{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import necessary libraries.\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "\n",
    "import os\n",
    "import tempfile\n",
    "\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "import sklearn\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "mpl.rcParams['figure.figsize'] = (12, 10)\n",
    "colors = plt.rcParams['axes.prop_cycle'].by_key()['color']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>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>V9</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>0.0</td>\n",
       "      <td>-1.359807</td>\n",
       "      <td>-0.072781</td>\n",
       "      <td>2.536347</td>\n",
       "      <td>1.378155</td>\n",
       "      <td>-0.338321</td>\n",
       "      <td>0.462388</td>\n",
       "      <td>0.239599</td>\n",
       "      <td>0.098698</td>\n",
       "      <td>0.363787</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018307</td>\n",
       "      <td>0.277838</td>\n",
       "      <td>-0.110474</td>\n",
       "      <td>0.066928</td>\n",
       "      <td>0.128539</td>\n",
       "      <td>-0.189115</td>\n",
       "      <td>0.133558</td>\n",
       "      <td>-0.021053</td>\n",
       "      <td>149.62</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.191857</td>\n",
       "      <td>0.266151</td>\n",
       "      <td>0.166480</td>\n",
       "      <td>0.448154</td>\n",
       "      <td>0.060018</td>\n",
       "      <td>-0.082361</td>\n",
       "      <td>-0.078803</td>\n",
       "      <td>0.085102</td>\n",
       "      <td>-0.255425</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.225775</td>\n",
       "      <td>-0.638672</td>\n",
       "      <td>0.101288</td>\n",
       "      <td>-0.339846</td>\n",
       "      <td>0.167170</td>\n",
       "      <td>0.125895</td>\n",
       "      <td>-0.008983</td>\n",
       "      <td>0.014724</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.358354</td>\n",
       "      <td>-1.340163</td>\n",
       "      <td>1.773209</td>\n",
       "      <td>0.379780</td>\n",
       "      <td>-0.503198</td>\n",
       "      <td>1.800499</td>\n",
       "      <td>0.791461</td>\n",
       "      <td>0.247676</td>\n",
       "      <td>-1.514654</td>\n",
       "      <td>...</td>\n",
       "      <td>0.247998</td>\n",
       "      <td>0.771679</td>\n",
       "      <td>0.909412</td>\n",
       "      <td>-0.689281</td>\n",
       "      <td>-0.327642</td>\n",
       "      <td>-0.139097</td>\n",
       "      <td>-0.055353</td>\n",
       "      <td>-0.059752</td>\n",
       "      <td>378.66</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.966272</td>\n",
       "      <td>-0.185226</td>\n",
       "      <td>1.792993</td>\n",
       "      <td>-0.863291</td>\n",
       "      <td>-0.010309</td>\n",
       "      <td>1.247203</td>\n",
       "      <td>0.237609</td>\n",
       "      <td>0.377436</td>\n",
       "      <td>-1.387024</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.108300</td>\n",
       "      <td>0.005274</td>\n",
       "      <td>-0.190321</td>\n",
       "      <td>-1.175575</td>\n",
       "      <td>0.647376</td>\n",
       "      <td>-0.221929</td>\n",
       "      <td>0.062723</td>\n",
       "      <td>0.061458</td>\n",
       "      <td>123.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.158233</td>\n",
       "      <td>0.877737</td>\n",
       "      <td>1.548718</td>\n",
       "      <td>0.403034</td>\n",
       "      <td>-0.407193</td>\n",
       "      <td>0.095921</td>\n",
       "      <td>0.592941</td>\n",
       "      <td>-0.270533</td>\n",
       "      <td>0.817739</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.009431</td>\n",
       "      <td>0.798278</td>\n",
       "      <td>-0.137458</td>\n",
       "      <td>0.141267</td>\n",
       "      <td>-0.206010</td>\n",
       "      <td>0.502292</td>\n",
       "      <td>0.219422</td>\n",
       "      <td>0.215153</td>\n",
       "      <td>69.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Time        V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0   0.0 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388  0.239599   \n",
       "1   0.0  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361 -0.078803   \n",
       "2   1.0 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
       "3   1.0 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
       "4   2.0 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921  0.592941   \n",
       "\n",
       "         V8        V9  ...       V21       V22       V23       V24       V25  \\\n",
       "0  0.098698  0.363787  ... -0.018307  0.277838 -0.110474  0.066928  0.128539   \n",
       "1  0.085102 -0.255425  ... -0.225775 -0.638672  0.101288 -0.339846  0.167170   \n",
       "2  0.247676 -1.514654  ...  0.247998  0.771679  0.909412 -0.689281 -0.327642   \n",
       "3  0.377436 -1.387024  ... -0.108300  0.005274 -0.190321 -1.175575  0.647376   \n",
       "4 -0.270533  0.817739  ... -0.009431  0.798278 -0.137458  0.141267 -0.206010   \n",
       "\n",
       "        V26       V27       V28  Amount  Class  \n",
       "0 -0.189115  0.133558 -0.021053  149.62      0  \n",
       "1  0.125895 -0.008983  0.014724    2.69      0  \n",
       "2 -0.139097 -0.055353 -0.059752  378.66      0  \n",
       "3 -0.221929  0.062723  0.061458  123.50      0  \n",
       "4  0.502292  0.219422  0.215153   69.99      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "file = tf.keras.utils\n",
    "raw_df = pd.read_csv('./creditcard.csv')\n",
    "raw_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Examples:\n",
      "    Total: 284807\n",
      "    Positive: 492 (0.17% of total)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "neg, pos = np.bincount(raw_df['Class'])\n",
    "total = neg + pos\n",
    "print('Examples:\\n    Total: {}\\n    Positive: {} ({:.2f}% of total)\\n'.format(\n",
    "    total, pos, 100 * pos / total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "cleaned_df = raw_df.copy()\n",
    "\n",
    "# You don't want the `Time` column.\n",
    "cleaned_df.pop('Time')\n",
    "\n",
    "# The `Amount` column covers a huge range. Convert to log-space.\n",
    "eps = 0.001 # 0 => 0.1¢\n",
    "cleaned_df['Log Ammount'] = np.log(cleaned_df.pop('Amount')+eps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use a utility from sklearn to split and shuffle your dataset.\n",
    "train_df, test_df = train_test_split(cleaned_df, test_size=0.2)\n",
    "train_df, val_df = train_test_split(train_df, test_size=0.2)\n",
    "\n",
    "# Form np arrays of labels and features.\n",
    "train_labels = np.array(train_df.pop('Class'))\n",
    "bool_train_labels = train_labels != 0\n",
    "val_labels = np.array(val_df.pop('Class'))\n",
    "test_labels = np.array(test_df.pop('Class'))\n",
    "\n",
    "train_features = np.array(train_df)\n",
    "val_features = np.array(val_df)\n",
    "test_features = np.array(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training labels shape: (182276,)\n",
      "Validation labels shape: (45569,)\n",
      "Test labels shape: (56962,)\n",
      "Training features shape: (182276, 29)\n",
      "Validation features shape: (45569, 29)\n",
      "Test features shape: (56962, 29)\n"
     ]
    }
   ],
   "source": [
    "scaler = StandardScaler()\n",
    "train_features = scaler.fit_transform(train_features)\n",
    "\n",
    "val_features = scaler.transform(val_features)\n",
    "test_features = scaler.transform(test_features)\n",
    "\n",
    "train_features = np.clip(train_features, -5, 5)\n",
    "val_features = np.clip(val_features, -5, 5)\n",
    "test_features = np.clip(test_features, -5, 5)\n",
    "\n",
    "\n",
    "print('Training labels shape:', train_labels.shape)\n",
    "print('Validation labels shape:', val_labels.shape)\n",
    "print('Test labels shape:', test_labels.shape)\n",
    "\n",
    "print('Training features shape:', train_features.shape)\n",
    "print('Validation features shape:', val_features.shape)\n",
    "print('Test features shape:', test_features.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/johnnydevriese/miniforge3/envs/pytorch_m1/lib/python3.8/site-packages/seaborn/_decorators.py:36: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n",
      "/Users/johnnydevriese/miniforge3/envs/pytorch_m1/lib/python3.8/site-packages/seaborn/_decorators.py:36: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pos_df = pd.DataFrame(train_features[ bool_train_labels], columns=train_df.columns)\n",
    "neg_df = pd.DataFrame(train_features[~bool_train_labels], columns=train_df.columns)\n",
    "\n",
    "sns.jointplot(pos_df['V5'], pos_df['V6'],\n",
    "              kind='hex', xlim=(-5,5), ylim=(-5,5))\n",
    "plt.suptitle(\"Positive distribution\")\n",
    "\n",
    "sns.jointplot(neg_df['V5'], neg_df['V6'],\n",
    "              kind='hex', xlim=(-5,5), ylim=(-5,5))\n",
    "_ = plt.suptitle(\"Negative distribution\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "METRICS = [\n",
    "      keras.metrics.TruePositives(name='true positive'),\n",
    "      keras.metrics.FalsePositives(name='false positive'),\n",
    "      keras.metrics.TrueNegatives(name='true negative'),\n",
    "      keras.metrics.FalseNegatives(name='false negative'), \n",
    "      keras.metrics.BinaryAccuracy(name='accuracy'),\n",
    "      keras.metrics.Precision(name='precision'),\n",
    "      keras.metrics.Recall(name='recall'),\n",
    "      keras.metrics.AUC(name='auc'),\n",
    "      keras.metrics.AUC(name='prc', curve='PR'), # precision-recall curve\n",
    "]\n",
    "\n",
