcDAAAwkSPDZmelGwAAAOZE0w0AAABzYrwcAACAibrtXj6riSvdVfWYqvrRqvqRqlpfVd9dVVdV1U9X1aGjigQAAICVaNp4+duSHJ3kuCTvSnJikouTVJJfnvTEqtpQVVuqast7tn9sCUoFAACAlWXaePmzu/vVVVVJPpvkxd3dVfXnST406YndvSnJpiT5X088t5ekWgAAAIaze/nsFrWRWnd3kqt3/fehtzXSAAAAMMG0pnvLQ/dud/e/e+jBqnpmkrvmWRgAAACsdBPHy7v7+x7h8Y9X1SnzKQkAAIDlpBfsXj6rfTmn+8VLVgUAAACsQvvSdF+yZFUAAADAKjRxvLyqrnqkS0mOXPpyAAAAYPWYdmTYKUnOS3L3Ho9XkpPmUhEAAADLSju7ambTmu5rk2zv7mv2vFBVN8ynJAAAAFgdpt3TfVOS+x/uQnefuvTlAAAAwOoxbaX7xiQXV9WTkrw9yZXd/cH5lwUAAMBy4ciw2U1c6e7ujd19cpIXJtma5LKqur6qLqyqZw+pEAAAAFaoaSvdSZLuviXJRUkuqqoTklya5A1J1i7m+ad/170zF7i3zvyN7cOy3vPzJw7LOvb7Lx+WdfNrnjssK0nWPOmocWEHHTQs6pI3Leqvx5L4vl9+3rCsPG7c5+uwb3nDsKwk2fa3vzYs633P/fFhWad++L8Py7r2eT86LOsVm188LCv37Lmf6JzdduuwqB96w8PeRTYXJ647cFjWl/blUNS99HX3L+rHqSVzYBaGZR166H3Dsj55+yHDsj63ZtwXyAMDFwef+ODYna7+/YMfHpb1O485blgWLLVF/StRVeuSvCTJ2UlOT3JNkjfOsS4AAACWCePls5t2TvcZSc5JclaS65JsTrKhu7cNqA0AAABWtGkr3RckuSLJ+d29dUA9AAAAsGpMbLq7+7RRhQAAALA89dgtA1aVgduMAAAAwKOLphsAAADmZOwZFwAAAKw4di+fnZVuAAAAmBNNNwAAAMyJphsAAADmxD3dAAAATNTtnu5ZWekGAACAOdnrpruqbpxHIQAAALDaTGy6q+quqrpz18tdVXVXkmc+9PiU526oqi1VteXSD35iSYsGAABgnF5Y/i/L1bSV7rcl+d0kz+ruw7r7sCSf3PX64ZOe2N2buvvE7j7x3x1/3BKVCwAAACvHxKa7u/9Dko1Jrqyq/1hVa5L0kMoAAABghZu6e3l3/01VvTjJa5Nck2T93KsCAABg2Viwe/nMFnVkWHcvJPmFqnpHkhPmWxIAAACsDnu1e3l3f7a7r06SqjpjPiUBAADA6rCole5HcEmSpy5VIQAAACxPbbx8ZhOb7qq66pEuJTly6csBAACA1WPaSvcpSc5Lcvcej1eSk+ZSEQAAAKwS05rua5Ns7+5r9rxQVTfMpyQAAACWk14wXj6raRup3ZTk/oe70N2nLn05AAAAsHpMa7pvTHJxVd1cVRdV1fEjigIAAIDVYGLT3d0bu/vkJC9MsjXJZVV1fVVdWFXPHlIhAAAA+1X38n9Zrqr3srqqOiHJpUme391rF/Ocf3z2mcP+CN6/7YhRUTnlsf80LOsDX3r8sKyvOvCOYVlJcsu9hw7Lev6Tx33O7t2+bljWA/cv6q/ikjj4MQ8My/rIF8b9fU6SYw/aNizrrvsOHJY10tf9w08Py3r/835sWNYBtTAsK0nW1LifHJ769NuHZb335mOGZZ14+BeHZX34S182LCtJDusdw7IOWfPgsKy7F8b9u3lIjfu4nvT4O4dlbd829t+Wz2wb9zPcEevuHZaVJCfe+rtuYN7D9c8a19PN6l999Opl+XmbNl6eJKmqdVX1sqq6PMm7s3Ps/JVzrQwAAABWuGnndJ+R5JwkZyW5LsnmJBu6e9xyEAAAAKxQ044MuyDJFUnO7+6tA+oBAABgmXFk2OwmNt3dfdqoQgAAAGC1WdQ93QAAAMDemzZeDgAAwKPcQhsvn5WVbgAAAJgTTTcAAADMifFyAAAAJmrj5TOz0g0AAABzoukGAACAOTFeDgAAwETd+7uClWviSndVPX+319dV1eur6qqq+m9V9Zj5lwcAAAAr17Tx8rft9vqbknx5kp9NcnCSX5n0xKraUFVbqmrLb97xyX0qEgAAAFaiaePlu29Rd3qSr+3uB6rqfUk+NOmJ3b0pyaYk+cdnn2kYAQAAYIVasHv5zKY13Y+tqpdn54r4Qd39QJJ0d1eVRhoAAAAmmNZ0X5PkW3e9fm1VHd3dn6uqJyb5wnxLAwAAgJVtYtPd3d/zCI/flp3j5gAAAMAjmPnIsKo6o7v/aCmLAQAAYPlp93TPbNru5ZNcsmRVAAAAwCo0caW7qq56pEtJjlz6cgAAAGD1mDZefkqS85LcvcfjleSkuVQEAADAstLOrprZtKb72iTbu/uaPS9U1Q3zKQkAAABWh2n3dN+U5P6Hu9Ddpy59OQAAALB6TFvpvjHJxVX1pCRvT3Jld39w/mUBAACwXCzYvXxmE1e6u3tjd5+c5IVJtia5rKqur6oLq+rZQyoEAACAFap6L++Ir6oTklya5PndvXYxz7nn13582G33p/zYX4yKyp9f+qphWc85b9wJbR95y3cMy0qSHR/YMixr7Te/ZFjWpu9537CsDW/5umFZdfTTh2U9+ZT/MiwrST77128dlvXOb3rLsKxXbH7xsKy/fvW7hmWd/A8XDcvqe+4alpUkve1Lw7K+8Zv+72FZG9Y8dVjWSAcvjM07aseDYwMHOfKA+4Zl3bZj/bCsx/S4L5DDD3jYu0Ln5ody57CsN+34smFZSfINt73Tsu4etjzl3yz7rdROvPV3l+Xnbdp4eZKkqtYleUmSs5OcnuSaJG+cY10AAAAsE228fGbTzuk+I8k5Sc5Kcl2SzUk2dPe2AbUBAADAijZtpfuCJFckOb+7tw6oBwAAAFaNiU13d582qhAAAACWJ7uXz27aOd0AAADAjDTdAAAAMCeL2r0cAACAR69lf17YMmalGwAAAOZE0w0AAABzoukGAACAOXFPNwAAABM5Mmx2VroBAABgTjTdAAAAMCcTm+6qem1VHbXr9S+vqvdV1Zeq6gNV9VVTnruhqrZU1ZZL3vvBpawZAACAgbpr2b8sV9NWun+gu7+w6/WNSX6uux+X5MeS/MqkJ3b3pu4+sbtP/N7Tjl+CUgEAAGBlmdZ0777R2hO6+3eSpLv/LMlh8yoKAAAAVoNpTfc7q+ptVfWMJL9TVf+5qp5aVd+T5JMD6gMAAGA/W1gBL8vVxCPDuvsnquq7k1yZ5JlJDkqyIcnvJvnOuVcHAAAAK9jUc7q7+21J3jb3SgAAAGCVmfnIsKo6YykLAQAAYHnq1LJ/WYyqeklV3VBVH6uq1z3M9YOq6u27rn+gqp6+r392+3JO9yX7Gg4AAAAjVNXaJG9O8tIkz0lyTlU9Z493+94kt3f3lyf5uSQX7WvuxPHyqrrqkS4lOXJfwwEAAGCQk5J8rLtvSpKq2pzk25J8ZLf3+bYkP7nr9Xcm+aWqqu7uWUOn3dN9SpLzkty9x+O1q2AAAABWuYWZW85xqmpDdm78/ZBN3b1pt7ePSfKp3d6+NckL9vjf/PP7dPeDVXVHdi44f2HWuqY13dcm2d7d1+x5oapumDUUAAAAltKuBnvThHd5uBu/9/x1wmLeZ69Mu6f7piT3P9yF7j51X4IBAABgoFuTHLvb209J8plHep+qOiDJY5Ns3ZfQaU33jUkurqqbq+qiqjp+X8IAAABgP/nrJM+qquOq6sAkZyfZcx+zq5J8167Xvz3Jn+7L/dzJlPHy7t6YZGNVPW1XQZdV1fokVybZ3N03Lipl/cH7UuNe2dELw7JGOmDN1CPVl84hh4/LSsbeILJ23J/j3TXw4zpw/bisgQ5au25s4IMPO9gzFzsWd6rF0rhnz2055ueAGvc9uO+5a1hWktTBh40L2/HAsKgDa9z3xQcGft0fvmNc1sLIv89J1u7blONeuavWDss6bMe4rDUD/4leN/D74kHrHhyWlSQH7Rj3/WNh4N9pHt7CIo/kWs523aP92iTvSbI2yaXd/eGq+qkkW7r7quw8pevXq+pj2bnCffa+5i7qb0p335KdW6VfVFUnJLk0yRt2FQoAq9rQhhsAmJvuvjrJ1Xs8duFur9+b5FVLmbmoc7qral1VvayqLk/y7uwcO3/lUhYCAAAAq820c7rPSHJOkrOSXJdkc5IN3b1tQG0AAAAsA70Kxsv3l2nj5RckuSLJ+d29Tzu2AQAAwKPNtI3UThtVCAAAAKw2A7fEBgAAYCVanWdEjbGojdQAAACAvafpBgAAgDkxXg4AAMBEdi+fnZVuAAAAmBNNNwAAAMyJ8XIAAAAmsnv57Kx0AwAAwJxMbLqr6rer6ryqOnRv/8dVtaGqtlTVlkv+eMvsFQIAAMAKNW2l+wVJ/k2ST1bVb1bVy6vqwMX8j7t7U3ef2N0nfu+LT9znQgEAAGClmdZ0f767vz3J05L8XpLvT/Lpqrqsqr557tUBAACw3y2sgJflalrT3UnS3Xd1969395lJviLJB5K8bt7FAQAAwEo2rem+e88Huntrd/9Kd3/TnGoCAACAVWHikWHdfeqoQgAAAFieOrW/S1ixZj4yrKrOWMpCAAAAYLXZl3O6L1myKgAAAGAVmjheXlVXPdKlJEcufTkAAAAsNwumy2c2selOckqS8/IvN1SrJCfNpSIAAABYJaY13dcm2d7d1+x5oapumE9JAAAAsDpMa7pvSnL/w12wszkAAMCjw4Ldy2c2bSO1G5NcXFU3V9VFVXX8iKIAAABgNZh2TvfGJBur6mlJzk5yWVWtT3Jlks3dfeOiUtbsyybpe2fdmmmL90vo/nuHRR12wMHDskZb+NK2YVlre2FY1voe+NvAbXeOyzpw3NfiJ971EznxFb84LK/v+PywrKFuu3VY1JrqYVm97UtDs2r9IcPy6tAjhmWtr3H/bq4b9+WRtRkXNnr1Z+RZufetGZf1/7F373GSV/Wd/1/vmWlmQBgXkKtkQfGyGxQHGEzUVUAkS2SNIoqgJhGMk103+cXsotGJq7n9Ng6bbMSY1UW5JIRLLl6iKAZjDMYYxNHghdVFuSk4CMjADPeZ6c/+UTVJZzJUdTf1PVPV83rOox5T/f121ftT3dXfqlPnfM9Zumhzs6ypTe2e97tPbXfQaCc2b1nE4x7XLm/vB9u9J2jXmpBGb1bP36q6parWVNURwKuBk4FvdlqZpJ1eywa3NEjLBrckzVfLBrd2PjUBl3E1q0Z3kqkkL0lyMXAFvWHnp3RamSRJkiRJE27YOt0nAKcDJwHXAJcBq6qq3XhgSZIkSZIm1LATWlYDlwBnVdXdDeqRJEmSJI2ZdjMjLTzDJlI7rlUhkiRJkiQtNE4EKEmSJElSR2x0S5IkSZLUkYaLWkuSJEmSJtF0sqNLmFj2dEuSJEmS1BEb3ZIkSZIkdcTh5ZIkSZKkgWpHFzDB7OmWJEmSJKkjNrolSZIkSerIwEZ3kicnOT/JbyXZPckHknwjyZ8lOaRNiZIkSZKkHWl6Ai7jalhP94XAl4D7gKuBbwE/CXwKOH/QDZOsSrI2ydrzPv2lEZQqSZIkSdJkGdbo3qOq3ldV7wKWV9XvVtX3quo8YM9BN6yqc6tqZVWtfP0JR4+sYEmSJEmSJsWw2cunkzwNeDywW5KVVbU2yVOAxd2XJ0mSJEna0aazoyuYXMMa3W8BPk5viPzLgLcleRawHHhDx7VJkiRJkjTRBja6q+ozwNNnbPp8kicA66tqS6eVSZIkSZI04Yb1dP8LVXUXQJITqurToy9JkiRJkjROpnF8+Xw9lnW6zxtZFZIkSZIkLUADe7qTfOzRdgF7j74cSZIkSZIWjmHDy58PvJbeOt0zBXh2JxVJkiRJkrRADGt0Xw08UFVXbbsjyf/tpiRJkiRJ0jipHV3ABBt2TveNwCPb21FVLxh9OZIkSZIkLRzDGt3XA7+T5OYka5KsaFGUJEmSJEkLwbB1us8BzklyMHAacEGSZcClwGVVdX2DGiVJkiRJO9C0K4bNW6rmNjo/yRHA+cDhVbV4Nrd52yGvbnYKwJuedFurKNbcdECzrHc8745mWf/l7/ZqlgVw2PSyZlk3LNru2RKd+G8H3dks61237tcs6x42Nct670893CwL4Of+Ytg0F6PzP/bZ2Czrt+9qt9jEfzuw3fP+pTdvaZa1S9o9NwCWNcz71LXvb5Z13VFvapa1667tjlVTU+2eiwC337VHs6w9lrV73fy7zY9vlvVjaXcM3rzlsazQOzd3bl7aLAvgx1aua5Z147Vt358efdtHbGJu44+e+NqxP637Z27747H8vc3qKJBkKslLklwMXEFv2PkpnVYmSZIkSdKEG7ZO9wnA6cBJwDXAZcCqqrq/QW2SJEmSpDEwvaMLmGDDxq+tBi4BzqqquxvUI0mSJEnSgjFsIrXjWhUiSZIkSdJC03ZmGEmSJEnSxBn7WdTGWLvpFCVJkiRJ2snY6JYkSZIkqSMOL5ckSZIkDTQ9litgTwZ7uiVJkiRJ6oiNbkmSJEmSOuLwckmSJEnSQNM7uoAJZk+3JEmSJEkdGdjTnWQR8DrgFOAgYDPwbeD9VfU3XRcnSZIkSdIkG9bTfR7wr4HfBj4LfKK/7e1JfnHQDZOsSrI2ydprN35nJMVKkiRJkjRJhp3TfVRVndG//vkkV1fVO5J8DrgW+P1Hu2FVnQucC/C2Q15dI6lWkiRJktSc53TP37Ce7k1JDgVIciTwCEBVPQzYkJYkSZIkaYBhPd1vBj6b5CFgCjgNIMk+wOUd1yZJkiRJ0kQb2Oiuqr9OcjCwd1XdNWP7ncBbui