    "def make_model(metrics=METRICS, output_bias=None):\n",
    "  if output_bias is not None:\n",
    "    output_bias = tf.keras.initializers.Constant(output_bias)\n",
    "  model = keras.Sequential([\n",
    "      keras.layers.Dense(\n",
    "          16, activation='relu',\n",
    "          input_shape=(train_features.shape[-1],)),\n",
    "      keras.layers.Dropout(0.5),\n",
    "      keras.layers.Dense(1, activation='sigmoid',\n",
    "                         bias_initializer=output_bias),\n",
    "  ])\n",
    "\n",
    "  model.compile(\n",
    "      optimizer=keras.optimizers.Adam(learning_rate=1e-3),\n",
    "      loss=keras.losses.BinaryCrossentropy(),\n",
    "      metrics=metrics)\n",
    "\n",
    "  return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Baseline Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "EPOCHS = 100\n",
    "BATCH_SIZE = 2048\n",
    "\n",
    "early_stopping = tf.keras.callbacks.EarlyStopping(\n",
    "    monitor='val_prc', \n",
    "    verbose=1,\n",
    "    patience=10,\n",
    "    mode='max',\n",
    "    restore_best_weights=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 16)                480       \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 497\n",
      "Trainable params: 497\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Create and train the model by calling the `make_model()` function.\n",
    "model = make_model()\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-01-05 15:45:40.957346: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n",
      "2022-01-05 15:45:40.957552: W tensorflow/core/platform/profile_utils/cpu_utils.cc:128] Failed to get CPU frequency: 0 Hz\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.5718938 ],\n",
       "       [0.9578531 ],\n",
       "       [0.48555294],\n",
       "       [0.38280708],\n",
       "       [0.39884338],\n",
       "       [0.18077362],\n",
       "       [0.3778038 ],\n",
       "       [0.85698396],\n",
       "       [0.5097802 ],\n",
       "       [0.41451403]], dtype=float32)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(train_features[:10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Set better initial bias "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.9362\n"
     ]
    }
   ],
   "source": [
    "results = model.evaluate(train_features, train_labels, batch_size=BATCH_SIZE, verbose=0)\n",
    "print(\"Loss: {:0.4f}\".format(results[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The correct bias to set can be derived from:\n",
    "\n",
    "$$ p_0 = pos/(pos + neg) = 1/(1+e^{-b_0}) $$\n",
    "$$ b_0 = -log_e(1/p_0 - 1) $$\n",
    "$$ b_0 = log_e(pos/neg)$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-6.35935934])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "initial_bias = np.log([pos/neg])\n",
    "initial_bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.00228986],\n",
       "       [0.00316817],\n",
       "       [0.00424775],\n",
       "       [0.00109661],\n",
       "       [0.00135592],\n",
       "       [0.00243115],\n",
       "       [0.00286371],\n",
       "       [0.00086367],\n",
       "       [0.00095445],\n",
       "       [0.00517303]], dtype=float32)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = make_model(output_bias=initial_bias)\n",
    "model.predict(train_features[:10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0160\n"
     ]
    }
   ],
   "source": [
    "results = model.evaluate(train_features, train_labels, batch_size=BATCH_SIZE, verbose=0)\n",
    "print(\"Loss: {:0.4f}\".format(results[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The initial loss is now 50x less than with naive initilization! "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# checkpoint weights\n",
    "\n",
    "initial_weights = os.path.join(tempfile.mkdtemp(), 'initial_weights')\n",
    "model.save_weights(initial_weights)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# confirm that bias init has helped "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = make_model()\n",
    "model.load_weights(initial_weights)\n",
    "model.layers[-1].bias.assign([0.0])\n",
    "zero_bias_history = model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    epochs=20,\n",
    "    validation_data=(val_features, val_labels), \n",
    "    verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = make_model()\n",
    "model.load_weights(initial_weights)\n",
    "careful_bias_history = model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    epochs=20,\n",
    "    validation_data=(val_features, val_labels), \n",
    "    verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_loss(history, label, n):\n",
    "  # Use a log scale on y-axis to show the wide range of values.\n",
    "  plt.semilogy(history.epoch, history.history['loss'],\n",
    "               color=colors[n], label='Train ' + label)\n",
    "  plt.semilogy(history.epoch, history.history['val_loss'],\n",
    "               color=colors[n], label='Val ' + label,\n",
    "               linestyle=\"--\")\n",
    "  plt.xlabel('Epoch')\n",
    "  plt.ylabel('Loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(zero_bias_history, \"Zero Bias\", 0)\n",
    "plot_loss(careful_bias_history, \"Careful Bias\", 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "90/90 [==============================] - 1s 4ms/step - loss: 0.0138 - true positive: 85.0000 - false positive: 41.0000 - true negative: 227411.0000 - false negative: 308.0000 - accuracy: 0.9985 - precision: 0.6746 - recall: 0.2163 - auc: 0.7032 - prc: 0.2490 - val_loss: 0.0086 - val_true positive: 1.0000 - val_false positive: 0.0000e+00 - val_true negative: 45487.0000 - val_false negative: 81.0000 - val_accuracy: 0.9982 - val_precision: 1.0000 - val_recall: 0.0122 - val_auc: 0.8469 - val_prc: 0.5781\n",
      "Epoch 2/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0092 - true positive: 78.0000 - false positive: 26.0000 - true negative: 181939.0000 - false negative: 233.0000 - accuracy: 0.9986 - precision: 0.7500 - recall: 0.2508 - auc: 0.7446 - prc: 0.3068 - val_loss: 0.0061 - val_true positive: 22.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 60.0000 - val_accuracy: 0.9986 - val_precision: 0.8462 - val_recall: 0.2683 - val_auc: 0.9023 - val_prc: 0.7164\n",
      "Epoch 3/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0072 - true positive: 124.0000 - false positive: 25.0000 - true negative: 181940.0000 - false negative: 187.0000 - accuracy: 0.9988 - precision: 0.8322 - recall: 0.3987 - auc: 0.8571 - prc: 0.4579 - val_loss: 0.0048 - val_true positive: 39.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 43.0000 - val_accuracy: 0.9990 - val_precision: 0.9070 - val_recall: 0.4756 - val_auc: 0.9144 - val_prc: 0.7398\n",
      "Epoch 4/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0066 - true positive: 130.0000 - false positive: 28.0000 - true negative: 181937.0000 - false negative: 181.0000 - accuracy: 0.9989 - precision: 0.8228 - recall: 0.4180 - auc: 0.8745 - prc: 0.5376 - val_loss: 0.0042 - val_true positive: 50.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 32.0000 - val_accuracy: 0.9992 - val_precision: 0.9259 - val_recall: 0.6098 - val_auc: 0.9205 - val_prc: 0.7593\n",
      "Epoch 5/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0060 - true positive: 152.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 159.0000 - accuracy: 0.9990 - precision: 0.8306 - recall: 0.4887 - auc: 0.8901 - prc: 0.6027 - val_loss: 0.0040 - val_true positive: 50.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 32.0000 - val_accuracy: 0.9992 - val_precision: 0.9259 - val_recall: 0.6098 - val_auc: 0.9205 - val_prc: 0.7783\n",
      "Epoch 6/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0059 - true positive: 148.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 163.0000 - accuracy: 0.9989 - precision: 0.8222 - recall: 0.4759 - auc: 0.8873 - prc: 0.5930 - val_loss: 0.0038 - val_true positive: 50.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 32.0000 - val_accuracy: 0.9992 - val_precision: 0.9259 - val_recall: 0.6098 - val_auc: 0.9205 - val_prc: 0.7827\n",
      "Epoch 7/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0058 - true positive: 154.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 157.0000 - accuracy: 0.9990 - precision: 0.8191 - recall: 0.4952 - auc: 0.8923 - prc: 0.5795 - val_loss: 0.0037 - val_true positive: 50.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 32.0000 - val_accuracy: 0.9992 - val_precision: 0.9259 - val_recall: 0.6098 - val_auc: 0.9205 - val_prc: 0.7922\n",
      "Epoch 8/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0054 - true positive: 156.0000 - false positive: 27.0000 - true negative: 181938.0000 - false negative: 155.0000 - accuracy: 0.9990 - precision: 0.8525 - recall: 0.5016 - auc: 0.9006 - prc: 0.6241 - val_loss: 0.0035 - val_true positive: 52.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 30.0000 - val_accuracy: 0.9993 - val_precision: 0.9286 - val_recall: 0.6341 - val_auc: 0.9205 - val_prc: 0.7938\n",
      "Epoch 9/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0051 - true positive: 157.0000 - false positive: 30.0000 - true negative: 181935.0000 - false negative: 154.0000 - accuracy: 0.9990 - precision: 0.8396 - recall: 0.5048 - auc: 0.9006 - prc: 0.6400 - val_loss: 0.0034 - val_true positive: 55.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 27.0000 - val_accuracy: 0.9993 - val_precision: 0.9322 - val_recall: 0.6707 - val_auc: 0.9265 - val_prc: 0.7984\n",