5OkiRJkrTjVXZ0BZNr6JJh1XPXttuTnNBNSZIkSZIkLQyPZZ3u80ZWhSRJkiRJC9Cwdbo/9mi7gL1HX44kSZIkadw4e/n8DZtI7fnAa4H7ttke4NmdVCRJkiRJ0gIxrNF9NfBAVV217Y4k/7ebkiRJkiRJWhiGNbpvpL8297aq6gWjL0eSJEmSNG4cXj5/wyZSux74nSQ3J1mTZEWLoiRJkiRJWggGNrqr6pyqeg5wDHA3cEGSbyZ5R5KnNalQkiRJkqQJNWx4OQBVdQuwBliT5AjgfOCdwOLZ3H63hiupP3LfrB7SSOz2mFZcm5vNd21plvXUWtYsC2Cvdg+N9YvaPT9uumGvZll7LG33XNydpc2y7v3i+mZZAM+tA5tl3fi9qWZZK6d2aZb12Zuf2CxrVbunPZvavYwBMFXtsq476k3Nsg778rubZZ151FnNst7zzLubZQH88AvtBnne+cCuzbKemYeaZW2k3XHxjrTLOnBRu58hwD23tHvPePC/aft3pn+p4UvTgjOrtyxJppK8JMnFwBX0hp2f0mllkiRJkiRNuGHrdJ8AnA6cBFwDXAasqqr7G9QmSZIkSdJEGzbWdjVwCXBWVTmmQ5IkSZKkORjY6K6q41oVIkmSJEkaT9ON5zdZSBpOQyNJkiRJ0s7FRrckSZIkSR1pt36SJEmSJGkitVuscOGxp1uSJEmSpI7Y6JYkSZIkqSMOL5ckSZIkDeTw8vmzp1uSJEmSpI7Y6JYkSZIkqSMDh5cnWQK8HjgZOBAo4PvAXwDnVdWmziuUJEmSJO1QtaMLmGDDerovAlYAvwa8GDgJ+HXgWcAfD7phklVJ1iZZu/a+74ygVEmSJEmSJsuwidSOrKqnb7PtVuDqJNcPumFVnQucC/CbB7/GD0YkSZIkSTudYY3u9UleCXyoqqYBkiwCXgms77o4SZIkSdKON50dXcHkGja8/DTgFcAPklzf792+HXh5f58kSZIkSXoUA3u6q+pm4FUASfYGUlV3NahLkiRJkqSJN3TJsCTLkxxaVT+c2eBOcni3pUmSJEmSNNkGNrqTnAp8C/hQkuuSHD1j94VdFiZJkiRJGg/TE3AZV8N6ulcDR1XVCuAM4KIkL+/v81R6SZIkSZIGGDZ7+ZKqWgdQVdckOQ64PMlBuD66JEmSJEkDDevp3pDk0K1f9BvgxwIvBQ7rsC5JkiRJ0pioCbiMq2GN7vXAgTM3VNVG4ETgzK6KkiRJkiRpIRjW6L4SODvJzUnWJFkBUFWbquri7suTJEmSJGlyDVun+xzgnCQHA6cBFyRZBlwKXFpV325QoyRJkiRpB5oe6wHc423YRGoAVNUtwBpgTZIjgPOBdwKLZ3P75z/0yLwLnKsfrNujWdZPbN7ULOubX92nWda/46FmWQCL0+4P+EkPDl2afmSuX7Jrs6xTubdZ1n0P79Is66GNU82yAA7c1O65+MjsDp8jcU+7pz3/YfcfNsv63H17N8tavqVZFACLG76x2XX3dq9lZx51VrOs87/8O82yVhx2erMsgI/ts7RZ1p137tks68Dl9zXLuve+Zc2yXnTM95tlTT/UdtGku77V7r3OTd9sd8wH2Ldpmha6Wb0VSzKV5CVJLgauAK4HTum0MkmSJEmSJtzAnu4kJwCnAycB1wCXAauq6v4GtUmSJEmSxkDbcRQLy7Dh5auBS4CzquruBvVIkiRJkrRgDJtI7bhWhUiSJEmStNDMaiI1SZIkSdLOy7nL56/hnLaSJEmSJO1cbHRLkiRJktQRh5dLkiRJkgZy9vL5s6dbkiRJkqSO2OiWJEmSJKkjNrolSZIkSeqI53RLkiRJkgaazo6uYHLZ0y1JkiRJUkfm3ehOcu6Q/auSrE2y9uMP3jjfGEmSJEmSJtbA4eVJ9nq0XcCLB922qs4FzgX4m/1eWfOqTpIkSZK0w01jk26+hp3TfSdwC71G9lbV/3rfroqSJEmSJGkhGNbovhE4vqq+u+2OJN/rpiRJkiRJkhaGYY3udwN7Av+i0Q2cPfpyJEmSJEnjxsHl8zew0V1VfzBg3++PvhxJkiRJkhaOobOXJ1me5NDtbD+8m5IkSZIkSVoYBja6k5wKfAv4UJLrkhw9Y/eFXRYmSZIkSRoP0xNwGVfDerpXA0dV1QrgDOCiJC/v78uj30ySJEmSJA2bSG1JVa0DqKprkhwHXJ7kIDyXXpIkSZKkgYY1ujckObSqbgCoqnVJjgU+ChzWdXGSJEmSpB1v2j7XeRs2vHw9cODMDVW1ETgROLOroiRJkiRJWgiGNbqvBM5OcnOSNUlWAFTVpqq6uPvyJEmSJEmaXMPW6T4HOCfJwcBpwAVJlgGXApdW1bdnE/IQix9zobO1aHOzKO5mqlnWv/1X65tlff3evZplAew+vaVZ1g8XDzujYkj0LpIAACAASURBVHSOnLq3WdZnpx/fLGtRux8hP35v23koH067+SH3mnqoWdaPP9Lul3bdI3s2y9p16KKXozPdeOrQ6YZzlU5NtTsGv+eZdzfLWnHY6c2yrr3u0mZZAH972FubZe2z5OFmWX/5ULv3H8/a0u5x/fbf798sa89q954b4NhNDzTL+sH0smZZ0qjN6i1LVd1SVWuq6gjg1cDJ9JYSkyRJkiQtcDUBl3E1q0Z3kqkkL0lyMXAFcD1wSqeVSZIkSZI04QaOOUxyAnA6cBJwDXAZsKqq7m9QmyRJkiRJE23YiX6rgUuAs6qq3YlYkiRJkqSx0XamnYVl2ERqx7UqRJIkSZKkhabh3K+SJEmSJO1cGi7+I0mSJEmaRNNjPT/4eLOnW5IkSZKkjtjoliRJkiSpIw4vlyRJkiQN5ODy+bOnW5IkSZKkjtjoliRJkiSpIw4vlyRJkiQNNL2jC5hgA3u6kyxO8vNJfjPJ87bZ9/Yht12VZG2StZ988IZR1CpJkiRJ0kQZNrz8fwPHAD8E3pPkf87Y9/JBN6yqc6tqZVWtfPGuhz7GMiVJkiRJmjzDhpc/u6oOB0jyXuB/JfkwcDqQrouTJEmSJO145fzl8zasp3uXrVeqanNVrQKuBf4a2L3LwiRJkiRJmnTDGt1rk5w4c0NV/QZwAXBIV0VJkiRJkrQQDGx0V9Vrq+pT29n+waqa6q4sSZIkSZIm39Alw5IsB/apqhu22X54VX2ts8okSZIkSWPBJcPmb9iSYacC3wI+lOS6JEfP2H1hl4VJkiRJkjTphp3TvRo4qqpWAGcAFyXZulSYs5dLkiRJkjTAsOHlS6pqHUBVXZPkOODyJAeBc8ZLkiRJ0s5g2ubfvA3r6d6Q5NCtX/Qb4McCLwUO67AuSZIkSZIm3rBG93rgwJkbqmojcCJwZldFSZIkSZK0EAxrdF8JnJ3k5iRrkqwAqKpNVXVx9+VJkiRJkna0moDLuBp4TndVnQOck+Rg4DTggiTLgEuBS6vq27MJGdayH6Uf2e+eZlkP3L53s6z1G3ZtlrWo2j5lVzz99mZZn/3OE5tlvZ92v7OTHt7SLGu64RSKj7C4XRiwX21qlnXXlmXNsnZpuMjHHrR7Lu62pe3iJYsbvpxXw7lKb79rj2ZZP/xCu9/Zx/ZZ2izrbw97a7MsgOdf965mWX932K80yzp6+sFmWbtMtTtWHf5Iu+fi03J/syyAjVummmWt26Vli0IarVk9e6vqlqpaU1VHAK8GTqa3lJgkSQteywa3JElaWGbV6E4yleQlSS4GrgCuB07ptDJJkiRJ0liYpsb+8lgk2SvJp5N8u///ngO+d3mS25K8dzb3PbDRneSEJOcDtwKrgE8Ch1bVq6rqo3N5EJIkSZIkjam3Ap+pqqcCn+l//Wh+E7hqtnc8rKd7NfD3wL+tqpdU1cVV1fZkEUmSJEmSuvVS4A/71/8QeNn2vinJUcB+9CYdn5VhE6kdN9s7kiRJkiQtTG2nL52fJKvojdDe6tyqOneWN9+vqtYBVNW6JPtu5/4XAb8L/DRw/GzrGtjoliRJkiRpEvQb2I/ayE7yV8D+29n1q7OMeCPwyar6XjL7lUZsdEuSJEmSFryqetGj7UvygyQH9Hu5DwDu2M63PQd4fpI3ArsDuyS5r6oGrh1po1uSJEmStLP7GPCzwLv6///Ftt9QVa/Zej3J64CVwxrcMMslwyRJkiRJO6+agH+P0buAE5J8Gzih/zVJVib54GO5Y3u6JUmSJEk7tar6IduZHK2q1gI/t53tFwIXzua+7emWJEmSJKkj9nRLkiRJkgaahCXDxpU93ZIkSZIkdWRgozvJbknekuTNSZYleV2SjyU5O8nuQ267KsnaJGs/+eANo61akiRJkqQJMKyn+0JgP+BJwCeAlcDvAAHeN+iGVXVuVa2sqpUv3vXQEZQqSZIkSdoRdvTM5A1mL+/MsHO6n1ZVpyYJsA54UVVVkr8Fvtp9eZIkSZIkTa5ZndNdVQV8sv//1q/H96MESZIkSZLGwLCe7rVJdq+q+6rqzK0bkxwKbOy2NEmSJEnSOHD28vkb2Oiuqn+xCHh/+w1Jnt9NSZIkSZIkLQxDh5cnWd7v2d7WMzuoR5IkSZKkBWNgT3eSU4F3A3ckmQJeV1Vf6u++EDiy2/IkSZIkSTvadDml13wN6+leDRxVVSuAM4CLkry8vy+dViZJkiRJ0oQbNpHakqpaB1BV1yQ5Drg8yUE4e7kkSZIkSQMN6+neMPN87n4D/FjgpcBhHdYlSZIkSdLEG9bTvR44ELhh64aq2pjkRODULguTJEmSJI0HhznP37Ce7iuBs5PcnGRNkhUAVbWpqi7uvjxJkiRJkibXwEZ3VZ1TVc8BjgHuBi5I8s0k70jy1CYVSpIkSZI0oVJznPo9yRHA+cDhVbV4Nrf5xH6nNxuNsCtbWkWxbFG7rNtrabOs+xYPXb59pA7YvKlZVjWcdH+fpQ82y7rvkalmWS1/huuyS7MsgIPT7ne2y+J2x4/dd3+4WdZd9zyuWdaG6XbP+wfS9rj48KJ2f2fPWLqhWdadD+zaLOuhoYP5RmefJe3+xgDu3zzs7MDRed51a5plffIZb2+WtajhQNmbp2b1dnkkfvSRdu+pAO5Nu8f2I4vbvUYDHH3bR1ypaRuvPvjksR9hfskt4/l7m9UrUpKpJC9JcjFwBXA9cEqnlUmSJEmSNOEGflSa5ATgdOAk4BrgMmBVVd3foDZJkiRJkibasPFJq4FLgLOq6u4G9UiSJEmSxkw5f/m8DWx0V9VxrQqRJEmSJGmhaTszjCRJkiRJO5F2019KkiRJkibS9I4uYILZ0y1JkiRJUkdsdEuSJEmS1BGHl0uSJEmSBpp29vJ5s6dbkiRJkqSO2OiWJEmSJKkjDi+XJEmSJA1UDi+ftzn3dCe5votCJEmSJElaaAY2upNsTLKhf9mYZCNw6NbtQ267KsnaJGs/9eB3Rlq0JEmSJEmTYFhP94XAR4GnVtUeVbUH8N3+9eWDblhV51bVyqpaeeKuTxlRuZIkSZIkTY6B53RX1S8mOQq4NMlHgfeCg/klSZIkaWcyvaMLmGBDz+muqi8DL+p/eRWwrNOKJEmSJElaIGY1e3lVTQPvSfJnwBHdliRJkiRJ0sIwtNGdZDmwT1XdUFXrgHX97YdX1de6LlCSJEmStGNVeZbxfA2bvfxU4FvAh5Jcl+ToGbsv7LIwSZIkSZIm3bBzulcDR1XVCuAM4KIkL+/vS6eVSZIkSZI04YYNL1/SH1JOVV2T5Djg8iQH4SzmkiRJkrRTmLb5N2/Dero3JDl06xf9BvixwEuBwzqsS5IkSZKkiTes0b0eOHDmhqraCJwInNlVUZIkSZIkLQTDhpdfCZyd5ADgT4BLq+raqtoEXNx5dZIkSZKkHW56RxcwwQb2dFfVOVX1HOAY4G7ggiTfTPKOJE9tUqEkSZIkSRNq6DrdAFV1C7AGWJPkCOB84J3A4tncfrrhPOcH7bmhWdb/Wb9Xs6w9akuzrKktbSdJuGPxVLOsQ6Yfapa17xM3Nsu668Z9mmUtTbvPOZ+86IFmWQCbpoedcTM6929q97z/7vrHNcvardo9P/Ze8nCzrD22zOrlbmSWLtrcLOvvNj++WdYz0+4YfODy+5pl/eVD7d4PABw9/WCzrE8+4+3Nsl78jd9qlvWNI3+5WdaK/do9Fzc91PZYdftdezTL+gbtXssAjh7+LdKszarRnWSK3nncpwHHA1cBv95hXZIkSZKkMVHOXj5vAxvdSU4ATgdOAq4BLgNWVdX9DWqTJEmSJGmiDevpXg1cApxVVXc3qEeSJEmSpAVjYKO7qo5rVYgkSZIkSQvNrM7pliRJkiTtvKY9p3ve2k3VK0mSJEnSTsZGtyRJkiRJHXF4uSRJkiRpoCqHl8+XPd2SJEmSJHXERrckSZIkSR1xeLkkSZIkaaDpHV3ABLOnW5IkSZKkjgxsdCc5fMb1qSRvT/KxJP89yW7dlydJkiRJ0uQa1tN94Yzr7wKeAvwusCvw/kE3TLIqydokaz/1wHceU5GSJEmSpB2nJuDfuBp2TndmXD8eOLqqNiX5HPDVQTesqnOBcwE+vv/p4/sTkCRJkiSpI8Ma3Y9PcjK9HvGlVbUJoKoqiQ1pSZIkSZIGGNbovgr4qf71q5PsV1U/SLI/cFe3pUmSJEmSxsH0GA/fHncDG91VdcajbL+d3nBzSZIkSZL0KIYuGZZkeZJDt7P98O19vyRJkiRJ6hnY053kVODdwB1JpoDXVdWX+rsvBI7stjxJkiRJ0o5W5fDy+RrW070aOKqqVgBnABcleXl/Xx79ZpIkSZIkadhEakuqah1AVV2T5Djg8iQHgWfSS5IkSZI0yLCe7g0zz+fuN8CPBV4KHNZhXZIkSZIkTbxhPd3rgQOBG7ZuqKqNSU4ETu2yMEmSJEnSeHDJsPkb1tN9JXB2kpuTrEmyAqCqNlXVxd2XJ0mSJEnS5BrY6K6qc6rqOcAxwN3ABUm+meQdSZ7apEJJkiRJkibUsOHlAFTVLcAaYE2SI4DzgXcCi2dz+30WPTzvAufqzvW7N8taXpubZR2y7z3Nsn549+OaZQFMbV7aLGu64Zz7d9y2R7OsW5dMNcvab3O75/1Be93bLAvgCz/ct1nWU/NAs6wfLBo2qGl0nrH5kWZZt29Z1ixrUeMRdVObZvXyPBI/tnRjs6yN7NIs69772j0/nrWl3fscgF2mtjTLWrS53ZP/G0f+crOsZ3zl95plrTvxDc2y7vhhu/fBAMuWtHtPsO8j7Z732r5yePm8zeqdWJKpJC9JcjFwBXA9cEqnlUmSJEmSNOEGfpSe5ATgdOAk4BrgMmBVVd3foDZJkiRJkibasPFrq4FLgLOq6u4G9UiSJEmSxsx0Obx8vgY2uqvquFaFSJIkSZK00LSbXUeSJEmSpJ1Mu+lRJUmSJEkTycHl82dPtyRJkiRJHbHRLUmSJElSRxxeLkmSJEkaaNoB5vNmT7ckSZIkSR2x0S1JkiRJUkdsdEuSJEmS1JGBje4kv5DkCf3rT0nyuST3JPlikme2KVGSJEmStCNNU2N/GVfDerr/U1Xd1b9+DvB7VfWvgF8B3j/ohklWJVmbZO1HH7hpBKVKkiRJkjRZhjW6Z85uvm9VfQSgqv4G2GPQDavq3KpaWVUrX7bbkx5blZIkSZIkTaBhS4b9eZILgd8APpLkTcCHgeOB73ZcmyRJkiRpDFSN7/DtcTew0V1Vv5rkdcClwKHAUmAV8FHgNZ1XJ0mSJEnSBBvW001VXQhc2HklkiRJkiQtMEMb3UmWA/tU1Q3bbD+8qr7WWWWSJEmSpLEwzrODj7thS4adCnwL+FCS65IcPWP3hV0WJkmSJEnSpBs2e/lq4KiqWgGcAVyU5OX9fem0MkmSJEmSJtyw4eVLqmodQFVdk+Q44PIkB4HjCyRJkiRpZ1A2/+ZtWE/3hiSHbv2i3wA/FngpcFiHdUmSJEmSNPGGNbrXAwfO3FBVG4ETgTO7KkqSJEmSpIVg2PDyK4GzkxwA/AlwaVVdW1WbgIs7r06SJEmStMNVObx8vgb2dFfVOVX1HOAY4G7ggiTfTPKOJE9tUqEkSZIkSRNq6DrdAFV1C7AGWJPkCOB84J3A4tnc/uAnr593gXP1sZuf2CzrBbve3SzrtjuXN8t60sHtHhfAd753QLOsFU+5vVlWhp28Mcqsm9p98rjscZubZWVR209U993S7rFtWdRuAYhNDdeaOGCfDc2yHrpjz2ZZU5lulgWw+9QjzbI2b2l3sLojuzTLetEx32+W9dt/v3+zLIDDH1naLOvuqWZRrNjvvmZZ6058Q7OsAz71gWZZy//z65tlAXzvmt2bZe1XDzbLkkZtVq+0SaaSvCTJxcAVwPXAKZ1WJkmSJEnShBvY053kBOB04CTgGuAyYFVV3d+gNkmSJEnSGJh2ybB5Gza8fDVwCXBWVbUdcyxJkiRJ0oQb2OiuquNaFSJJkiRJ0kIzq4nUJEmSJEk7L5cMm7+G8ytLkiRJkrRzsdEtSZIkSVJHHF4uSZIkSRrI2cvnz55uSZIkSZI6YqNbkiRJkqSOOLxckiRJkjRQObx83uzpliRJkiSpIza6JUmSJEnqyMBGd5IPJ3ltkt3nesdJViVZm2TtH9/+/flXKEmSJEnaoaarxv4yrob1dP8Y8DLgu0n+NMnJSXaZzR1X1blVtbKqVr52/wMfc6GSJEmSJE2aYY3uO6rqFcDBwMeBNwC3JbkgyU90Xp0kSZIkSRNsWKO7AKpqY1VdVFUvBp4OfBF4a9fFSZIkSZJ2vJqAf+NqWKP7vm03VNXdVfX+qnphRzVJkiRJkrQgDGx0V9ULWhUiSZIkSdJCM3TJsCTLkxy6ne2Hd1OSJEmSJEkLw5JBO5OcCrwbuCPJFPC6qvpSf/eFwJHdlidJkiRJ2tHGeUmucTesp3s1cFRVrQDOAC5K8vL+vnRamSRJkiRJE25gTzewpKrWAVTVNUmOAy5PchCM8fRwkiRJkiSNgWGN7g1JDq2qGwCqal2SY4GPAod1XZwkSZIkaccb5yW5xt2w4eXrgQNnbqiqjcCJwJldFSVJkiRJ0kIwrNF9JXB2kpuTrEmyAqCqNlXVxd2XJ0mSJEnS5Bo4vLyqzgHOSXIwcBpwQZJlwKXApVX17VmF7Dr9mAudrYM2bWmW9bj9H26Wdd19ezbLespU26EjT9iyuVnWpgcXN8u64bt7Ncva//H3N8tasqTd39hNt7d73gMsX/xIs6yb2LVZ1v6b2/1NP3D/Ls2yli9p9/sCWDrV7li1dFm7rBvWP75Z1oGLHmqWNf1Qu/cee1a71xaAp6XdMX/DI+3+pjc91O7neMcPd2+Wtfw/v75Z1uP+4LxmWQCbjvjlZll3bFrWLEvb5+zl8zd0nW6AqrqlqtZU1RHAq4GTgW91WpkkSWOiZYNbkiQtLLNqdCeZSvKSJBcDVwDXA6d0WpkkSZIkSRNu4PDyJCcApwMnAdcAlwGrqqrduCZJkiRJ0g7l7OXzN2zJsNXAJcBZVXV3g3okSZIkSVowhk2kdlyrQiRJkiRJWmiG9XRLkiRJknZyzl4+f7OaSE2SJEmSJM2djW5JkiRJkjpio1uSJEmSpI54TrckSZIkaSCXDJs/e7olSZIkSeqIjW5JkiRJkjri8HJJkiRJ0kBV0zu6hIk1sKc7yZOTnJ/kt5LsnuQDSb6R5M+SHDLktquSrE2y9o9uXTfKmiVJkiRJmgjDhpdfCHwJuA+4GvgW8JPAp4DzB92wqs6tqpVVtfJnDjpgBKVKkiRJkjRZhg0v36Oq3geQ5I1V9bv97ecl+YVuS5MkSZIkjYNpZy+ft2E93dNJnpbkaGC3JCsBkjwFWNx5dZIkSZIkTbBhPd1vAT4OTAMvA96W5FnAcuANHdcmSZIkSdJEG9jorqrPAE+fsenzSZ4ArK+qLZ1WJkmSJEkaC1UOL5+voet0J1me5NCtX1fVXVW1Jcnh3ZYmSZIkSdJkG7Zk2Kn0Ziz/UJLr+ud2b3Vhl4VJkiRJkjTphp3TvRo4qqrWJXk2cFGS1VX1YSDdlydJkiRJ2tGcvXz+hjW6l1TVOoCquibJccDlSQ4Cf+qSJEmSJA0y7JzuDducz70OOBZ4KXBYh3VJkiRJkjTxhvV0rwcOBG7YuqGqNiY5ETi1y8IkSZIkSePB2cvnb1hP95XA2UluTrImyQqAqtpUVRd3X54kSZIkSZMrs/nEIsnBwGn9yzLgUuDSqvr2bELOOOSUZh+LvPenF7eK4qcuuKdZ1uV/8KJmWW/9pa80ywLYwKZmWRurXdbbmyXBf93yULOsjdMPN8u66j0nNMsCWPmfP94s62ye3CzrjZuva5b1gSU/2izrN5fc2SwLYOmiYYPDRmfvRbs2y/pfz1jfLOueW5Y1y9qypd18r3ffu1uzLIAHtrR7Lt6Tdln7p93ry7Ilm5tlTS2ZbpYFsGnz0BWBR+aZ//B7zbK+cvhZzbIAfuz7H3bS6G08cc/Dxr6r+7b1143l721Wf5VVdUtVramqI4BXAyfTW0pMkqQFr2WDW5Lmq2WDW9LszepdRJIp4ER6Pd3HA1cBv95hXZIkSZKkMTHtOd3zNrDRneQE4HTgJOAa4DJgVVXd36A2SZIkSZIm2rCe7tXAJcBZVXV3g3okSZIkSVowBja6q+q4VoVIkiRJksZT4fDy+XK2BUmSJEmSOmKjW5IkSZKkjrgGiiRJkiRpoHL28nmzp1uSJEmSpI7Y6JYkSZIkqSMOL5ckSZIkDTS9wGcvT7IX8CfAIcDNwKlVtX4733c2cBK9DuxPA79UQ8be29MtSZIkSdrZvRX4TFU9FfhM/+t/JslzgecBhwPPAI4Gjhl2xza6JUmSJEk7u5cCf9i//ofAy7bzPQUsA3YBlgJTwA+G3fGsG91JfinJ8vScl+QrSX5itreXJEmSJE2mqhr7S5JVSdbOuKyaw0Pcr6rW9R/rOmDf7fwM/h74LLCuf/nLqvrmsDueyzndZ1bVOUn+PbAPcAZwAXDl9r65/wBXATxnryN4+h5PmkOUJEmSJEmzV1XnAuc+2v4kfwXsv51dvzqb+0/yFODfAgf1N306yQuq6nODbjeXRnf6/78YuKCqvpokj/bNMx/wGYecsrDPupckSZIkjbWqetGj7UvygyQHVNW6JAcAd2zn204Grq6q+/q3uQL4cWBgo3su53R/OcmV9Brdf5lkD2B6DreXJEmSJGkcfQz42f71nwX+Yjvf813gmCRLkkzRm0RtpMPLXw+sAG6sqgf6U6qfMYfbS5IkSZIm0PTgVbEWgncBf5rk9fQa168ESLIS+I9V9XPAnwMvBL5Ob1K1T1XVx4fd8Vwa3c8Brq2q+5O8FjgSOGdOD0OSJEmSpDFTVT8Ejt/O9rXAz/WvbwF+fq73PZfh5e8DHkjyLOAtwC3AH801UJIkSZKkncVcero3V1UleSlwTlWdl+Rnh95KkiRJkjTRauEPL+/MXBrdG5O8DXgt8IIki+ktBi5JkiRJkrZjLsPLXwU8DLy+qm4Hngj8j06qkiRJkiRpAZh1T3e/of0/Z3z9XTynW5IkSZIWvGkcXj5fs+7pTvLjSb6U5L4kjyTZkuTeLouTJEmSJGmSzWV4+XuB04FvA7vSmzb9D7ooSpIkSZKkhWAuE6lRVd9Jsri/PtkFSb7QUV2SJEmSpDHh7OXzl9n+8JJ8DngR8EHgdmAd8Lqqetaw2356v1c1+w393bI5fY7wmBz/8MPNsuYyJOGxumnRsoZpcPyTb2uW9eC97Sbcv/XOxzfLWrpoS7OsLZVmWfdMt10gYe/FjzTLunfLLs2ynrDLg82yphs+Px7e3O54P90sqaflMX+Xxe2OHwf/m7ubZd30zb2bZd0+3fZ1c91Uu2fIUdP3N8v6Bo9rlrXv5nbP+/2m2h2D79jU9rnY8nXzyK/9TrMsgKknPLndC9qEWP64J499q3vD/TeO5e9tLkftnwYWA78A3A/8CHBKF0VJkiRJkrQQzGX28lv6Vx8Efr2bciRJkiRJ42ba4eXzNrTRneTr8Ojzw1fV4SOtSJIkSZKkBWI2Pd0vB/YDvrfN9oOB74+8IkmSJEmSFojZnNP9e8CGqrpl5gV4oL9PkiRJkiRtx2x6ug+pqq9tu7Gq1iY5ZOQVSZIkSZLGSj36GccaYjY93YPWHth1VIVIkiRJkrTQzKbR/aUkb9h2Y5LXA18efUmSJEmSJC0Msxle/ibgI0lewz81slcCuwAnd1WYJEmSJGk8uGTY/A1tdFfVD4DnJjkOeEZ/8yeq6q87rUySJEmSpAk3m55uAKrqs8BnO6xFkiRJkqQFZTbndAOQ5Hmz2SZJkiRJWliqauwv42rWjW7g92e5TZIkSZIkMYvh5UmeAzwX2CfJf5mxazmweMDtVgGrAH5pj6M4addDH2OpkiRJkiRNltmc070LsHv/e/eYsX0D8IpHu1FVnQucC/Dp/V41vn39kiRJkqSBCpt08zWb2cuvAq5KcmFV3dKgJkmSJEmSFoRZz14OLE1yLnDIzNtV1QtHXZQkSZIkSQvBXBrdfwa8H/ggsKWbciRJkiRJ42acZwcfd3NpdG+uqvd1VokkSZIkSQvMXJYM+3iSNyY5IMleWy+dVSZJkiRJ0oSbS0/3z/b/f/OMbQU8eXTlSJIkSZLGjcPL52/Wje6qelKXhUiSJEmStNDMutGd5Ge2t72q/mh05UiSJEmStHDMZXj50TOuLwOOB74C2OiWJEmSJGk75jK8/Bdnfp3k8cBFI69IkiRJkjRWPKN7/uYye/m2HgCeOqpCJEmSJElacKpqVhfg48DH+pdPADcC75rt7edzAVZ1ef9mTW7WQn5sZk1W1kJ+bGZNXp5ZZo1LnlmTlbWQH1vrn6MXL9u7pGp2AwWSHDPjy83ALVV166xuPE9J1lbVyi4zzJrMrNZ5Zpk1LnlmTVZW6zyzzBqXPLMmK6t13kLNkh7NrIeXV9VVwLeAPYA9gUe6KkqSJEmSpIVg1o3uJKcC1wCvBE4FvpjkFV0VJkmSJEnSpJvLkmG/ChxdVXcAJNkH+Cvgz7sorO/cDu/brMnOap1nllnjkmfWZGW1zjPLrHHJM2uyslrnLdQsabvmck7316vqmTO+XgR8deY2SZIkSZL0T+bS0/2pJH8JXNr/+lXAFaMvSZIkSZKkhWHWPd0ASU4BngcE+FxVfaSrwiRJkiRJmnRzanQDJFnOjB7yqrp71EVJkiRJkrQQzHp4eZKfB34DeBCYptfbXcCTuylNkiRJkqTJNuslw4CzgMOq6pCqenJVPamqOm9wJzmhg/tcnuTQ7Ww/vIOs/ZPs37++T5KXJzls1DmPkv3fG+U8qf+4/k1H9/+vkyzrX0+SM5L8fpL/lGQu8xLMJuuntma1kOQFIvArFAAAE8ZJREFUSZ7ev/7vkpyV5KSOsnZP8ookv5zkF5Oc2J8QUZIkqTNJfqn//jtJzkvylSQ/0WHe82azTWplLm+4bwAe6KqQAc4b5Z311xv/FvChJNclOXrG7gtHnPXzwN8DVyf5T8DlwH8APpzk9SPOes82l98H3rj16xFnfXTG9ZcCfw28BPiLJK8bZVbfJ/mn5+q7gJOALwJHM/plIP4EuDXJRUlenGTxiO//HyV5N73Hc1GS3wTOBnYFfjnJ/xhx1qnAZ4ETgV8Ang38NHBtkpGuQJBkSZKfT/KpJF9L8tUkVyT5j0mmRpk1i1pG+vxIsrj/2H5z2xfvJG8fcdZuSd