      "Epoch 10/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0051 - true positive: 172.0000 - false positive: 30.0000 - true negative: 181935.0000 - false negative: 139.0000 - accuracy: 0.9991 - precision: 0.8515 - recall: 0.5531 - auc: 0.9104 - prc: 0.6561 - val_loss: 0.0033 - val_true positive: 54.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 28.0000 - val_accuracy: 0.9993 - val_precision: 0.9310 - val_recall: 0.6585 - val_auc: 0.9326 - val_prc: 0.8092\n",
      "Epoch 11/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0049 - true positive: 157.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 154.0000 - accuracy: 0.9990 - precision: 0.8351 - recall: 0.5048 - auc: 0.9089 - prc: 0.6437 - val_loss: 0.0032 - val_true positive: 57.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 25.0000 - val_accuracy: 0.9994 - val_precision: 0.9344 - val_recall: 0.6951 - val_auc: 0.9266 - val_prc: 0.8052\n",
      "Epoch 12/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0050 - true positive: 169.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 142.0000 - accuracy: 0.9990 - precision: 0.8325 - recall: 0.5434 - auc: 0.9024 - prc: 0.6560 - val_loss: 0.0032 - val_true positive: 55.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 27.0000 - val_accuracy: 0.9993 - val_precision: 0.9322 - val_recall: 0.6707 - val_auc: 0.9205 - val_prc: 0.8066\n",
      "Epoch 13/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0052 - true positive: 149.0000 - false positive: 33.0000 - true negative: 181932.0000 - false negative: 162.0000 - accuracy: 0.9989 - precision: 0.8187 - recall: 0.4791 - auc: 0.8911 - prc: 0.6368 - val_loss: 0.0032 - val_true positive: 56.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 26.0000 - val_accuracy: 0.9993 - val_precision: 0.9333 - val_recall: 0.6829 - val_auc: 0.9266 - val_prc: 0.8133\n",
      "Epoch 14/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0051 - true positive: 164.0000 - false positive: 37.0000 - true negative: 181928.0000 - false negative: 147.0000 - accuracy: 0.9990 - precision: 0.8159 - recall: 0.5273 - auc: 0.8912 - prc: 0.6336 - val_loss: 0.0031 - val_true positive: 56.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 26.0000 - val_accuracy: 0.9993 - val_precision: 0.9333 - val_recall: 0.6829 - val_auc: 0.9326 - val_prc: 0.8187\n",
      "Epoch 15/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0046 - true positive: 173.0000 - false positive: 33.0000 - true negative: 181932.0000 - false negative: 138.0000 - accuracy: 0.9991 - precision: 0.8398 - recall: 0.5563 - auc: 0.9041 - prc: 0.6660 - val_loss: 0.0031 - val_true positive: 57.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 25.0000 - val_accuracy: 0.9994 - val_precision: 0.9344 - val_recall: 0.6951 - val_auc: 0.9327 - val_prc: 0.8207\n",
      "Epoch 16/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0049 - true positive: 165.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 146.0000 - accuracy: 0.9990 - precision: 0.8418 - recall: 0.5305 - auc: 0.9040 - prc: 0.6574 - val_loss: 0.0030 - val_true positive: 55.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 27.0000 - val_accuracy: 0.9993 - val_precision: 0.9322 - val_recall: 0.6707 - val_auc: 0.9327 - val_prc: 0.8243\n",
      "Epoch 17/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0048 - true positive: 162.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 149.0000 - accuracy: 0.9990 - precision: 0.8351 - recall: 0.5209 - auc: 0.9058 - prc: 0.6674 - val_loss: 0.0030 - val_true positive: 54.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 28.0000 - val_accuracy: 0.9993 - val_precision: 0.9474 - val_recall: 0.6585 - val_auc: 0.9266 - val_prc: 0.8232\n",
      "Epoch 18/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0045 - true positive: 170.0000 - false positive: 33.0000 - true negative: 181932.0000 - false negative: 141.0000 - accuracy: 0.9990 - precision: 0.8374 - recall: 0.5466 - auc: 0.9122 - prc: 0.6770 - val_loss: 0.0029 - val_true positive: 59.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 23.0000 - val_accuracy: 0.9994 - val_precision: 0.9516 - val_recall: 0.7195 - val_auc: 0.9327 - val_prc: 0.8282\n",
      "Epoch 19/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0044 - true positive: 168.0000 - false positive: 27.0000 - true negative: 181938.0000 - false negative: 143.0000 - accuracy: 0.9991 - precision: 0.8615 - recall: 0.5402 - auc: 0.9187 - prc: 0.6887 - val_loss: 0.0028 - val_true positive: 59.0000 - val_false positive: 5.0000 - val_true negative: 45482.0000 - val_false negative: 23.0000 - val_accuracy: 0.9994 - val_precision: 0.9219 - val_recall: 0.7195 - val_auc: 0.9327 - val_prc: 0.8284\n",
      "Epoch 20/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0049 - true positive: 155.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 156.0000 - accuracy: 0.9990 - precision: 0.8201 - recall: 0.4984 - auc: 0.9072 - prc: 0.6393 - val_loss: 0.0028 - val_true positive: 60.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 22.0000 - val_accuracy: 0.9994 - val_precision: 0.9375 - val_recall: 0.7317 - val_auc: 0.9327 - val_prc: 0.8291\n",
      "Epoch 21/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0049 - true positive: 147.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 164.0000 - accuracy: 0.9989 - precision: 0.8122 - recall: 0.4727 - auc: 0.9007 - prc: 0.6365 - val_loss: 0.0028 - val_true positive: 61.0000 - val_false positive: 5.0000 - val_true negative: 45482.0000 - val_false negative: 21.0000 - val_accuracy: 0.9994 - val_precision: 0.9242 - val_recall: 0.7439 - val_auc: 0.9326 - val_prc: 0.8283\n",
      "Epoch 22/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0046 - true positive: 173.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 138.0000 - accuracy: 0.9991 - precision: 0.8439 - recall: 0.5563 - auc: 0.8959 - prc: 0.6617 - val_loss: 0.0028 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9326 - val_prc: 0.8286\n",
      "Epoch 23/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0046 - true positive: 175.0000 - false positive: 36.0000 - true negative: 181929.0000 - false negative: 136.0000 - accuracy: 0.9991 - precision: 0.8294 - recall: 0.5627 - auc: 0.9187 - prc: 0.6665 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9327 - val_prc: 0.8308\n",
      "Epoch 24/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0042 - true positive: 180.0000 - false positive: 30.0000 - true negative: 181935.0000 - false negative: 131.0000 - accuracy: 0.9991 - precision: 0.8571 - recall: 0.5788 - auc: 0.9268 - prc: 0.7088 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9531 - val_recall: 0.7439 - val_auc: 0.9327 - val_prc: 0.8330\n",
      "Epoch 25/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0045 - true positive: 174.0000 - false positive: 40.0000 - true negative: 181925.0000 - false negative: 137.0000 - accuracy: 0.9990 - precision: 0.8131 - recall: 0.5595 - auc: 0.9282 - prc: 0.6740 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9326 - val_prc: 0.8312\n",
      "Epoch 26/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0044 - true positive: 185.0000 - false positive: 33.0000 - true negative: 181932.0000 - false negative: 126.0000 - accuracy: 0.9991 - precision: 0.8486 - recall: 0.5949 - auc: 0.9234 - prc: 0.6859 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9531 - val_recall: 0.7439 - val_auc: 0.9327 - val_prc: 0.8324\n",
      "Epoch 27/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 176.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 135.0000 - accuracy: 0.9991 - precision: 0.8462 - recall: 0.5659 - auc: 0.9219 - prc: 0.7069 - val_loss: 0.0027 - val_true positive: 60.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 22.0000 - val_accuracy: 0.9995 - val_precision: 0.9524 - val_recall: 0.7317 - val_auc: 0.9327 - val_prc: 0.8333\n",
      "Epoch 28/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 180.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 131.0000 - accuracy: 0.9991 - precision: 0.8491 - recall: 0.5788 - auc: 0.9284 - prc: 0.7011 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9387 - val_prc: 0.8376\n",
      "Epoch 29/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0046 - true positive: 169.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 142.0000 - accuracy: 0.9990 - precision: 0.8408 - recall: 0.5434 - auc: 0.9076 - prc: 0.6798 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9531 - val_recall: 0.7439 - val_auc: 0.9327 - val_prc: 0.8328\n",
      "Epoch 30/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0040 - true positive: 183.0000 - false positive: 28.0000 - true negative: 181937.0000 - false negative: 128.0000 - accuracy: 0.9991 - precision: 0.8673 - recall: 0.5884 - auc: 0.9334 - prc: 0.7300 - val_loss: 0.0027 - val_true positive: 61.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9531 - val_recall: 0.7439 - val_auc: 0.9387 - val_prc: 0.8372\n",
      "Epoch 31/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0044 - true positive: 178.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 133.0000 - accuracy: 0.9991 - precision: 0.8396 - recall: 0.5723 - auc: 0.9122 - prc: 0.6779 - val_loss: 0.0026 - val_true positive: 61.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9531 - val_recall: 0.7439 - val_auc: 0.9387 - val_prc: 0.8372\n",
      "Epoch 32/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0038 - true positive: 196.0000 - false positive: 28.0000 - true negative: 181937.0000 - false negative: 115.0000 - accuracy: 0.9992 - precision: 0.8750 - recall: 0.6302 - auc: 0.9333 - prc: 0.7405 - val_loss: 0.0026 - val_true positive: 57.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 25.0000 - val_accuracy: 0.9994 - val_precision: 0.9500 - val_recall: 0.6951 - val_auc: 0.9387 - val_prc: 0.8414\n",