6S5M1JliV5XZKPJTk7ye6jzHqU/Os7ut/DZ1yfSvL2/uP670l2G3HWLyR5Qv/6U5J8Lsk9Sb446ud9P+PDSV7b6Pfz5CTnJ/mt/gdqH0jyjSR/luSQEWctSnJmkk/0/56/nOSyJMeOMqefNRbHj1EfO/r36fHjsd/vgjx+tDx29POaHT9mOLOqNgA/AewDnEHvPVBXfn+W26Qm5rJk2BHABfQaOw9v3V5V/99jLiL52KPtAl5YVY97rBkzsq4FfrKq1iV5NvBHwOqq+nCSf6iqI0aY9XXgx+g1pG4BnlJVtyfZE/hsVa0YYdatwN8AV9L7uQH8Dr0RClTVH44w6x9/Tkm+ALymqm7qvzh9pqqeNaqsfsb/qaof7V//Mr314qf7X391lHlJ/gF4IfAK4DTgGcBHgEur6qpR5fSzruvf/67AbcATq+qB/hvLf6iqZ4ww62vAj/fv/wnAxVX17/tvYN5fVc8dYdalwD3AHwK39jcfBPwssFdVvWpUWf28vR5tF71lDQ8aYdYHgd2Aa+h9aHFVVf2X/r6vVNWRI8z6U+B79J4fTwe+CfwpvQ+49q+qnx5h1kZ6pwvBPx0/dqP3QWtV1fIRZv3jzynJ7wJ703tteRmwd1X9zAizrquqw/rXPwF8sKo+0m8s/v9VNdJejyS30fug9YXAX9Fb7eMTVfXIKHP6WZ/r3//jgdfS+xn+Kb03tK+pqheOMOsCeq9hf0Xv2LgB+FvgV4C/qKqRvZFtefxoeezo53n8eOxZC/L40fLY0c9rdvyYkfm1qjo8yTnA3/R/liN9393PeQ7wXOBNwO/N2LUcOHnU71Gl2ZrL0Nz/Ta9H8+v0zukepefT+6O/b5vtodcjN0pLqmodQFVdk+Q44PIkB/FPLxqjsrmqHgAeSHJDVd3ez12fZNRZP0rvnPsTgTdX1W1J3jnKxvYMM2tfUlU3AVTVXUlG/dwA+F6SF1bVXwM3Az8C3JJk7w6yqqrWAx8APpDeqQGnAu9KclBV/ciIs2rGz2zrz3WauY1CmY3Qm48B4H5g334BX0tvcsRROrKqnr7NtlvpjfjoovfjTnoNgszYVv2v9x1x1rOr6nCAJO8F/leSDwOnb5M/Ck+rqlOTBFgHvKj/fPlb4KsjzrqQ3puvN1fVDwCS3FRVTxpxDvzzn9Px9D5E29R/EzjqxzXzNW7frStuVNXfJNljxFkAd1TVK/r3/TLgDcC5SS6n98HdlSPM2qOq3geQ5I1V9bv97ecl+YUR5gAcVVVn9K9/PsnVVfWO/u/sWkbbe9Ty+NHy2AEeP0ZhoR4/Wh47oO3xY6svJ7kSeBLwtv5j7eI94y7A7vR+fzN/TxvofWgo7RBzaXRv3vqJbAeuBh7YXk9ikv874qwNSQ6tqhsA+j3exwIfBUZ9rvWWJFNVtYnekGgA0jtneKSNqv6QnTclOQr44/6nsl2dr3t4kg30XvyWJdm/34O/C9DFcOyfA/4oya8B99IbEv0PwJ5AV89JAPoflLwHeE+Sg0d8959I8nlgKfBB4E+TXA0cA3xu1FnAp5JcBfwk8Gfwjz09o36ztz7JK4EPzRiRsAh4JbB+xFkANwLHV9V3t92R5Hsjztpl65Wq2gysSvIOeh9IdjIssP9G+ZPVH5bU/3qkH9pV1S/2jx2Xpnf6yHsZ/YeQWz0+ycn0jk9L+8fHTh4X8OdJLqT3geRHkrwJ+DC9N+v/4vkyAlt/RxuBi+idOrIXvQ/u3kpvJNKoTCd5Gr3Gzm5JVlbV2iRPYfTH4U1bXzeTHAk8AlBVD3fwO2t5/Gh57ACPH6OwUI8fLY8d0Pb4sdXrgRXAjf1Rd3vRG2I+Uv22xFVJLqyqW0Z9/9J8zaXR/dkkq4CP88+Hl49iybAb6b+Ib6uqXjCC+59pPXAgvXPUt2ZsTHIivYPbKH2N3vDyz1fVrTO27w3811EG9T81v6SqvpDkhcAbgc+PMmOG/93P+rtttu8G/HwHeb8CvJ3e7+6p9D5VvxX40tY3ZSO0LMlzq+oL2+7o4OC9O/A24JGq+mJ6k/udTK8B/ucjztqD3vPhYeDXq+qv+tvvAUY2pLHvNGANvV6crW+S/xW9c8pPG3EWwLvpfQCzvTdBZ484a22SE6vqU1s3VNVvJPk+8L4Osnavqvuq6sytG/vPk40jzqKqvpzkRfTO+b8K6GpCwauAn+pfvzrJflX1g/6okrtGGVRVv5rePBOXAofS+4BrFb0PWV8zyqy+bUdrbX2NfH//Mkpvofd6PE2vZ+xtSZ5FbwjlG0ac9WZ67wEeAqbo/x0n2YfeXCWj1PL40fLYAR4/RmGhHj9aHjug7fFjq+cA11bV/UleS++9xzkdZQEsTW9uhkP450sdj3zovDQbczmn+6btbK4awQzmSX6J3ovpAfQmsrq0qq59rPdr1uRntc7bCbIOBC7rMmub3L3pHWdG+mZoZ5ckNduD9/zu/wDgiKr6ZFcZGr305mtYX1VbOrjv0DtnttnfssePbnj80PZ0efzo3//XgGcBh9PrzT8PeHlVHdNR3lfpfWDxZeAfH1NVfbmLPGmYWTe6W+gP3z2tf1lG79PFy6pq5OeBPkrWpVX17UZZLR/XxGe1zjOrG0lOqKpPt8hqnWeWWeOSN+lZ6c0zsc/W08BmbD+8qr42qVmt88wya8zyvlJVR/ZPqbitqs7LiCcQ3Cbvy1V1VBf3Lc1LVc36Qm+m5VOBn9l6mcvt55h1BPAPwJauMsyazKyF/NgWataMzO+2ymqdZ5ZZ45I3yVn99xjfpzdB23X0Jsrauu8rk5q1kB+bWZOVtSPy+vd7Fb1T6q4H9qd37vjXu8jq5/0avVMtDwD22nrpKs+Ll2GXWZ/TneSdwLH0Zsn+JL3JmD5Pb8mtkUhvqaQT6fXCHU/vD/TXR3X/Zk1uVus8sx5TxqAlAEc+43zLPLPMGpe8hZoFrKY3W/rWZT0vSrK6qj7M6Cd9bJnVOs8ss8YpD+BVwKv/X3v3F+JZXcZx/P1xM/VmZ2/Cu1wsL1y3WQ0DSROVEiQMItGu+oNQd0FSEN1Igt3Y6kUgxNpKRv9LU8MKwg0RK1l3bbf1ypA1KlkKzdYGZduni/P75bSNO3Nmf+ec35/3CwZ2vrMzn+di9+KZ7znPA9xazfDddwJ3dZQFzZpBaGZSjBVwxq/FSpvRZpDaTTTvYhysqk8nOZ9m4NMZS/IhmpUZH6bZX/l94DNV9dokfr5Zs5vVd55ZE9HnCsC+88wya1ry5jWrz7WefWb1nWeWWdOURzXbYO5e9fmLTPDibo28LlbWSZvWpuleqaqTSU6M3gM5xuR+W/Rl4LvAF2oy09DNmp+svvPMOnN9rgDsO88ss6Ylb16z+lzr2WdW33lmmTVNeSS5Avg6cDHN+rwtwPGqWuoo7xNrnVdVZ42+dDptmu79SbYBe2gmAR6nuTE7Y1V17SR+jlnzl9V3nlkT0ecKwL7zzDJrWvLmNavPtZ59ZvWdZ5ZZ05QHzf72jwM/Ai6nmQ11UUdZAO9b9edzaV6pO0CHt+vS6WxqenmS7cDWWjXdMMklVXVkcqVJmkVxpZxZC5jVd55Zs5XVd55ZZk1T3ihzf1VdnuRQVS2Pzp6qqvd3mbsqfwn4dlV9ZN2/LHVgYivD0uHYf0mzJ66UM2sBs/rOW7CsPtd6dpLVd55ZZk1LXpIngA/SzIN6Cfgr8Kmq2jXprLfIPxs4VFUX95EnnWqSTffBqrpsIj9M0lxJchmwF1iuqi3zlGeWWdOSZ9ZsZfWdZ5ZZQ+aNGvxjwNnA54El4N6qen7SWaO8R3lzKNwWmnfJf1hVX+oiT1rPWRP8WZ1MO5Q0m5KcneTGJN8Bfk6zm/Nj85BnllnTkmfWbGX1nWeWWdOSV1VHq2qlql6tqq9U1W1dNdwjXwN2jz6+Clxtw60h+Xi5pInK2uvJflr9rpTrJM8ss6Ylz6zZyuo7zyyzpiUvyWFOczE3fr+7C2nWG48Hqj1dVce6ypLWM8mm+7dVdcVEfpikmZVkH816sp9UDyvl+swzy6xpyTNrtrL6zjPLrGnJS3IRcD7wp1O+dAHwlw4fL78ZuAv4NRDgA8AXq+rHXeRJ61m36U5y2tvrqjow0YokSZIkzbwkPwO+vHrj0ej8cuD2qrqxo9zfAx8a324neQfwq74Gt0mn2sie7t2n+VoB102oFkmSJEnzY/upDTdAVe1Ps4K4K2ed8jj535nsLCuplXWb7qq6to9CJEmSJM2Vc0/ztfM6zP1Fkl/SrEEDuIVmWJw0iFbvdCfZCexg1X+gqnqgg7okSZIkzbAk3wMer6o9p5zfClxfVbd0mP0x4Eqad7qfqKqHusqS1rPhpjvJ7cA1NE33Y8ANwJNVdVNn1UmSJEmaSaMJ4g8BbwDPjI4vB94OfLSqXuo4fyurnuztY1CdtJY2TfdhYBdwsKp2jf4T3dfVAARJkiRJsy/JtcDO0adHqurxjvM+C9wBrAAnaW67q6ou7DJXeisbGaQ2tlJVJ5OcGP3W6BjgP1xJkiRJb6mq9gH7eoz8AnBJVf2tx0zpLbVpuvcn2QbsoXk85DjwdCdVSZIkSdLm/BH419BFSGOtBqn995uaEf9b11oBIEmSJElDSXIZcD/wO+D18XlVfW6worTQ2tx0k2QZ2D7+viTvrqoHO6hLkiRJkjbjG8DjwGGad7qlQW246U6yF1gGjvDmP94CbLolSZIkTYsTVXXb0EVIY22mlz9XVTs6rkeSJEmSNi3JncBR4FH+9/FyV4ZpEG2a7m8Cu6vquW5LkiRJkqTNSfLCGseuDNNg2jTdV9P8tuglmt8YjffdLXdXniRJkiRJs6tN0/08cBunDCSoqqPdlCZJkiRJ7SXZCewAzh2fVdUDw1WkRdZmevmLVfVIZ5VIkiRJ0hlKcjtwDU3T/RhwA/AkYNOtQbS56b4X2Mb/DyRwerkkSZKkqZDkMLALOFhVu5KcD9xXVTcOXJoWVJub7vNomu3rV525MkySJEnSNFmpqpNJTiTZChwDHKKmwWyo6U6yBThUVfd0XI8kSZIknYn9SbYBe4BngOPA08OWpEXW5vHyfVV1bcf1SJIkSdJEJNkObK2qQ6vOLqmqI4MVpYXTpum+E1gCfgC8Nj6vqgPdlCZJkiRJk5XkQFW9d+g6tDha3XSvcVxVdd1kS5IkSZKkbiQ5WFWXDV2HFseGB6n5aLkkSZKkObCxW0dpQs7a6F9MspTk7iT7Rx+7kyx1WZwkSZIkSbNsw003sBf4J3Dz6ONV4P4uipIkSZKkjrwxdAFaLG3e6X62qi5d70ySJEmS+pbktMPRHACtoWz4nW5gJclVVfUkQJIrgZVuypIkSZKkVnaf5msFOABag2hz030p8C2atWEALwOfXL3zTpIkSZIkvalN030OcBPwLmAb8A+alWF3dFeeJEmSJLWTZCewAzh3fFZVDwxXkRZZm8fLHwZeAQ4Af+6mHEmSJEnavCS3A9fQNN2PATcATwI23RpEm5vuP1TVzo7rkSRJkqRNS3IY2AUcrKpdSc4H7quqGwcuTQuqzcqwp5K8p7NKJEmSJOnMrVTVSeBEkq3AMeDCgWvSAmvzePlVwKeSvAC8DoTmne7lTiqTJEmSpPb2J9kG7AGeAY4DTw9bkhZZm8fLL1jrvKqOTrQiSZIkSZqAJNuBrW5c0pA23HRLkiRJ0ixIsgxsZ9WTvVX14GAFaaG1ebxckiRJkqZakr3AMnAEODk6LsCmW4PwpluSJEnS3EjyXFXtGLoOaazN9HJJkiRJmna/SWLTranhTbckSZKkuZHkauBR4CXcuqQpYNMtSZIkaW4keR64DTjMm+90u3VJg3GQmiRJkqR58mJVPTJ0EdKYN92SJEmS5kaSe4FtNI+Yvz4+d2WYhuJNtyRJkqR5ch5Ns339qjNXhmkwNt2SJEmS5kKSLcChqrpn6FqkMVeGSZIkSZoLVfVv4CND1yGt5jvdkiRJkuZGkjuBJeAHwGvj86o6MFhRWmg23ZIkSZLmRpJ9axxXVV3XezESNt2SJEmSJHXGd7olSZIkzY0kS0nuTrJ/9LE7ydLQdWlx2XRLkiRJmid7gX8CN48+XgXuH7QiLTQfL5ckSZI0N5I8W1WXrncm9cWbbkmSJEnzZCXJVeNPklwJrAxYjxacN92SJEmS5kaSS4Fv0awNA3gZ+GRVHRquKi0ym25JkiRJcyPJOcBNwLuAbcA/aFaG3TFoYVpYbxu6AEmSJEmaoIeBV4ADwJ8HrkXypluSJEnS/Ejyh6raOXQd0piD1CRJkiTNk6eSvGfoIqQxb7olSZIkzY0kzwHvBl4AXgdC80738qCFaWHZdEuSJEmaG0kuWOu8qo72XYsENt2SJEmSJHXGd7olSZIkSeqITbckSZIkSR2x6ZYkSZIkqSM23ZIkSZIkdeQ/g4XMdnce5/UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1800x1080 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# correlation matrix\n",
    "corrmat =under_sample.corr()\n",
    "fig,ax= plt.subplots()\n",
    "fig.set_size_inches(25,15)\n",
    "sns.heatmap(corrmat,square=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=None, max_features='sqrt', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=-1,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(n_estimators=100, \n",
    "                            criterion='gini', \n",
    "                            max_features='sqrt',\n",
    "                            n_jobs=-1)\n",
    "rf.fit(X_under_train, y_under_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature ranking:\n",
      "1. V14 (0.163468)\n",
      "2. V10 (0.142475)\n",
      "3. V4 (0.135820)\n",
      "4. V17 (0.077558)\n",
      "5. V3 (0.066908)\n",
      "6. V12 (0.063024)\n",
      "7. V11 (0.059748)\n",
      "8. V7 (0.035103)\n",
      "9. V2 (0.029975)\n",
      "10. V16 (0.025015)\n",
      "11. V18 (0.020058)\n",
      "12. V8 (0.014964)\n",
      "13. V6 (0.014854)\n",
      "14. V5 (0.011935)\n",
      "15. V23 (0.011827)\n",
      "16. V9 (0.011814)\n",
      "17. V19 (0.011713)\n",
      "18. V13 (0.011508)\n",
      "19. V1 (0.010258)\n",
      "20. V27 (0.010227)\n",
      "21. V20 (0.010050)\n",
      "22. normal_amount (0.009788)\n",
      "23. V21 (0.008971)\n",
      "24. V22 (0.008304)\n",
      "25. V15 (0.008091)\n",
      "26. V28 (0.006979)\n",
      "27. V26 (0.006722)\n",
      "28. V25 (0.006679)\n",
      "29. V24 (0.006168)\n"
     ]
    },
    {
     "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": [
    "feature_importance(rf,X_under_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### From the above heeatmap we can say that many of features are correlated but we are more interested in correlation of features with class. So I am going to list those features whose correlation coefficient w.r.t class is less than -0.6 or greater than 0.6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>V10</th>\n",
       "      <td>-0.633606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V12</th>\n",
       "      <td>-0.677653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V14</th>\n",
       "      <td>-0.740321</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Class\n",
       "V10 -0.633606\n",
       "V12 -0.677653\n",
       "V14 -0.740321"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#negative correlations smaller than -0.5\n",
    "corr = under_sample.corr()\n",
    "corr = corr[['Class']]\n",
    "corr[corr.Class < -0.6]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>V4</th>\n",
       "      <td>0.698864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>V11</th>\n",
       "      <td>0.688469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Class</th>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Class\n",
       "V4     0.698864\n",
       "V11    0.688469\n",
       "Class  1.000000"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#positive correlations greater than 0.5\n",
    "corr[corr.Class > 0.6]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### BoxPlots\n",
    "We will use boxplots to have a better understanding of the distribution of these features in fradulent and non fradulent transactions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1872x1152 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#visualizing the features with high correlation\n",
    "f, axes = plt.subplots(nrows=2, ncols=3, figsize=(26,16))\n",
    "f.suptitle('Features With High Correlation', size=35)\n",
    "sns.boxplot(x=\"Class\", y=\"V10\", data=under_sample, ax=axes[0,0])\n",
    "sns.boxplot(x=\"Class\", y=\"V12\", data=under_sample, ax=axes[0,1])\n",
    "sns.boxplot(x=\"Class\", y=\"V14\", data=under_sample, ax=axes[0,2])\n",
    "sns.boxplot(x=\"Class\", y=\"V4\", data=under_sample, ax=axes[1,0])\n",
    "sns.boxplot(x=\"Class\", y=\"V11\", data=under_sample, ax=axes[1,1])\n",
    "f.delaxes(axes[1,2])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Box plots provide us with a good intuition of whether we need to worry about outliers as all transactions outside of 1.5 times the IQR (Inter-Quartile Range) are usually considered to be outliers. However, removing all transactions outside of 1.5 times the IQR would dramatically decrease  training data size, which is not very large, to begin with. Thus, I decided to only focus on extreme outliers outside of 2.5 times the IQR."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(844, 30)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "under_sample.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Removing extreme outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "v14_fraud = under_sample['V14'].loc[under_sample['Class'] == 1].values\n",
    "q25, q75 = np.percentile(v14_fraud, 25), np.percentile(v14_fraud, 75)\n",
    "v14_iqr = q75 - q25\n",
    "v14_cut_off = v14_iqr * 2.5\n",
    "v14_lower, v14_upper = q25 - v14_cut_off, q75 + v14_cut_off\n",
    "under_sample= under_sample.drop(under_sample[(under_sample['V14'] > v14_upper) | (under_sample['V14'] < v14_lower)].index)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "v12_fraud = under_sample['V12'].loc[under_sample['Class'] == 1].values\n",
    "q25, q75 = np.percentile(v12_fraud, 25), np.percentile(v12_fraud, 75)\n",
    "v12_iqr = q75 - q25\n",
    "v12_cut_off = v12_iqr * 2.5\n",
    "v12_lower, v12_upper = q25 - v12_cut_off, q75 + v12_cut_off\n",
    "under_sample= under_sample.drop(under_sample[(under_sample['V12'] > v12_upper) | (under_sample['V12'] < v12_lower)].index)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "v10_fraud = under_sample['V10'].loc[under_sample['Class'] == 1].values\n",
    "q25, q75 = np.percentile(v10_fraud, 25), np.percentile(v10_fraud, 75)\n",
    "v10_iqr = q75 - q25\n",
    "v10_cut_off = v10_iqr * 2.5\n",
    "v10_lower, v10_upper = q25 - v10_cut_off, q75 + v10_cut_off\n",
    "under_sample= under_sample.drop(under_sample[(under_sample['V10'] > v10_upper) | (under_sample['V10'] < v10_lower)].index)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "v4_fraud = under_sample['V4'].loc[under_sample['Class'] == 1].values\n",
    "q25, q75 = np.percentile(v4_fraud, 25), np.percentile(v4_fraud, 75)\n",
    "v4_iqr = q75 - q25\n",
    "v4_cut_off = v4_iqr * 2.5\n",
    "v4_lower, v4_upper = q25 - v4_cut_off, q75 + v4_cut_off\n",
    "under_sample= under_sample.drop(under_sample[(under_sample['V4'] > v4_upper) | (under_sample['V4'] < v4_lower)].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "v11_fraud = under_sample['V11'].loc[under_sample['Class'] == 1].values\n",
    "q25, q75 = np.percentile(v11_fraud, 25), np.percentile(v11_fraud, 75)\n",
    "v11_iqr = q75 - q25\n",
    "v11_cut_off = v11_iqr * 2.5\n",
    "v11_lower, v11_upper = q25 - v11_cut_off, q75 + v11_cut_off\n",