      "Epoch 33/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0042 - true positive: 172.0000 - false positive: 29.0000 - true negative: 181936.0000 - false negative: 139.0000 - accuracy: 0.9991 - precision: 0.8557 - recall: 0.5531 - auc: 0.9234 - prc: 0.7009 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9394 - val_recall: 0.7561 - val_auc: 0.9387 - val_prc: 0.8396\n",
      "Epoch 34/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0039 - true positive: 186.0000 - false positive: 30.0000 - true negative: 181935.0000 - false negative: 125.0000 - accuracy: 0.9991 - precision: 0.8611 - recall: 0.5981 - auc: 0.9236 - prc: 0.7301 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9538 - val_recall: 0.7561 - val_auc: 0.9387 - val_prc: 0.8408\n",
      "Epoch 35/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 189.0000 - false positive: 33.0000 - true negative: 181932.0000 - false negative: 122.0000 - accuracy: 0.9991 - precision: 0.8514 - recall: 0.6077 - auc: 0.9204 - prc: 0.7083 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9538 - val_recall: 0.7561 - val_auc: 0.9387 - val_prc: 0.8425\n",
      "Epoch 36/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 192.0000 - false positive: 27.0000 - true negative: 181938.0000 - false negative: 119.0000 - accuracy: 0.9992 - precision: 0.8767 - recall: 0.6174 - auc: 0.9187 - prc: 0.7068 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9538 - val_recall: 0.7561 - val_auc: 0.9387 - val_prc: 0.8418\n",
      "Epoch 37/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 180.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 131.0000 - accuracy: 0.9991 - precision: 0.8411 - recall: 0.5788 - auc: 0.9284 - prc: 0.7236 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9538 - val_recall: 0.7561 - val_auc: 0.9387 - val_prc: 0.8423\n",
      "Epoch 38/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 181.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 130.0000 - accuracy: 0.9991 - precision: 0.8419 - recall: 0.5820 - auc: 0.9284 - prc: 0.7024 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9538 - val_recall: 0.7561 - val_auc: 0.9387 - val_prc: 0.8438\n",
      "Epoch 39/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0040 - true positive: 183.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 128.0000 - accuracy: 0.9991 - precision: 0.8433 - recall: 0.5884 - auc: 0.9203 - prc: 0.7192 - val_loss: 0.0026 - val_true positive: 58.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 24.0000 - val_accuracy: 0.9994 - val_precision: 0.9508 - val_recall: 0.7073 - val_auc: 0.9387 - val_prc: 0.8440\n",
      "Epoch 40/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0039 - true positive: 194.0000 - false positive: 26.0000 - true negative: 181939.0000 - false negative: 117.0000 - accuracy: 0.9992 - precision: 0.8818 - recall: 0.6238 - auc: 0.9139 - prc: 0.7160 - val_loss: 0.0026 - val_true positive: 59.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 23.0000 - val_accuracy: 0.9994 - val_precision: 0.9516 - val_recall: 0.7195 - val_auc: 0.9387 - val_prc: 0.8440\n",
      "Epoch 41/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0039 - true positive: 184.0000 - false positive: 33.0000 - true negative: 181932.0000 - false negative: 127.0000 - accuracy: 0.9991 - precision: 0.8479 - recall: 0.5916 - auc: 0.9235 - prc: 0.7267 - val_loss: 0.0026 - val_true positive: 65.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9420 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8492\n",
      "Epoch 42/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0038 - true positive: 192.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 119.0000 - accuracy: 0.9992 - precision: 0.8571 - recall: 0.6174 - auc: 0.9365 - prc: 0.7470 - val_loss: 0.0026 - val_true positive: 65.0000 - val_false positive: 5.0000 - val_true negative: 45482.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9286 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8479\n",
      "Epoch 43/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0039 - true positive: 186.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 125.0000 - accuracy: 0.9991 - precision: 0.8571 - recall: 0.5981 - auc: 0.9299 - prc: 0.7197 - val_loss: 0.0026 - val_true positive: 63.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 19.0000 - val_accuracy: 0.9995 - val_precision: 0.9545 - val_recall: 0.7683 - val_auc: 0.9448 - val_prc: 0.8502\n",
      "Epoch 44/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 176.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 135.0000 - accuracy: 0.9991 - precision: 0.8502 - recall: 0.5659 - auc: 0.9155 - prc: 0.7109 - val_loss: 0.0026 - val_true positive: 63.0000 - val_false positive: 3.0000 - val_true negative: 45484.0000 - val_false negative: 19.0000 - val_accuracy: 0.9995 - val_precision: 0.9545 - val_recall: 0.7683 - val_auc: 0.9448 - val_prc: 0.8501\n",
      "Epoch 45/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0038 - true positive: 187.0000 - false positive: 26.0000 - true negative: 181939.0000 - false negative: 124.0000 - accuracy: 0.9992 - precision: 0.8779 - recall: 0.6013 - auc: 0.9380 - prc: 0.7407 - val_loss: 0.0026 - val_true positive: 66.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 16.0000 - val_accuracy: 0.9996 - val_precision: 0.9429 - val_recall: 0.8049 - val_auc: 0.9448 - val_prc: 0.8486\n",
      "Epoch 46/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 202.0000 - false positive: 28.0000 - true negative: 181937.0000 - false negative: 109.0000 - accuracy: 0.9992 - precision: 0.8783 - recall: 0.6495 - auc: 0.9348 - prc: 0.7476 - val_loss: 0.0026 - val_true positive: 64.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 18.0000 - val_accuracy: 0.9995 - val_precision: 0.9412 - val_recall: 0.7805 - val_auc: 0.9448 - val_prc: 0.8490\n",
      "Epoch 47/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 187.0000 - false positive: 35.0000 - true negative: 181930.0000 - false negative: 124.0000 - accuracy: 0.9991 - precision: 0.8423 - recall: 0.6013 - auc: 0.9397 - prc: 0.7424 - val_loss: 0.0026 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9448 - val_prc: 0.8481\n",
      "Epoch 48/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0040 - true positive: 183.0000 - false positive: 30.0000 - true negative: 181935.0000 - false negative: 128.0000 - accuracy: 0.9991 - precision: 0.8592 - recall: 0.5884 - auc: 0.9268 - prc: 0.7195 - val_loss: 0.0026 - val_true positive: 64.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 18.0000 - val_accuracy: 0.9995 - val_precision: 0.9412 - val_recall: 0.7805 - val_auc: 0.9448 - val_prc: 0.8497\n",
      "Epoch 49/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0041 - true positive: 182.0000 - false positive: 35.0000 - true negative: 181930.0000 - false negative: 129.0000 - accuracy: 0.9991 - precision: 0.8387 - recall: 0.5852 - auc: 0.9283 - prc: 0.6969 - val_loss: 0.0026 - val_true positive: 65.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9420 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8478\n",
      "Epoch 50/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0039 - true positive: 192.0000 - false positive: 40.0000 - true negative: 181925.0000 - false negative: 119.0000 - accuracy: 0.9991 - precision: 0.8276 - recall: 0.6174 - auc: 0.9317 - prc: 0.7317 - val_loss: 0.0026 - val_true positive: 60.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 22.0000 - val_accuracy: 0.9994 - val_precision: 0.9375 - val_recall: 0.7317 - val_auc: 0.9448 - val_prc: 0.8488\n",
      "Epoch 51/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0038 - true positive: 184.0000 - false positive: 25.0000 - true negative: 181940.0000 - false negative: 127.0000 - accuracy: 0.9992 - precision: 0.8804 - recall: 0.5916 - auc: 0.9333 - prc: 0.7476 - val_loss: 0.0026 - val_true positive: 63.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 19.0000 - val_accuracy: 0.9995 - val_precision: 0.9403 - val_recall: 0.7683 - val_auc: 0.9388 - val_prc: 0.8465\n",
      "Epoch 52/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0040 - true positive: 189.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 122.0000 - accuracy: 0.9992 - precision: 0.8591 - recall: 0.6077 - auc: 0.9267 - prc: 0.7210 - val_loss: 0.0026 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9448 - val_prc: 0.8505\n",
      "Epoch 53/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0036 - true positive: 194.0000 - false positive: 22.0000 - true negative: 181943.0000 - false negative: 117.0000 - accuracy: 0.9992 - precision: 0.8981 - recall: 0.6238 - auc: 0.9365 - prc: 0.7656 - val_loss: 0.0025 - val_true positive: 67.0000 - val_false positive: 5.0000 - val_true negative: 45482.0000 - val_false negative: 15.0000 - val_accuracy: 0.9996 - val_precision: 0.9306 - val_recall: 0.8171 - val_auc: 0.9448 - val_prc: 0.8505\n",
      "Epoch 54/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0039 - true positive: 201.0000 - false positive: 32.0000 - true negative: 181933.0000 - false negative: 110.0000 - accuracy: 0.9992 - precision: 0.8627 - recall: 0.6463 - auc: 0.9349 - prc: 0.7285 - val_loss: 0.0026 - val_true positive: 63.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 19.0000 - val_accuracy: 0.9995 - val_precision: 0.9403 - val_recall: 0.7683 - val_auc: 0.9448 - val_prc: 0.8492\n",
      "Epoch 55/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 204.0000 - false positive: 29.0000 - true negative: 181936.0000 - false negative: 107.0000 - accuracy: 0.9993 - precision: 0.8755 - recall: 0.6559 - auc: 0.9366 - prc: 0.7507 - val_loss: 0.0026 - val_true positive: 63.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 19.0000 - val_accuracy: 0.9995 - val_precision: 0.9403 - val_recall: 0.7683 - val_auc: 0.9448 - val_prc: 0.8519\n",