    "under_sample= under_sample.drop(under_sample[(under_sample['V11'] > v11_upper) | (under_sample['V11'] < v11_lower)].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(976, 30)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "under_sample.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### After removing outliers our under sample dataframe is reduced to "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I have tried to run this notebook without removing outliers and after removing outliers but I got better result before removing outliers.Before even thinking of removing outliers there should be enough evidence that these observations are actual outliers Before removing outliers. It should be done by in depth statistical analysis to make sure that these observations are actual outliers because different ML methods used for detecting fraud, are based on anomaly detection and they treat such extreme outliers as frauds. So, by deleting it we delete the most important observations, that have higher probability of being frauds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Splitting under sampled dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_under = under_sample.loc[:,under_sample.columns != 'Class']\n",
    "y_under = under_sample.loc[:,under_sample.columns == 'Class']\n",
    "X_under_train, X_under_test, y_under_train, y_under_test = train_test_split(X_under,y_under,test_size = 0.3, random_state = 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3] Logistic regression with C=0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_under_C1 = LogisticRegression(C=0.01,penalty = 'l1')\n",
    "lr_under_C1.fit(X_under_train,y_under_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  96.23%\n",
      "Recall: 0.9115646258503401\n",
      "Log Loss: 3.5007281778367507\n",
      "Precision: 0.01526022093155677\n",
      "Accurcay: 0.8986458808796508\n",
      "AUC:  0.9050941212162975\n",
      "F1 Score: 0.030017921146953407\n",
      "Confusion matrix:\n",
      " [[76649  8647]\n",
      " [   13   134]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Prediction on original dataframe\n",
    "y_pred_full_model1 = lr_under_C1.predict(X_test)\n",
    "cross_validation_metrics(lr_under_C1,X_under_train,y_under_train)\n",
    "print_accuracy_metrics(y_test,y_pred_full_model1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4] Logistic regression with C=0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=0.1, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_under_C2 = LogisticRegression(C=0.1,penalty = 'l1')\n",
    "lr_under_C2.fit(X_under_train,y_under_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  88.99%\n",
      "Recall: 0.9115646258503401\n",
      "Log Loss: 0.6209142741882784\n",
      "Precision: 0.08086904043452021\n",
      "Accurcay: 0.9820231031213792\n",
      "AUC:  0.9468545789165413\n",
      "F1 Score: 0.14855875831485585\n",
      "Confusion matrix:\n",
      " [[83773  1523]\n",
      " [   13   134]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Prediction on original dataset\n",
    "y_pred_full_model2 = lr_under_C2.predict(X_test)\n",
    "cross_validation_metrics(lr_under_C2,X_under_train,y_under_train)\n",
    "print_accuracy_metrics(y_test,y_pred_full_model2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5] Logistic regression with C=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_under_C3 = LogisticRegression(C=1,penalty = 'l1')\n",
    "lr_under_C3.fit(X_under_train,y_under_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  91.19%\n",
      "Recall: 0.9259259259259259\n",
      "Log Loss: 1.4306248896684717\n",
      "Precision: 0.03190810465858328\n",
      "Accurcay: 0.9585801233609739\n",
      "AUC:  0.9422770694605426\n",
      "F1 Score: 0.061690314620604564\n",
      "Confusion matrix:\n",
      " [[70301  3034]\n",
      " [    8   100]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Prediction on original dataset\n",
    "y_pred_full_model3 = lr_under_C3.predict(X_test)\n",
    "cross_validation_metrics(lr_under_C3,X_under_train,y_under_train)\n",
    "print_accuracy_metrics(y_test,y_pred_full_model3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6] Logistic regreesion with C=10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=10, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_under_C4 = LogisticRegression(C=10,penalty = 'l1')\n",
    "lr_under_C4.fit(X_under_train,y_under_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  90.51%\n",
      "Recall: 0.9259259259259259\n",
      "Log Loss: 1.6827008224584545\n",
      "Precision: 0.027247956403269755\n",
      "Accurcay: 0.951281946543578\n",
      "AUC:  0.9386226070619608\n",
      "F1 Score: 0.05293806246691372\n",
      "Confusion matrix:\n",
      " [[69765  3570]\n",
      " [    8   100]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Prediction on original dataset\n",
    "y_pred_full_model4 = lr_under_C4.predict(X_test)\n",
    "cross_validation_metrics(lr_under_C4,X_under_train,y_under_train)\n",
    "print_accuracy_metrics(y_test,y_pred_full_model4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7] Decision Tree Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,\n",
       "            max_features=None, max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, presort=False, random_state=None,\n",
       "            splitter='best')"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DecisionTreeClassifier= DecisionTreeClassifier()\n",
    "DecisionTreeClassifier.fit(X_under_train,y_under_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Cross Validation Score(Recall):  87.12%\n",
      "Recall: 0.9722222222222222\n",
      "Log Loss: 3.9090580892603124\n",
      "Precision: 0.012479201331114808\n",
      "Accurcay: 0.886823795324265\n",
      "AUC:  0.9294601259062294\n",
      "F1 Score: 0.02464210279277165\n",
      "Confusion matrix:\n",
      " [[65026  8309]\n",
      " [    3   105]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Prediction on original dataset\n",
    "y_pred_DecisionTree = DecisionTreeClassifier.predict(X_test)\n",
    "cross_validation_metrics(DecisionTreeClassifier,X_under_train,y_under_train)\n",
    "print_accuracy_metrics(y_test,y_pred_DecisionTree)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x2293b290>"
      ]
     },
     "execution_count": 45,
     "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": [
    "plt.figure(0).clf()\n",
    "\n",
    "fpr, tpr, thresh = metrics.roc_curve(y_test, y_balanced_pred)\n",
    "auc = metrics.roc_auc_score(y_test, y_balanced_pred)\n",
    "plt.plot(fpr,tpr,label=\"Logistic Regtession Class weight, auc=\"+ '{0:.3f}'.format(auc))\n",
    "\n",
    "fpr, tpr, thresh = metrics.roc_curve(y_test, y_pred_full_model1)\n",
    "auc = metrics.roc_auc_score(y_test, y_pred_full_model1)\n",
    "plt.plot(fpr,tpr,label=\"Logistic regression(C=0.01), auc=\"+ '{0:.3f}'.format(auc))\n",
    "\n",
    "fpr, tpr, thresh = metrics.roc_curve(y_test, y_pred_full_model2)\n",
    "auc = metrics.roc_auc_score(y_test, y_pred_full_model2)\n",
    "plt.plot(fpr,tpr,label=\"Logistic regression(C=0.1), auc=\"+'{0:.3f}'.format(auc))\n",
    "\n",
    "fpr, tpr, thresh = metrics.roc_curve(y_test, y_pred_full_model3)\n",
    "auc = metrics.roc_auc_score(y_test, y_pred_full_model3)\n",
    "plt.plot(fpr,tpr,label=\"Logistic regression(C=1), auc=\"+'{0:.3f}'.format(auc))\n",
    "\n",
    "fpr, tpr, thresh = metrics.roc_curve(y_test, y_pred_full_model4)\n",
    "auc = metrics.roc_auc_score(y_test, y_pred_full_model4)\n",
    "plt.plot(fpr,tpr,label=\"Logistic regression(C=10), auc=\"+'{0:.3f}'.format(auc))\n",
    "\n",
    "plt.plot(fpr, tpr)\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.0])\n",
    "plt.title('ROC Curve\\n AUC={auc}'.format(auc = auc))\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.grid(True)\n",
    "plt.legend(loc=\"lower right\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Obviously, trying to increase recall, tends to come with a decrease of precision. However, in our case, if we predict that a transaction is fraudulent and turns out not to be, is not a massive problem compared to the opposite."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predictions on new samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test=pd.read_csv('creditcard_test_dataset.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>...</th>\n",
       "      <th>V20</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>30808</td>\n",
       "      <td>109847</td>\n",
       "      <td>1.930307</td>\n",
       "      <td>-0.234417</td>\n",
       "      <td>-1.583368</td>\n",
       "      <td>0.024786</td>\n",
       "      <td>1.016470</td>\n",
       "      <td>1.179664</td>\n",
       "      <td>-0.386242</td>\n",
       "      <td>0.327416</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.437585</td>\n",
       "      <td>0.211402</td>\n",
       "      <td>1.015098</td>\n",
       "      <td>0.073241</td>\n",
       "      <td>-0.804496</td>\n",
       "      <td>0.072202</td>\n",
       "      <td>-0.383211</td>\n",
       "      <td>0.016787</td>\n",
       "      <td>-0.081815</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>174948</td>\n",
       "      <td>84730</td>\n",
       "      <td>-5.053316</td>\n",
       "      <td>-3.617236</td>\n",
       "      <td>-0.323455</td>\n",
       "      <td>1.447171</td>\n",
       "      <td>-0.097495</td>\n",
       "      <td>-0.154917</td>\n",
       "      <td>-2.076441</td>\n",
       "      <td>1.331307</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.351008</td>\n",
       "      <td>-0.202483</td>\n",
       "      <td>0.036442</td>\n",
       "      <td>-1.901898</td>\n",
       "      <td>0.090641</td>\n",
       "      <td>-1.777126</td>\n",
       "      <td>0.947972</td>\n",
       "      <td>0.954071</td>\n",
       "      <td>-1.577919</td>\n",
       "      <td>333.48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>203982</td>\n",
       "      <td>95957</td>\n",
       "      <td>0.090684</td>\n",
       "      <td>1.197902</td>\n",
       "      <td>-1.370219</td>\n",
       "      <td>0.770437</td>\n",
       "      <td>0.857203</td>\n",
       "      <td>-0.698486</td>\n",
       "      <td>1.350617</td>\n",
       "      <td>-0.952780</td>\n",
       "      <td>...</td>\n",
       "      <td>0.617074</td>\n",
       "      <td>-0.255027</td>\n",
       "      <td>0.378177</td>\n",
       "      <td>-0.126596</td>\n",
       "      <td>0.572250</td>\n",
       "      <td>0.347829</td>\n",
       "      <td>0.538705</td>\n",
       "      <td>-0.211582</td>\n",
       "      <td>-0.267241</td>\n",
       "      <td>50.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>283146</td>\n",
       "      <td>142109</td>\n",
       "      <td>1.808668</td>\n",
       "      <td>-1.197522</td>\n",
       "      <td>-2.291089</td>\n",
       "      <td>-0.690236</td>\n",
       "      <td>-0.312654</td>\n",
       "      <td>-1.756994</td>\n",
       "      <td>0.631548</td>\n",
       "      <td>-0.725965</td>\n",
       "      <td>...</td>\n",
       "      <td>0.457894</td>\n",
       "      <td>0.498576</td>\n",
       "      <td>0.927097</td>\n",
       "      <td>-0.283918</td>\n",
       "      <td>0.134011</td>\n",
       "      <td>0.372327</td>\n",
       "      <td>0.050185</td>\n",
       "      <td>-0.121994</td>\n",
       "      <td>-0.038088</td>\n",
       "      <td>254.04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>47316</td>\n",
       "      <td>129034</td>\n",
       "      <td>-0.920372</td>\n",
       "      <td>1.003867</td>\n",
       "      <td>-0.110709</td>\n",
       "      <td>-2.771173</td>\n",
       "      <td>0.841827</td>\n",
       "      <td>-0.497278</td>\n",
       "      <td>0.769851</td>\n",
       "      <td>0.474075</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.175248</td>\n",
       "      <td>-0.238352</td>\n",
       "      <td>-0.849203</td>\n",
       "      <td>-0.124786</td>\n",
       "      <td>0.086646</td>\n",
       "      <td>-0.007944</td>\n",
       "      <td>0.360582</td>\n",
       "      <td>0.090110</td>\n",
       "      <td>0.091000</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       id    Time        V1        V2        V3        V4        V5        V6  \\\n",
       "0   30808  109847  1.930307 -0.234417 -1.583368  0.024786  1.016470  1.179664   \n",
       "1  174948   84730 -5.053316 -3.617236 -0.323455  1.447171 -0.097495 -0.154917   \n",
       "2  203982   95957  0.090684  1.197902 -1.370219  0.770437  0.857203 -0.698486   \n",
       "3  283146  142109  1.808668 -1.197522 -2.291089 -0.690236 -0.312654 -1.756994   \n",
       "4   47316  129034 -0.920372  1.003867 -0.110709 -2.771173  0.841827 -0.497278   \n",
       "\n",
       "         V7        V8  ...       V20       V21       V22       V23       V24  \\\n",
       "0 -0.386242  0.327416  ... -0.437585  0.211402  1.015098  0.073241 -0.804496   \n",
       "1 -2.076441  1.331307  ... -1.351008 -0.202483  0.036442 -1.901898  0.090641   \n",
       "2  1.350617 -0.952780  ...  0.617074 -0.255027  0.378177 -0.126596  0.572250   \n",
       "3  0.631548 -0.725965  ...  0.457894  0.498576  0.927097 -0.283918  0.134011   \n",
       "4  0.769851  0.474075  ... -0.175248 -0.238352 -0.849203 -0.124786  0.086646   \n",
       "\n",
       "        V25       V26       V27       V28  Amount  \n",
       "0  0.072202 -0.383211  0.016787 -0.081815    1.00  \n",
       "1 -1.777126  0.947972  0.954071 -1.577919  333.48  \n",
       "2  0.347829  0.538705 -0.211582 -0.267241   50.14  \n",
       "3  0.372327  0.050185 -0.121994 -0.038088  254.04  \n",
       "4 -0.007944  0.360582  0.090110  0.091000    1.00  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>...</th>\n",
       "      <th>V20</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>39995</th>\n",
       "      <td>33211</td>\n",
       "      <td>87180</td>\n",
       "      <td>-0.227015</td>\n",
       "      <td>1.066681</td>\n",
       "      <td>0.167350</td>\n",
       "      <td>0.847756</td>\n",
       "      <td>0.918577</td>\n",
       "      <td>-0.569415</td>\n",
       "      <td>1.412774</td>\n",
       "      <td>-0.317352</td>\n",
       "      <td>...</td>\n",
       "      <td>0.007818</td>\n",
       "      <td>0.115365</td>\n",
       "      <td>0.685676</td>\n",
       "      <td>-0.300523</td>\n",
       "      <td>-0.008611</td>\n",
       "      <td>-0.105765</td>\n",
       "      <td>-0.433028</td>\n",
       "      <td>0.129265</td>\n",
       "      <td>0.010438</td>\n",
       "      <td>13.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39996</th>\n",
       "      <td>226678</td>\n",
       "      <td>66735</td>\n",
       "      <td>1.080631</td>\n",
       "      <td>0.072792</td>\n",
       "      <td>0.193645</td>\n",
       "      <td>1.189376</td>\n",
       "      <td>-0.379073</td>\n",
       "      <td>-0.941658</td>\n",
       "      <td>0.310233</td>\n",
       "      <td>-0.237765</td>\n",
       "      <td>...</td>\n",
       "      <td>0.009391</td>\n",
       "      <td>0.075056</td>\n",
       "      <td>0.056154</td>\n",
       "      <td>-0.170659</td>\n",
       "      <td>0.388427</td>\n",
       "      <td>0.646854</td>\n",
       "      <td>-0.336122</td>\n",
       "      <td>-0.004275</td>\n",
       "      <td>0.034426</td>\n",
       "      <td>89.16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39997</th>\n",
       "      <td>91144</td>\n",
       "      <td>69391</td>\n",
       "      <td>-0.366346</td>\n",
       "      <td>-0.123654</td>\n",
       "      <td>1.574127</td>\n",
       "      <td>-1.620753</td>\n",
       "      <td>-1.152640</td>\n",
       "      <td>-0.860367</td>\n",
       "      <td>-0.407476</td>\n",
       "      <td>-0.123524</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.106231</td>\n",
       "      <td>-0.167896</td>\n",
       "      <td>-0.034830</td>\n",
       "      <td>-0.125103</td>\n",
       "      <td>0.357849</td>\n",
       "      <td>-0.137647</td>\n",
       "      <td>-0.134737</td>\n",
       "      <td>0.253740</td>\n",
       "      <td>0.180461</td>\n",
       "      <td>10.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39998</th>\n",
       "      <td>151948</td>\n",
       "      <td>84607</td>\n",
       "      <td>1.245859</td>\n",
       "      <td>-0.208636</td>\n",
       "      <td>-0.083039</td>\n",
       "      <td>-0.792677</td>\n",
       "      <td>-0.621157</td>\n",
       "      <td>-1.219052</td>\n",
       "      <td>0.085329</td>\n",
       "      <td>-0.130222</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.217895</td>\n",
       "      <td>0.033446</td>\n",
       "      <td>0.241614</td>\n",
       "      <td>-0.169362</td>\n",
       "      <td>0.550034</td>\n",
       "      <td>0.853734</td>\n",
       "      <td>-0.576327</td>\n",
       "      <td>0.022538</td>\n",
       "      <td>0.001399</td>\n",
       "      <td>7.71</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39999</th>\n",
       "      <td>80174</td>\n",
       "      <td>70403</td>\n",
       "      <td>-0.633590</td>\n",
       "      <td>0.812096</td>\n",
       "      <td>2.561737</td>\n",
       "      <td>0.446063</td>\n",
       "      <td>-0.448034</td>\n",
       "      <td>-0.683693</td>\n",
       "      <td>0.267501</td>\n",
       "      <td>-0.008028</td>\n",
       "      <td>...</td>\n",
       "      <td>0.149068</td>\n",
       "      <td>0.007932</td>\n",
       "      <td>-0.015742</td>\n",
       "      <td>-0.131304</td>\n",
       "      <td>0.736886</td>\n",
       "      <td>-0.036651</td>\n",
       "      <td>0.325312</td>\n",
       "      <td>0.051134</td>\n",
       "      <td>0.087213</td>\n",
       "      <td>6.47</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           id   Time        V1        V2        V3        V4        V5  \\\n",
       "39995   33211  87180 -0.227015  1.066681  0.167350  0.847756  0.918577   \n",
       "39996  226678  66735  1.080631  0.072792  0.193645  1.189376 -0.379073   \n",
       "39997   91144  69391 -0.366346 -0.123654  1.574127 -1.620753 -1.152640   \n",
       "39998  151948  84607  1.245859 -0.208636 -0.083039 -0.792677 -0.621157   \n",
       "39999   80174  70403 -0.633590  0.812096  2.561737  0.446063 -0.448034   \n",
       "\n",
       "             V6        V7        V8  ...       V20       V21       V22  \\\n",
       "39995 -0.569415  1.412774 -0.317352  ...  0.007818  0.115365  0.685676   \n",
       "39996 -0.941658  0.310233 -0.237765  ...  0.009391  0.075056  0.056154   \n",
       "39997 -0.860367 -0.407476 -0.123524  ... -0.106231 -0.167896 -0.034830   \n",
       "39998 -1.219052  0.085329 -0.130222  ... -0.217895  0.033446  0.241614   \n",
       "39999 -0.683693  0.267501 -0.008028  ...  0.149068  0.007932 -0.015742   \n",
       "\n",
       "            V23       V24       V25       V26       V27       V28  Amount  \n",
       "39995 -0.300523 -0.008611 -0.105765 -0.433028  0.129265  0.010438   13.99  \n",
       "39996 -0.170659  0.388427  0.646854 -0.336122 -0.004275  0.034426   89.16  \n",
       "39997 -0.125103  0.357849 -0.137647 -0.134737  0.253740  0.180461   10.00  \n",
       "39998 -0.169362  0.550034  0.853734 -0.576327  0.022538  0.001399    7.71  \n",
       "39999 -0.131304  0.736886 -0.036651  0.325312  0.051134  0.087213    6.47  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_test.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test['normal_amount'] = StandardScaler().fit_transform(data_test['Amount'].values.reshape(-1,1))\n",
    "data_test1 = data_test.drop(['Amount','Time','id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_pred_class = lr_under_C2.predict(data_test1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 0, 0, 0], dtype=int64)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_pred_class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_pred_class_prob = lr_balanced.predict_proba(data_test1)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.04646066, 0.00050197, 0.05586541, ..., 0.01563743, 0.03944736,\n",
       "       0.06019283])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_pred_class_prob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.DataFrame({'ID':data_test.id,'Class':new_pred_class}).set_index('ID').to_csv('submission_credit1.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.DataFrame({'ID':data_test.id,'Class':new_pred_class_prob}).set_index('ID').to_csv('submission_credit_prob1.csv')"
   ]
  }
 ],
 "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}