      "Epoch 56/100\n",
      "90/90 [==============================] - 0s 2ms/step - loss: 0.0038 - true positive: 191.0000 - false positive: 26.0000 - true negative: 181939.0000 - false negative: 120.0000 - accuracy: 0.9992 - precision: 0.8802 - recall: 0.6141 - auc: 0.9268 - prc: 0.7399 - val_loss: 0.0025 - val_true positive: 65.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9420 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8501\n",
      "Epoch 57/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0036 - true positive: 197.0000 - false positive: 25.0000 - true negative: 181940.0000 - false negative: 114.0000 - accuracy: 0.9992 - precision: 0.8874 - recall: 0.6334 - auc: 0.9413 - prc: 0.7611 - val_loss: 0.0025 - val_true positive: 65.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9420 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8495\n",
      "Epoch 58/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 205.0000 - false positive: 34.0000 - true negative: 181931.0000 - false negative: 106.0000 - accuracy: 0.9992 - precision: 0.8577 - recall: 0.6592 - auc: 0.9348 - prc: 0.7422 - val_loss: 0.0026 - val_true positive: 63.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 19.0000 - val_accuracy: 0.9995 - val_precision: 0.9403 - val_recall: 0.7683 - val_auc: 0.9448 - val_prc: 0.8500\n",
      "Epoch 59/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 192.0000 - false positive: 27.0000 - true negative: 181938.0000 - false negative: 119.0000 - accuracy: 0.9992 - precision: 0.8767 - recall: 0.6174 - auc: 0.9316 - prc: 0.7437 - val_loss: 0.0025 - val_true positive: 65.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9420 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8501\n",
      "Epoch 60/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 193.0000 - false positive: 30.0000 - true negative: 181935.0000 - false negative: 118.0000 - accuracy: 0.9992 - precision: 0.8655 - recall: 0.6206 - auc: 0.9349 - prc: 0.7384 - val_loss: 0.0025 - val_true positive: 64.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 18.0000 - val_accuracy: 0.9995 - val_precision: 0.9412 - val_recall: 0.7805 - val_auc: 0.9448 - val_prc: 0.8501\n",
      "Epoch 61/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0035 - true positive: 200.0000 - false positive: 27.0000 - true negative: 181938.0000 - false negative: 111.0000 - accuracy: 0.9992 - precision: 0.8811 - recall: 0.6431 - auc: 0.9350 - prc: 0.7603 - val_loss: 0.0025 - val_true positive: 65.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9420 - val_recall: 0.7927 - val_auc: 0.9387 - val_prc: 0.8458\n",
      "Epoch 62/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0038 - true positive: 190.0000 - false positive: 36.0000 - true negative: 181929.0000 - false negative: 121.0000 - accuracy: 0.9991 - precision: 0.8407 - recall: 0.6109 - auc: 0.9381 - prc: 0.7410 - val_loss: 0.0026 - val_true positive: 62.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 20.0000 - val_accuracy: 0.9995 - val_precision: 0.9394 - val_recall: 0.7561 - val_auc: 0.9448 - val_prc: 0.8510\n",
      "Epoch 63/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0036 - true positive: 194.0000 - false positive: 26.0000 - true negative: 181939.0000 - false negative: 117.0000 - accuracy: 0.9992 - precision: 0.8818 - recall: 0.6238 - auc: 0.9350 - prc: 0.7508 - val_loss: 0.0025 - val_true positive: 65.0000 - val_false positive: 5.0000 - val_true negative: 45482.0000 - val_false negative: 17.0000 - val_accuracy: 0.9995 - val_precision: 0.9286 - val_recall: 0.7927 - val_auc: 0.9448 - val_prc: 0.8507\n",
      "Epoch 64/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0037 - true positive: 188.0000 - false positive: 31.0000 - true negative: 181934.0000 - false negative: 123.0000 - accuracy: 0.9992 - precision: 0.8584 - recall: 0.6045 - auc: 0.9334 - prc: 0.7449 - val_loss: 0.0026 - val_true positive: 66.0000 - val_false positive: 5.0000 - val_true negative: 45482.0000 - val_false negative: 16.0000 - val_accuracy: 0.9995 - val_precision: 0.9296 - val_recall: 0.8049 - val_auc: 0.9448 - val_prc: 0.8510\n",
      "Epoch 65/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.0034 - true positive: 206.0000 - false positive: 27.0000 - true negative: 181938.0000 - false negative: 105.0000 - accuracy: 0.9993 - precision: 0.8841 - recall: 0.6624 - auc: 0.9382 - prc: 0.7614 - val_loss: 0.0026 - val_true positive: 61.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 21.0000 - val_accuracy: 0.9995 - val_precision: 0.9385 - val_recall: 0.7439 - val_auc: 0.9448 - val_prc: 0.8518\n",
      "Restoring model weights from the end of the best epoch.\n",
      "Epoch 00065: early stopping\n"
     ]
    }
   ],
   "source": [
    "model = make_model()\n",
    "model.load_weights(initial_weights)\n",
    "baseline_history = model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    epochs=EPOCHS,\n",
    "    callbacks=[early_stopping],\n",
    "    validation_data=(val_features, val_labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Check training history "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_metrics(history):\n",
    "  metrics = ['loss', 'prc', 'precision', 'recall']\n",
    "  for n, metric in enumerate(metrics):\n",
    "    name = metric.replace(\"_\",\" \").capitalize()\n",
    "    plt.subplot(2,2,n+1)\n",
    "    plt.plot(history.epoch, history.history[metric], color=colors[0], label='Train')\n",
    "    plt.plot(history.epoch, history.history['val_'+metric],\n",
    "             color=colors[0], linestyle=\"--\", label='Val')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel(name)\n",
    "    if metric == 'loss':\n",
    "      plt.ylim([0, plt.ylim()[1]])\n",
    "    elif metric == 'auc':\n",
    "      plt.ylim([0.8,1])\n",
    "    else:\n",
    "      plt.ylim([0,1])\n",
    "\n",
    "    plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_metrics(baseline_history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Diagnostic Metrics "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_predictions_baseline = model.predict(train_features, batch_size=BATCH_SIZE)\n",
    "test_predictions_baseline = model.predict(test_features, batch_size=BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_cm(labels, predictions, p=0.5):\n",
    "  cm = confusion_matrix(labels, predictions > p)\n",
    "  plt.figure(figsize=(5,5))\n",
    "  sns.heatmap(cm, annot=True, fmt=\"d\")\n",
    "  plt.title('Confusion matrix @{:.2f}'.format(p))\n",
    "  plt.ylabel('Actual label')\n",
    "  plt.xlabel('Predicted label')\n",
    "\n",
    "  print('Legitimate Transactions Detected (True Negatives): ', cm[0][0])\n",
    "  print('Legitimate Transactions Incorrectly Detected (False Positives): ', cm[0][1])\n",
    "  print('Fraudulent Transactions Missed (False Negatives): ', cm[1][0])\n",
    "  print('Fraudulent Transactions Detected (True Positives): ', cm[1][1])\n",
    "  print('Total Fraudulent Transactions: ', np.sum(cm[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss :  0.002845881273970008\n",
      "true positive :  76.0\n",
      "false positive :  8.0\n",
      "true negative :  56855.0\n",
      "false negative :  23.0\n",
      "accuracy :  0.9994557499885559\n",
      "precision :  0.9047619104385376\n",
      "recall :  0.7676767706871033\n",
      "auc :  0.9340019822120667\n",
      "prc :  0.8226545453071594\n",
      "\n",
      "Legitimate Transactions Detected (True Negatives):  56855\n",
      "Legitimate Transactions Incorrectly Detected (False Positives):  8\n",
      "Fraudulent Transactions Missed (False Negatives):  23\n",
      "Fraudulent Transactions Detected (True Positives):  76\n",
      "Total Fraudulent Transactions:  99\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU8AAAFNCAYAAACABe35AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAklUlEQVR4nO3debxd873/8dc7owiJSCSNxBBNUPTiUk1NRVpCq6GlYiitEENUe6vX0IHS3pZf62q1hgatOYSaSk2Nmq4hA9FIRIVoZCAiQYQMJ+fz+2N9T+wcZ1z2Pvts+/30WI+s/V1rfffnnON8zndY67sVEZiZWet0KHcAZmaVyMnTzCwHJ08zsxycPM3McnDyNDPLwcnTzCwHJ08zsxycPNshSd0k/VXSO5Ju+Rj1HCnpgWLGVi6S9pD0YrnjMKvj5PkxSDpC0mRJ70laIOleSbsXoepDgH5A74g4NG8lEXFDROxbhHhKSlJIGtzUORHxWERs9THfZ9/0R+l1SQslPS5plKQO9c7bUNLtkpZJ+rekI5qo89uSVqf/B+q2vfLUZZXFyTMnST8Afgv8kizRbQpcCowoQvWbAf+KiJoi1FXxJHUqQh3/j+xndSWwNdAfOAXYG7hbUteC0y8BVpL9XI8ELpO0bRPVPxkR6xVsD3+MuqxSRIS3Vm5AT+A94NAmzulKllznp+23QNd0bC9gLnAasBBYAHwnHTuX7JdtVXqPUcDPgOsL6t4cCKBTev1t4BVgKTAbOLKg/PGC63YFJgHvpH93LTj2MPBz4P9SPQ8AfRr52uriP70g/oOAA4B/AYuBHxWcvwvwJPB2OvcPQJd07NH0tSxLX+9hBfWfAbwOXFdXlq75dHqP/0yvNwYWAXs1Eu/R6evp2sjxXwNnp/3u6fu/ZcHx64DzG7l2re9xvWOtqstbZW1lD6ASN2A4UFOXvBo55zzgKaAvsBHwBPDzdGyvdP15QOeUdN4HeqXj9ZNlo8kz/YK+C2yVjvUHtk37a36xgQ2BJcC30nWHp9e90/GHgZeBLYFu6XVjCaMu/rNT/McDbwI3AusD2wLLgS3S+TsBQ9P7bg68AHy/oL4ABjdQ/wVkf4S6FSbPdM7xqZ51gfuB3zTxs3gJ2CTtX0CWwP8PuCh9P7oBL6fjOwIf1Lv+h8BfG6n722SJfxHZH46f8uEftVbV5a2yNnfb8+kNLIqmu9VHAudFxMKIeJOsRfmtguOr0vFVEfE3slZX3jG9WmA7Sd0iYkFETG/gnK8AL0XEdRFRExHjgJnAgQXn/Dki/hURHwDjgR2aeM9VwP9ExCrgJqAP8LuIWJrefzrwHwARMSUinkrv+yrwR+CLLfiazomIFSmetUTEFWRJ8WmyPxg/bqiSNJY6PyJek7Q/sD+wPXAwMAzomOpfLKkPsB5Zy7zQO2R/FBryKLAd2R/Jb5D9UfrvdKy1dVkFcfLM5y2gTzNjcRsD/y54/e9UtqaOesn3fbJftlaJiGVkXd0TgQWS7pG0dQviqYtpQMHr11sRz1sRsTrt1yW3NwqOf1B3vaQtJd2dJmreJRt77NNE3QBvRsTyZs65gixx/T4iVjRyTl9gXtr/LHBf+oO2ELgvxdcB6EU2FPAe0KNeHT3IhjI+IiJeiYjZEVEbEdPIehOHpMOtqssqi5NnPk+SdUsPauKc+WQTP3U2TWV5LCPrntb5VOHBiLg/Ir5M1gKbSZZUmounLqZ5DZxbbJeRxTUkInoAPwLUzDVNrpUoaT2yceSrgJ9J2rCRUxeRfV8ApgH7SeorqS/Z8Et34FfA3yKilqzr3UnSkII6tidrSbdE8OHX9nHrsnbMyTOHiHiHbLzvEkkHSVpXUmdJ+6dZXYBxwE8kbZS6g2cD1+d8y6nAnpI2ldQTOKvugKR+kr4mqTuwgqy1s7qBOv4GbJlur+ok6TBgG+DunDG1xvpk47LvpVbxSfWOvwFs0co6fwdMiYjjgHuAyxs6KSL+BWwiqX9E3EvW2nwOuIusy30SWUvwh+n8ZcBtwHmSukvajewOiusaqj/9zPul/a3JxjzvzFOXVZhyD7pW8kY2rjmZrGX4Otkv8a7p2DrAxWSTEwvS/jrp2F4UTH6ksleBL6X9n1EwQZTKLiGbrZ5FNllSN2HUH3iEbCztbbKJnm3SNd9m7dn23YEp6dwpwO4Fxx4Gjit4vda19WJZK/4URwCbF5Q9DhyV9vcka3m+BzxG1rUtjOvE9D16G/hmI9+fNWVkCWgesGF6vV76vhzZSLyj08/mIxN8jZRtCNyRfq5zgCMKjm2avo5N0+vfkCX/ZWR3PJwHdG5JXd4qe1P6AZt9okn6A1mX+WyyYZcOZHc5XAAMi2wiy6zFnDytakg6GBhDlkQhu33sgoh4onxRWaVy8jQzy8ETRmZmOTh5mpnl8LEXXCiVVYte8XhCBeu28R7lDsFyqlk5r7l7cBuU93e2c58tcr1fubXb5GlmFaa2oduLP7mcPM2sOKK23BG0KSdPMyuOWidPM7NWC7c8zcxycMvTzCwHtzzNzHLwbLuZWQ5V1vL0E0ZmZjm45WlmxeEJIzOz1vOtSmZmebjlaWaWg1ueZmY5+FYlM7Mc3PI0M8vBY55mZjm45WlmloNbnmZmrRfhCSMzs9Zzt93MLAd3283McnDL08wsB98kb2aWg1ueZmY5VNmYpxdDNjPLwS1PMysOd9vNzHJwt93MLIfa2nxbC0h6VdI0SVMlTU5lG0p6UNJL6d9eBeefJWmWpBcl7VdQvlOqZ5akiyUplXeVdHMqf1rS5s3F5ORpZkURsTrX1gp7R8QOEbFzen0mMCEihgAT0mskbQOMBLYFhgOXSuqYrrkMGA0MSdvwVD4KWBIRg4GLgAuaC8bJ08yKo4Qtz0aMAK5J+9cABxWU3xQRKyJiNjAL2EVSf6BHRDwZEQFcW++aurpuBYbVtUob4+RpZsURtfm2FtYOPCBpiqTRqaxfRCwASP/2TeUDgNcKrp2bygak/frla10TETXAO0DvpgLyhJGZFUfOVmRKhqMLisZGxNh6p+0WEfMl9QUelDSzqSobKIsmypu6plFOnmZWHDlvVUqJsn6yrH/O/PTvQkm3A7sAb0jqHxELUpd8YTp9LrBJweUDgfmpfGAD5YXXzJXUCegJLG4qJnfbzaw4SjTmKam7pPXr9oF9geeBu4Bj0mnHAHem/buAkWkGfRDZxNDE1LVfKmloGs88ut41dXUdAjyUxkUb5ZanmRVH6W6S7wfcnuZvOgE3RsR9kiYB4yWNAuYAhwJExHRJ44EZQA0wJj6c1j8JuBroBtybNoCrgOskzSJrcY5sLig1k1zLZtWiV9pnYNYi3Tbeo9whWE41K+c1OcvcmA/uvTjX72y3/U/N9X7l5panmRVHlT1h5ORpZsXhZ9vNzHJwy9PMLAe3PM3Mcqiylqfv8zQzy8EtTzMrDnfbzcxyqLJuu5OnmRWHk6eZWQ7t9GnFUnHyNLPicMvTzCwHJ08zsxw8225mloNbnmZmOXjCyMwsB7c8zcxycPI0M8vBE0ZmZq0XtR7zNDNrPXfbzcxycLfdzCyHKuu2ezFkM7Mc3PI0s+LwmKeZWQ5OntYa+37jGLqvuy4dOnSgY8eOjP/TxQDccMudjPvLX+nYsSN77roLp40ZxaqaGs751W954V8vU7N6NV8bPozjjz4MgG+fcjqLFi2ma9euAIz97f/Qu9cG3HHPg1x46ZX07dMHgMO/cSCHfG14eb5Y43unHs+xxx5ORPD88zMZddwPWLFiRbnDah/8eKa11p9+fz69Nui55vXEKc/xj8ef4rZrL6VLly68teRtAB546DFWrlrF7dddxgfLlzPiyBM44Mt7MaB/PwDOP+d0tvvMlh+pf/g+X+THp53cJl+LNW7jjT/FKWOO5bPb783y5csZd+PlHPbNEVx73fhyh9Y+uOVpH9fNd9zDqKO+SZcuXQDo3WsDACTxwfLl1NSsZsWKlXTu3Jn1uq9bxkittTp16kS3buuwatUq1u3WjQULXi93SO1Hlc22lyx5StoaGAEMAAKYD9wVES+U6j3LQRKj/+vHSOLQEftz6IgDeHXOPKY89zwXj72Grl06c9opx/HZz2zFl/fenYcee5K9RxzB8uUrOP3U0fTssf6aun76y4vo0KEDX95rN0749uFIAuDBRx5n8nPT2HyTAZx+6gn077dRub7cqjZ//uv870WXM/vliXzwwXIe/PsjPPj3R8sdVvtRZfd5luRWJUlnADcBAiYCk9L+OElnluI9y+W6yy7klj//gcsu/DnjbrubyVOnsXr1at5d+h43jr2I08Ycxw9/+isigmkzXqRjhw48dOcN3Hfr1Vwz7jZem7cAgAvOOZ3br7uMay/9NVOee5677psAwF67f54Hbr2a26+9jKE778iPf3FhOb/cqrbBBj352oH7MXjLoWyy2X/Svfu6HHHE18sdVvtRG/m2ClWq+zxHAZ+LiPMj4vq0nQ/sko41SNJoSZMlTb7y2nElCq24+m7UG8i65sP23JVpM16kX98+fOmLuyGJz26zFZJY8vY7/O3Bh9lt6M507tSJ3r02YIf/2IbpM18CoN9G2YRQ9+7r8pUv783zM/4FwAY9e6zp/h/yteHMePGlMnyVBjBs2B7MfnUOixYtpqamhtvvuJcvDN253GG1G1Fbm2urVKVKnrXAxg2U90/HGhQRYyNi54jY+bijDy9RaMXz/gfLWbbs/TX7T0x8hiFbbM4+e3yBiVOmAvDqnLmsqqmh1wY96d9vIyZOeY6I4P0PlvPP6TMZtNkm1NSsZsnb7wCwqqaGR554msFbbAbAm4sWr3m/fzz+FFtstknbfpG2xmtz5vH5z/8n3bqtA8A+e+/OzJn+Y7ZGlbU8SzXm+X1ggqSXgNdS2abAYOCUEr1nm3tr8RK+96OfA7C6ZjUH7LsXuw/dmVWrVvGTX17EQUedSOfOnfjlT05DEod//UB+8sv/5aCjTiQIDjpgX7YaPIj3P1jOCT/4CatqaqhdXcvQz+245nak62+5k4cff4qOnTrSc/31+cVPTivnl1zVJk56lttuu4dJE++npqaGqVOnc8WVN5Q7rPajysY8FSW6N0tSB7Ju+gCy8c65wKSIWN2S61cteqVy/yQZ3Tbeo9whWE41K+cpz3XLzjsy1+9s97NvyPV+5Vay2faIqAWeKlX9ZtbOVPD4ZR6+z9PMiqOCxy/zcPI0s+KosjFPJ08zKw63PM3MWq+S79nMw4shm1m7J6mjpGcl3Z1ebyjpQUkvpX97FZx7lqRZkl6UtF9B+U6SpqVjFys9/yypq6SbU/nTkjZvSUxOnmZWHKW9Sf57QOG6GGcCEyJiCDAhvUbSNsBIYFtgOHCppI7pmsuA0cCQtNWt7TgKWBIRg4GLgAtaEpCTp5kVR4mSp6SBwFeAKwuKRwDXpP1rgIMKym+KiBURMRuYBewiqT/QIyKejOzm9mvrXVNX163AsLpWaVOcPM2sOKI239a83wKns/aj3f0iYgFA+rdvKh/Ah081QvZwzoC0zW2gfK1rIqIGeAfo3VxQTp5mVhw5W56FCwKlbXRdlZK+CiyMiCktjKKhFmM0Ud7UNU3ybLuZFUXkvFUpIsYCYxs5vBvwNUkHAOsAPSRdD7whqX9ELEhd8oXp/LlA4eo5A8nWEp6b9uuXF14zV1InoCewmGa45WlmxVGCMc+IOCsiBkbE5mQTQQ9FxFHAXcAx6bRjgDvT/l3AyDSDPohsYmhi6tovlTQ0jWceXe+auroOSe/hlqeZtZG2vc/zfGC8pFHAHOBQgIiYLmk8MAOoAcYULEZ0EnA10A24N20AVwHXSZpF1uIc2ZIASraq0sflVZUqm1dVqlx5V1VaevL+uX5n17/0Xq+qZGZVzI9nmpm1XnvtxZaKk6eZFYdbnmZmOTh5mpm1Xt77PCuVk6eZFYeTp5lZDtW1nKeTp5kVh7vtZmZ5VFny9LPtZmY5uOVpZsXhMU8zs9bzmKeZWR5ueZqZtZ5bnmZmebjlaWbWei37LLdPDidPMysOJ08zs9Zzy9PMLA8nTzOz1nPL08wsBydPM7McnDzNzPKIivwE4dwaTZ6SlgJ1jwzUfVci7UdE9ChxbGZWQdzyTCJi/bYMxMwqW9RWV8uzRet5Stpd0nfSfh9Jg0oblplVmqjNt1WqZpOnpHOAM4CzUlEX4PpSBmVm1t61ZMLoYGBH4BmAiJgvyV16M1tLeMLoI1ZGREgKAEndSxyTmVWgSu6C59GS5Dle0h+BDSQdDxwLXFHasMys0lTbhFGzyTMifiPpy8C7wJbA2RHxYMkjM7OKEtW1FnKLb5KfBnQju89zWunCMbNKVW0tz5bMth8HTAS+DhwCPCXp2FIHZmaVJWqVa6tULWl5/jewY0S8BSCpN/AE8KdSBmZmlcXd9o+aCywteL0UeK004ZhZparkVmQeTT3b/oO0Ow94WtKdZGOeI8i68WZma/g+zw/V3Qj/ctrq3Fm6cMysUvk+zyQizm3LQMysstW65bk2SRsBpwPbAuvUlUfEPiWMy8wqTLV121uyqtINwExgEHAu8CowqYQxmVkFKtWtSpLWkTRR0nOSpks6N5VvKOlBSS+lf3sVXHOWpFmSXpS0X0H5TpKmpWMXS1Iq7yrp5lT+tKTNm4urJcmzd0RcBayKiEci4lhgaAuuM7MqEpFva4EVwD4RsT2wAzBc0lDgTGBCRAwBJqTXSNoGGEnWWx4OXCqpY6rrMmA0MCRtw1P5KGBJRAwGLgIuaC6oliTPVenfBZK+ImlHYGALrjOzKlKqlmdk3ksvO6et7s6fa1L5NcBBaX8EcFNErIiI2cAsYBdJ/YEeEfFkRARwbb1r6uq6FRhW1yptTEvu8/yFpJ7AacDvgR7Af7XgOjOrInknjCSNJmsN1hkbEWPrndMRmAIMBi6JiKcl9YuIBQARsUBS33T6AOCpgsvnprJVab9+ed01r6W6aiS9A/QGFjUWd0sWBrk77b4D7N3c+WZmrZES5dhmzlkN7CBpA+B2Sds1cXpDWTyaKG/qmkY1dZP875u6OCJObapiM6subTHbHhFvS3qYbKzyDUn9U6uzP7AwnTYX2KTgsoHA/FQ+sIHywmvmSuoE9AQWNxVLU2Oek8mayY1tZmZrlGrCSNJGqcWJpG7Al8juALoLOCaddgwfPsBzFzAyzaAPIpsYmpi6+EslDU3jmUfXu6aurkOAh9K4aKOaukn+msaOmZnVV8Kb5PsD16Rxzw7A+Ii4W9KTZIu1jwLmAIcCRMR0SeOBGUANMCZ1+wFOAq4mW2Lz3rQBXAVcJ2kWWYtzZHNBqZnkWjarFr3SPgOzFum28R7lDsFyqlk5L1cWfHbTEbl+Z3ecc2dF3l3f0sWQzcya1E7bYSXj5GlmReFn25Nyz7a722dWWart2famWp6T2ywKM6t4bnkmnm03s9aosiHPFi9JdwawDV6SzswaUW0tz5YuSfcCXpLOzJoQoVxbpfKSdGZWFLU5t0rVkluV1lqSjuxZUC9JZ2ZriQbX1vjk8pJ0ZlYUtVU2Y+Ql6cysKGrd8lybpD/TwF0IaezTzAxwt70hdxfsrwMczIdr4JmZVaWWdNv/Uvha0jjg7yWLyMwqUiXPnOeRZ2GQIcCmxQ7EzCqbu+31SFrK2mOer5M9cWRmtoZbnvVExPptEYiZVbZqS57NPmEkaUJLysysugXKtVWqptbzXAdYF+gjqRcffjRnD2DjNojNzCpIbeXmwVya6rafAHyfLFFO4cPk+S5wSWnDMrNK45vkk4j4HfA7Sd+NiN+3YUxmVoGq7OnMFq2qVFv3mckAknpJOrl0IZlZJaq2VZVakjyPj4i3615ExBLg+JJFZGYVqVbKtVWqltwk30GSIn3Ae/rg+S6lDcvMKk21ddtbkjzvB8ZLupzs+3MicF9JozKzilPJXfA8WpI8zwBGAyeRzbg/AFxRyqDMrPJU261KzY55RkRtRFweEYdExDeA6WSLIpuZrVGLcm2VqkULg0jaATgcOAyYDdxWwpjMrAJ5zDORtCUwkixpvgXcDCgivJq8mX1EtXXbm2p5zgQeAw6MiFkAkvzZRWZmND3m+Q2y5ef+IekKScOgggcozKykfJN8EhG3R8RhwNbAw2SfmNlP0mWS9m2j+MysQkTOrVK1ZLZ9WUTcEBFfJfu89qnAmaUOzMwqS63ybZWqJY9nrhERiyPijxGxT6kCMrPKVG3d9jyfYWRm9hGVnAjzcPI0s6KICu6C5+HkaWZF4ZanmVkO1ZY8WzVhZGbWmFLdqiRpE0n/kPSCpOmSvpfKN5T0oKSX0r+9Cq45S9IsSS9K2q+gfCdJ09Kxi6VsQVFJXSXdnMqflrR5c3E5eZpZUZTwVqUa4LSI+AwwFBgjaRuyWyYnRMQQYEJ6TTo2EtgWGA5cmtYhBriMbJW4IWkbnspHAUsiYjBwEXBBc0E5eZpZUZTqVqWIWBARz6T9pcALwABgBHBNOu0a4KC0PwK4KSJWRMRsYBawi6T+QI+IeDIt7n5tvWvq6roVGFbXKm2Mk6eZFUVb3OeZutM7Ak8D/SJiAWQJFuibThsAvFZw2dxUNiDt1y9f65qIqAHeAXo3FYuTp5kVRd4xT0mjJU0u2EY3VL+k9YC/AN+PiHebCKWhFmM0Ud7UNY3ybLuZFUXeRy0jYiwwtqlzJHUmS5w3RETdesJvSOofEQtSl3xhKp8LbFJw+UBgfiof2EB54TVzJXUCegKLm4rJLU8zK4pSddvT2ONVwAsR8b8Fh+4Cjkn7xwB3FpSPTDPog8gmhiamrv1SSUNTnUfXu6aurkOAh+o+9LIxbnmaWVGUcIWk3YBvAdMkTU1lPwLOJ/twylHAHOBQgIiYLmk8MINspn5MRKxO150EXA10A+5NG2TJ+TpJs8hanCObC0rNJNey6dRlQPsMzOwTrmblvFwd8P/Z7Mhcv7M//vcNFflgp7vtZmY5uNtuZkVRbY9nOnmaWVFU2zibk6eZFYVbnmZmOVTyR2rk4eRpZkVRW2UddydPMyuK6kqdTp5mViQe8zQzy8HddjOzHKordTp5mlmRuNtuZpaDu+1mZjlUV+p08jSzInG33cwsh6iytqeTp5kVhVueZmY5VNuEkRdDNjPLwcmzxAYO3Ji/P3AL0/75MM9NfYjvnjIKgHN/9t88M+VBJk96gHvvuZH+/fuVOVJryJZbfprJkx5Ysy1eNJNTv3scAGNO/g7Tn3+U56Y+xPm/+nGZIy2/vB89XKn8GUYl9qlP9aX/p/ry7NTnWW+97kx8+j6+ccixzJ27gKVL3wPglDHH8pnPbMmYU84sc7TWlA4dOjDn1SnsuvtX2WLQZpx15qkcOOJoVq5cyUYb9ebNN98qd4hFkfczjE7Y/NBcv7N/fPWWilzMzi3PEnv99YU8O/V5AN57bxkzZ77EgI0/tSZxAnTvvi7t9Y+YfWjYPrvzyiv/Zs6ceZxwwtH8v19fwsqVKwE+MYnz4yjVRw+3V06ebWizzQayw/bb8fTEZwH4+XlnMPvlSRx++MH87Nxflzk6a843vzmCm26+A4AhQ7Zg99134YnH/8pDf7+VnXfavrzBtQOR879K1ebJU9J32vo924Pu3ddl/M1X8IMfnrOm1fnTsy9g0Kc/x7hxtzPm5Kr8tlSMzp07c+BX9+XWv9wNQKdOHdlgg57suvuBnHHmLxh34+VljrD83PIsvXMbOyBptKTJkibX1i5ry5hKqlOnTtxy8xWMG3c7d9xx70eOj7vpdg4++IAyRGYtNXz43jz77DQWLlwEwLy5C9b8LCdNnkptbS19+mxYzhDLrtpaniW5z1PSPxs7BDQ6rRwRY4Gx8MmZMAK4YuyFvDBzFr/93dg1ZYMHD2LWrNkAHPjVfXnxxZfLFZ61wMjDDlrTZQe486772Xvv3Xjk0ScZMmQLunTpwqJFi8sXYDtQya3IPEp1k3w/YD9gSb1yAU+U6D3bpd12/RzfOuoQ/jltBpMnPQDAT396Pt/5zki23PLT1NbWMmfOPE4e45n29qpbt3X40rA9OenkM9aU/fnqm7jyiguZ+uwEVq5cxbGjvl++ANuJ2iqb9CzJrUqSrgL+HBGPN3Dsxog4ork6PkktT7NKkvdWpaM2+3qu39nr/31bRd6qVJKWZ0SMauJYs4nTzCpPtT2e6WfbzawoKnnyJw8nTzMrCk8YmZnl4G67mVkO7rabmeXgbruZWQ7VtriNFwYxM8vBLU8zKwpPGJmZ5eAxTzOzHDzbbmaWQ7V12z1hZGZFERG5tuZI+pOkhZKeLyjbUNKDkl5K//YqOHaWpFmSXpS0X0H5TpKmpWMXS1Iq7yrp5lT+tKTNW/L1OnmaWVGUcCX5q4Hh9crOBCZExBBgQnqNpG2AkcC26ZpLJXVM11wGjAaGpK2uzlHAkogYDFwEXNCSoJw8zawoSrWSfEQ8CtRfaXoEcE3avwY4qKD8pohYERGzgVnALpL6Az0i4snImrvX1rumrq5bgWF1rdKmeMzTzIqijcc8+0XEAoCIWCCpbyofADxVcN7cVLYq7dcvr7vmtVRXjaR3gN7AoqYCcMvTzIoi75hn4WeXpW30xwijoRZjNFHe1DVNcsvTzIoib8uz8LPLWuENSf1Tq7M/sDCVzwU2KThvIDA/lQ9soLzwmrmSOgE9+egwwUe45WlmRdHGn555F3BM2j8GuLOgfGSaQR9ENjE0MXXxl0oamsYzj653TV1dhwAPRQtuA3DL08yKolQfACdpHLAX0EfSXOAc4HxgvKRRwBzgUICImC5pPDADqAHGRMTqVNVJZDP33YB70wZwFXCdpFlkLc6RLYqrva6E4g+AMyuPvB8At8eAYbl+Zx+bN8EfAGdm1avanjBy8jSzonDyNDPLob0OAZaKZ9vNzHJwy9PMisLddjOzHLyep5lZDtU25unkaWZF4W67mVkObnmameXglqeZWQ6eMDIzy6FUC4O0V06eZlYUbnmameXglqeZWQ5ueZqZ5eCWp5lZDm55mpnl4JanmVkObnmameUQUVvuENqUF0M2M8vBLU8zKwo/225mloNXVTIzy8EtTzOzHNzyNDPLwfd5mpnl4Ps8zcxycLfdzCwHTxiZmeXglqeZWQ6eMDIzy8EtTzOzHDzmaWaWg1ueZmY5eMzTzCwH3yRvZpaDW55mZjlU25inV5I3M8vBLU8zKwqPeZqZ5VBt3XYnTzMrCidPM7Mcqit1gqrtr0V7IWl0RIwtdxyWj39+5tn28hld7gDsY/HPr8o5eZqZ5eDkaWaWg5Nn+Xi8rLL551flPGFkZpaDW55mZjk4ebYxScMlvShplqQzyx2PtY6kP0laKOn5csdi5eXk2YYkdQQuAfYHtgEOl7RNeaOyVroaGF7uIKz8nDzb1i7ArIh4JSJWAjcBI8ock7VCRDwKLC53HFZ+Tp5tawDwWsHruanMzCqMk2fbUgNlvt3BrAI5ebatucAmBa8HAvPLFIuZfQxOnm1rEjBE0iBJXYCRwF1ljsnMcnDybEMRUQOcAtwPvACMj4jp5Y3KWkPSOOBJYCtJcyWNKndMVh5+wsjMLAe3PM3McnDyNDPLwcnTzCwHJ08zsxycPM3McnDy/ISQtFrSVEnPS7pF0rofo66rJR2S9q9savESSXtJ2jXHe7wqqU9Ly+ud814r3+tnkn7Y2hjNmuLk+cnxQUTsEBHbASuBEwsPphWdWi0ijouIGU2cshfQ6uRpVumcPD+ZHgMGp1bhPyTdCEyT1FHSryVNkvRPSScAKPMHSTMk3QP0ratI0sOSdk77wyU9I+k5SRMkbU6WpP8rtXr3kLSRpL+k95gkabd0bW9JD0h6VtIfafg5/7VIukPSFEnTJY2ud+zCFMsESRulsk9Lui9d85ikrYvy3TRrQKdyB2DFJakT2Xqh96WiXYDtImJ2SkDvRMTnJHUF/k/SA8COwFbAZ4F+wAzgT/Xq3Qi4Atgz1bVhRCyWdDnwXkT8Jp13I3BRRDwuaVOyp6k+A5wDPB4R50n6Ci376N5j03t0AyZJ+ktEvAV0B56JiNMknZ3qPoXsc4VOjIiXJH0euBTYJ8e30axZTp6fHN0kTU37jwFXkXWnJ0bE7FS+L/AfdeOZQE9gCLAnMC4iVgPzJT3UQP1DgUfr6oqIxta0/BKwjbSmYdlD0vrpPb6err1H0pIWfE2nSjo47W+SYn0LqAVuTuXXA7dJWi99vbcUvHfXFryHWS5Onp8cH0TEDoUFKYksKywCvhsR99c77wCaXxpPLTgHsqGgL0TEBw3E0uJngSXtRZaIvxAR70t6GFinkdMjve/b9b8HZqXiMc/qcj9wkqTOAJK2lNQdeBQYmcZE+wN7N3Dtk8AXJQ1K126YypcC6xec9wBZF5p03g5p91HgyFS2P9CrmVh7AktS4tyarOVbpwNQ13o+gmw44F1gtqRD03tI0vbNvIdZbk6e1eVKsvHMZ9IHmP2RrPdxO/ASMA24DHik/oUR8SbZOOVtkp7jw27zX4GD6yaMgFOBndOE1Aw+nPU/F9hT0jNkwwdzmon1PqCTpH8CPweeKji2DNhW0hSyMc3zUvmRwKgU33T8ESdWQl5VycwsB7c8zcxycPI0M8vBydPMLAcnTzOzHJw8zcxycPI0M8vBydPMLAcnTzOzHP4/SLJGmN3aFxwAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 360x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate your model on test dataset.\n",
    "baseline_results = model.evaluate(test_features, test_labels,\n",
    "                                  batch_size=BATCH_SIZE, verbose=0)\n",
    "for name, value in zip(model.metrics_names, baseline_results):\n",
    "  print(name, ': ', value)\n",
    "print()\n",
    "\n",
    "plot_cm(test_labels, test_predictions_baseline)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reciever Operatering Curve "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_roc(name, labels, predictions, **kwargs):\n",
    "  fp, tp, _ = sklearn.metrics.roc_curve(labels, predictions)\n",
    "\n",
    "  plt.plot(100*fp, 100*tp, label=name, linewidth=2, **kwargs)\n",
    "  plt.xlabel('False positives [%]')\n",
    "  plt.ylabel('True positives [%]')\n",
    "  plt.xlim([-0.5,20])\n",
    "  plt.ylim([80,100.5])\n",
    "  plt.grid(True)\n",
    "  ax = plt.gca()\n",
    "  ax.set_aspect('equal')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x14a7bd940>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\n",
    "plot_roc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\n",
    "plt.legend(loc='lower right')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot the Area Under Precision Recall Curve (AUPRC)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_prc(name, labels, predictions, **kwargs):\n",
    "    precision, recall, _ = sklearn.metrics.precision_recall_curve(labels, predictions)\n",
    "\n",
    "    plt.plot(precision, recall, label=name, linewidth=2, **kwargs)\n",
    "    plt.xlabel('Recall')\n",
    "    plt.ylabel('Precision')\n",
    "    plt.grid(True)\n",
    "    ax = plt.gca()\n",
    "    ax.set_aspect('equal')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x176a7d610>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_prc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\n",
    "plot_prc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\n",
    "plt.legend(loc='lower right')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using class weights for imabalanced dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The goal is to identify fraudulent transactions, but you don't have very many of those positive samples to work with, so you would want to have the classifier heavily weight the few examples that are available. You can do this by passing Keras weights for each class through a parameter. These will cause the model to \"pay more attention\" to examples from an under-represented class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Examples:\n",
      "    Total: 284807\n",
      "    Positive: 492 (0.17% of total)\n",
      "\n",
      "Weight for class 0: 0.50\n",
      "Weight for class 1: 289.44\n"
     ]
    }
   ],
   "source": [
    "# Scaling by total/2 helps keep the loss to a similar magnitude.\n",
    "# The sum of the weights of all examples stays the same.\n",
    "neg, pos = np.bincount(raw_df['Class'])\n",
    "total = neg + pos\n",
    "print('Examples:\\n    Total: {}\\n    Positive: {} ({:.2f}% of total)\\n'.format(\n",
    "    total, pos, 100 * pos / total))\n",
    "    \n",
    "weight_for_0 = (1 / neg) * (total / 2.0)\n",
    "weight_for_1 = (1 / pos) * (total / 2.0)\n",
    "\n",
    "class_weight = {0: weight_for_0, 1: weight_for_1}\n",
    "\n",
    "print('Weight for class 0: {:.2f}'.format(weight_for_0))\n",
    "print('Weight for class 1: {:.2f}'.format(weight_for_1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Now we use these weights and re-train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "90/90 [==============================] - 1s 4ms/step - loss: 2.7256 - true positive: 121.0000 - false positive: 167.0000 - true negative: 238661.0000 - false negative: 289.0000 - accuracy: 0.9981 - precision: 0.4201 - recall: 0.2951 - auc: 0.7303 - prc: 0.2434 - val_loss: 0.0088 - val_true positive: 19.0000 - val_false positive: 4.0000 - val_true negative: 45483.0000 - val_false negative: 63.0000 - val_accuracy: 0.9985 - val_precision: 0.8261 - val_recall: 0.2317 - val_auc: 0.8830 - val_prc: 0.4929\n",
      "Epoch 2/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 1.4418 - true positive: 133.0000 - false positive: 467.0000 - true negative: 181498.0000 - false negative: 178.0000 - accuracy: 0.9965 - precision: 0.2217 - recall: 0.4277 - auc: 0.7733 - prc: 0.2486 - val_loss: 0.0092 - val_true positive: 50.0000 - val_false positive: 7.0000 - val_true negative: 45480.0000 - val_false negative: 32.0000 - val_accuracy: 0.9991 - val_precision: 0.8772 - val_recall: 0.6098 - val_auc: 0.9389 - val_prc: 0.6460\n",
      "Epoch 3/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.9799 - true positive: 171.0000 - false positive: 948.0000 - true negative: 181017.0000 - false negative: 140.0000 - accuracy: 0.9940 - precision: 0.1528 - recall: 0.5498 - auc: 0.8620 - prc: 0.3486 - val_loss: 0.0122 - val_true positive: 63.0000 - val_false positive: 19.0000 - val_true negative: 45468.0000 - val_false negative: 19.0000 - val_accuracy: 0.9992 - val_precision: 0.7683 - val_recall: 0.7683 - val_auc: 0.9512 - val_prc: 0.7248\n",
      "Epoch 4/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.6913 - true positive: 213.0000 - false positive: 1596.0000 - true negative: 180369.0000 - false negative: 98.0000 - accuracy: 0.9907 - precision: 0.1177 - recall: 0.6849 - auc: 0.8846 - prc: 0.3961 - val_loss: 0.0166 - val_true positive: 66.0000 - val_false positive: 42.0000 - val_true negative: 45445.0000 - val_false negative: 16.0000 - val_accuracy: 0.9987 - val_precision: 0.6111 - val_recall: 0.8049 - val_auc: 0.9571 - val_prc: 0.7487\n",
      "Epoch 5/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.6498 - true positive: 218.0000 - false positive: 2577.0000 - true negative: 179388.0000 - false negative: 93.0000 - accuracy: 0.9854 - precision: 0.0780 - recall: 0.7010 - auc: 0.8868 - prc: 0.3237 - val_loss: 0.0230 - val_true positive: 67.0000 - val_false positive: 117.0000 - val_true negative: 45370.0000 - val_false negative: 15.0000 - val_accuracy: 0.9971 - val_precision: 0.3641 - val_recall: 0.8171 - val_auc: 0.9630 - val_prc: 0.7428\n",
      "Epoch 6/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.5167 - true positive: 239.0000 - false positive: 3392.0000 - true negative: 178573.0000 - false negative: 72.0000 - accuracy: 0.9810 - precision: 0.0658 - recall: 0.7685 - auc: 0.9075 - prc: 0.3229 - val_loss: 0.0289 - val_true positive: 68.0000 - val_false positive: 199.0000 - val_true negative: 45288.0000 - val_false negative: 14.0000 - val_accuracy: 0.9953 - val_precision: 0.2547 - val_recall: 0.8293 - val_auc: 0.9668 - val_prc: 0.7134\n",
      "Epoch 7/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.4706 - true positive: 241.0000 - false positive: 4093.0000 - true negative: 177872.0000 - false negative: 70.0000 - accuracy: 0.9772 - precision: 0.0556 - recall: 0.7749 - auc: 0.9161 - prc: 0.3109 - val_loss: 0.0356 - val_true positive: 72.0000 - val_false positive: 303.0000 - val_true negative: 45184.0000 - val_false negative: 10.0000 - val_accuracy: 0.9931 - val_precision: 0.1920 - val_recall: 0.8780 - val_auc: 0.9673 - val_prc: 0.6864\n",
      "Epoch 8/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.4091 - true positive: 252.0000 - false positive: 4972.0000 - true negative: 176993.0000 - false negative: 59.0000 - accuracy: 0.9724 - precision: 0.0482 - recall: 0.8103 - auc: 0.9251 - prc: 0.2675 - val_loss: 0.0421 - val_true positive: 72.0000 - val_false positive: 387.0000 - val_true negative: 45100.0000 - val_false negative: 10.0000 - val_accuracy: 0.9913 - val_precision: 0.1569 - val_recall: 0.8780 - val_auc: 0.9674 - val_prc: 0.6955\n",
      "Epoch 9/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.3433 - true positive: 263.0000 - false positive: 5409.0000 - true negative: 176556.0000 - false negative: 48.0000 - accuracy: 0.9701 - precision: 0.0464 - recall: 0.8457 - auc: 0.9415 - prc: 0.2660 - val_loss: 0.0470 - val_true positive: 72.0000 - val_false positive: 446.0000 - val_true negative: 45041.0000 - val_false negative: 10.0000 - val_accuracy: 0.9900 - val_precision: 0.1390 - val_recall: 0.8780 - val_auc: 0.9680 - val_prc: 0.6971\n",
      "Epoch 10/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.3899 - true positive: 261.0000 - false positive: 5740.0000 - true negative: 176225.0000 - false negative: 50.0000 - accuracy: 0.9682 - precision: 0.0435 - recall: 0.8392 - auc: 0.9279 - prc: 0.2692 - val_loss: 0.0502 - val_true positive: 72.0000 - val_false positive: 476.0000 - val_true negative: 45011.0000 - val_false negative: 10.0000 - val_accuracy: 0.9893 - val_precision: 0.1314 - val_recall: 0.8780 - val_auc: 0.9681 - val_prc: 0.6993\n",
      "Epoch 11/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.2993 - true positive: 272.0000 - false positive: 5800.0000 - true negative: 176165.0000 - false negative: 39.0000 - accuracy: 0.9680 - precision: 0.0448 - recall: 0.8746 - auc: 0.9481 - prc: 0.2743 - val_loss: 0.0513 - val_true positive: 72.0000 - val_false positive: 484.0000 - val_true negative: 45003.0000 - val_false negative: 10.0000 - val_accuracy: 0.9892 - val_precision: 0.1295 - val_recall: 0.8780 - val_auc: 0.9679 - val_prc: 0.7016\n",
      "Epoch 12/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.3726 - true positive: 260.0000 - false positive: 5906.0000 - true negative: 176059.0000 - false negative: 51.0000 - accuracy: 0.9673 - precision: 0.0422 - recall: 0.8360 - auc: 0.9302 - prc: 0.2504 - val_loss: 0.0546 - val_true positive: 72.0000 - val_false positive: 514.0000 - val_true negative: 44973.0000 - val_false negative: 10.0000 - val_accuracy: 0.9885 - val_precision: 0.1229 - val_recall: 0.8780 - val_auc: 0.9678 - val_prc: 0.7037\n",
      "Epoch 13/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.3607 - true positive: 261.0000 - false positive: 6167.0000 - true negative: 175798.0000 - false negative: 50.0000 - accuracy: 0.9659 - precision: 0.0406 - recall: 0.8392 - auc: 0.9303 - prc: 0.2405 - val_loss: 0.0587 - val_true positive: 72.0000 - val_false positive: 564.0000 - val_true negative: 44923.0000 - val_false negative: 10.0000 - val_accuracy: 0.9874 - val_precision: 0.1132 - val_recall: 0.8780 - val_auc: 0.9678 - val_prc: 0.6732\n",
      "Epoch 14/100\n",
      "90/90 [==============================] - 0s 1ms/step - loss: 0.3630 - true positive: 263.0000 - false positive: 6462.0000 - true negative: 175503.0000 - false negative: 48.0000 - accuracy: 0.9643 - precision: 0.0391 - recall: 0.8457 - auc: 0.9313 - prc: 0.2402 - val_loss: 0.0599 - val_true positive: 72.0000 - val_false positive: 580.0000 - val_true negative: 44907.0000 - val_false negative: 10.0000 - val_accuracy: 0.9871 - val_precision: 0.1104 - val_recall: 0.8780 - val_auc: 0.9705 - val_prc: 0.6745\n",
      "Restoring model weights from the end of the best epoch.\n",
      "Epoch 00014: early stopping\n"
     ]
    }
   ],
   "source": [
    "# Train the model with class weights.\n",
    "weighted_model = make_model()\n",
    "weighted_model.load_weights(initial_weights)\n",
    "\n",
    "weighted_history = weighted_model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    epochs=EPOCHS,\n",
    "    callbacks=[early_stopping],\n",
    "    validation_data=(val_features, val_labels),\n",
    "    # The class weights go here\n",
    "    class_weight=class_weight) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_metrics(weighted_history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Diagnostic Metrics -- weighted class "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_predictions_weighted = weighted_model.predict(train_features, batch_size=BATCH_SIZE)\n",
    "test_predictions_weighted = weighted_model.predict(test_features, batch_size=BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss :  0.017389826476573944\n",
      "true positive :  78.0\n",
      "false positive :  65.0\n",
      "true negative :  56798.0\n",
      "false negative :  21.0\n",
      "accuracy :  0.9984902143478394\n",
      "precision :  0.5454545617103577\n",
      "recall :  0.7878788113594055\n",
      "auc :  0.9571707844734192\n",
      "prc :  0.7003578543663025\n",
      "\n",
      "Legitimate Transactions Detected (True Negatives):  56798\n",
      "Legitimate Transactions Incorrectly Detected (False Positives):  65\n",
      "Fraudulent Transactions Missed (False Negatives):  21\n",
      "Fraudulent Transactions Detected (True Positives):  78\n",
      "Total Fraudulent Transactions:  99\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "weighted_results = weighted_model.evaluate(test_features, test_labels,\n",
    "                                           batch_size=BATCH_SIZE, verbose=0)\n",
    "for name, value in zip(weighted_model.metrics_names, weighted_results):\n",
    "  print(name, ': ', value)\n",
    "print()\n",
    "\n",
    "plot_cm(test_labels, test_predictions_weighted)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# compared to metrics *without* class weights\n",
    "\n",
    "loss :  0.0025574234314262867\n",
    "true positive :  69.0\n",
    "false positive :  6.0\n",
    "true negative :  56862.0\n",
    "false negative :  25.0\n",
    "accuracy :  0.9994557499885559\n",
    "precision :  0.9200000166893005\n",
    "recall :  0.7340425252914429\n",
    "auc :  0.9412895441055298\n",
    "prc :  0.8399971127510071\n",
    "\n",
    "Legitimate Transactions Detected (True Negatives):  56862\n",
    "Legitimate Transactions Incorrectly Detected (False Positives):  6\n",
    "Fraudulent Transactions Missed (False Negatives):  25\n",
    "Fraudulent Transactions Detected (True Positives):  69\n",
    "Total Fraudulent Transactions:  94"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here you can see that with class weights the accuracy and precision are lower because there are more false positives, but conversely the recall and AUC are higher because the model also found more true positives. Despite having lower accuracy, this model has higher recall (and identifies more fraudulent transactions). Of course, there is a cost to both types of error (you wouldn't want to bug users by flagging too many legitimate transactions as fraudulent, either). Carefully consider the trade-offs between these different types of errors for your application."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot the Reciever Operator Curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x17683f4c0>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmoAAAJNCAYAAACBe1nxAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAABubklEQVR4nO3deXxU1f3/8dfJHpKwhx0FZJE9QEAWwVAVUXGpluLWL2pbsLWu37Zq/Vlpba2ttlXUfpUqLi1112rVWsQaURGUVdkE2ZewQ0ggIdv5/XEnk5DMTCbLzJ3MvJ+PRx5z595zz/3MYQgfzj33HGOtRUREREQiT5zbAYiIiIiIb0rURERERCKUEjURERGRCKVETURERCRCKVETERERiVBK1EREREQiVILbAYRD+/btbY8ePep1zrFjx0hLSwtNQM2M2sKhdnCoHRxqhypqC4fawaF2qLJs2bID1trMxtQRE4lajx49WLp0ab3Oyc3NJScnJzQBNTNqC4fawaF2cKgdqqgtHGoHh9qhijFmW2Pr0K1PERERkQilRE1EREQkQilRExEREYlQStREREREIpQSNREREZEIpURNREREJEIpURMRERGJUCFL1Iwxc40x+4wxq6vta2uMed8Ys9Hz2qbasbuMMd8YY742xpznp06/54uIiIhEm1D2qD0LTK6x707gA2ttH+ADz3uMMQOAK4CBnnP+YoyJ91Gnz/NFREREolHIEjVr7ULgUI3dlwDPebafAy6ttv9Fa+0Ja+0W4BtglI9q/Z0vIiIiEnXCPUato7U2D8Dz2sGzvyuwo1q5nZ59wZ4vIiIiEnUiZa1P42OfbVSFxswAZgB07NiR3Nzcep1fWFhY73OildrCoXZwqB0caocqaguH2sGhdmha4U7U9hpjOltr84wxnYF9nv07ge7VynUDdtfj/FqstXOAOQDZ2dm2vgvEalHZKmoLh9rBoXZwqB2qqC0cageH2qFphfvW51vAdM/2dODNavuvMMYkG2N6An2Az+txvoiIiEjUCeX0HC8AnwH9jDE7jTHfBx4AzjXGbATO9bzHWrsGeBlYC7wH3GitLffU85QxJttTrc/zRURERKJRyG59Wmuv9HPobD/lfwv81sf+H1TbPujvfBEREZFoo5UJRERERCKUEjURERGRCKVETURERCRCKVETERERiVBK1EREREQilBI1ERERkQilRE1EREQkQilRExEREYlQStREREREIpQSNREREZEIpURNREREJEIpURMRERGJUErURERERCKUEjURERGRCKVETURERCRCKVETERERiVBK1EREREQilBI1ERERkQilRE1EREQkQilRExEREYlQStREREREIlSC2wGIiIgEa/P+Qu587Su/x++/bDC9O6QDMGfhJhas3eezXI/2LfjDd4Z633/3ic/81vmD8T2ZNLATAO+v3ctfF272We5IfhFnnWUxxgBw52tfsnn/MZ9lz+7fgZlnnRbxnwngpZmj9Zka8JkOFpb4LVsfStRERKTZOF5SzudbDwU4Xubd3nLguN+yx6qVAwLWecmwLt7t/QUnApatbs3uo3y1K9/nsdM6pFWLWZ8pGj/TnvyioK5fFyVqIiIS8d75Mg+L5ay+mbw0Y7Tfcr0y073bPxzfk0uzuvgsl5Z88j9/gers2b7qH+tz+nfgtEzfZVeuXHnS+99dNphjJ8p8ls3MSPZu92ifFrGfqSZ9puA/U2m55dmgIghMiZqIiES8W15cQVmFZeNvz+eMXu2COqdXZvpJ/3gGEmydHVqm0KFlis9jRdvjvbfTAAZ1bRVUnenJCRH7mWrSZwr+MzUVPUwgIiIiEqGUqImISMSzbgcg4hIlaiIi0myYuouIRBWNURMRkYhnbWz2qRUUl/KvVXkUlZa7HUrQvtlayqZPtrgdRtRQoiYiIs1G9cH6seDZT7fyx/c3uB1G/a1f63YEUUOJmoiISIQ6UlQKwMgebYJ+OtFtO3fupFu3bm6HERFmNUEdStRERCTirfjlJADiYqtDjco7vucN7MQPxvdyN5gg5ebuJydnoNthRIRZTVCHEjUREYl4rVIT3Q7BFRWeTC3WbvlKFT31KSIiEuGUpsUu9aiJiEjEu/qpxQD8/ftnxFTvkvX2qLkciLhGiZqIiES8T785CMTeLcAKzxi1uBj73FJFtz5FRCSixeocagAW9ajFOvWoiYiIX6XlFXzv6SV8s6/Qb5mSkhKSPnk/ZDFUT9SyfxO66zRWKNqhoLgM0Bi1WKZETURE/Np28BiLNx+qu2BJSeiDAQ4Uhuc6DRaCdkiKj6N/55ZNXq80D0rURETEr8oxUr3ap/HSzDE+yyxatIixY8eGMAbLGfd/AMAXd58Tsus0VqjaoUVSPGnJ+uc6VulPXkRE/KqcxysxPo7MjGSfZVolG7/HmiSGiqpxWqG8TmOFuh0kNilRExERvyoqnFe3B7NfOLizuwGIuESJmoiI+FX11KF7mVpcnOHxq4e7dn0RN2l6DhER8ct65/FyNw6RWKUeNRER8atyjJqbE65aa9l+6DgGwyntWrgWh4gblKiJSEQ6dqKMBev2cqK0wu1QfFq/s5R9X+xwO4yQ23H4OODuGLWyCstZD+aSEGf45v4L3AtExAVK1EQkIj25cDOzP9jodhiBrf7S7QjCJiUh3u0QRGKSEjURiUiHj1VNHDp1RDcXI/Ftz549dOrUye0wwiLOGL6THXl/BiKxQImaiES0X18ykP8Z08PtMGrJzT1MTs5QV67976/y+PmrX3L16FO58/zTAVi06QDXzv3C7znzb5tAj/ZpANzy4gr+/dUen+XO6NWWv33/DACOFpeSfd8CAN5YsatW2T9Py+LCIc60GU99vJk/vPe1zzrTUxJYfs+53vfn/Okjth887rPs9Wf2rPWZKp88FYlFStRERJqZhRv3U3CijPlr9niTGiyUlAc3nq+swvotW1Z+clIUqM6Kamtwlgeos7Ts5P2l5RV+y1avs+ZnGte7vd9YRKKVEjURkWamchLa687s6d03ulc7vv7NZL/nJMVXzcb08LQs/vRd372Bptry3xnJCQHrTIirqvP7Z/bk2nE96godcHr3/Imv9tRCzc9U/TOIxAolaiIizUzlrcDEapObxcUZkuOCG/CfGGTCY4whOciHCBLi44L+ByXYOuvzmUSilf57IiLSzFRNQqtZaEWinRI1EZFmpqJyGJfyNJGop1ufIjFgwdq9/P699ZQGOdg8kKKiIlK/+LAJogrsYGFJ3YVilHf9TZfjEJHQU6ImEgP+uXIXG/cVNl2Fx31PrdDUjIHemelhuVZz8sPxvbhoSBdO75zhdigiEmJK1ERiQOWUC7MuGsBZ/To0qq4lS5ZwxhlnNEVYdcpISaB9enJYrtWc9O/ckv6dW7odhoiEgRI1kRhQ7hl93rl1Kj09k5421La0uEbXISIiwXElUTPG3AL8EGeIxV+ttQ8bY14C+nmKtAaOWGuzfJy7FSgAyoEya212OGIWac7KPaPP4/WUYFR4+8vdfL2ngAsGd1bPmkiUC3uiZowZhJOkjQJKgPeMMe9Ya6dVK/NHID9ANROttQdCG6lI9PAmavFK1KLBe6v38PaXefTpmKFETSTKuTE9R39gsbX2uLW2DPgI+HblQWOMAb4LvOBCbCJRST1q0aVyHjX9aYpEPzcStdXABGNMO2NMC+ACoHu14+OBvdbajX7Ot8B8Y8wyY8yMEMcqEhXKPGsOJcTpn/Zo4J2eQ3+cIlHPWGvrLtXUFzXm+8CNQCGwFiiy1t7mOfZ/wDfW2j/6ObeLtXa3MaYD8D5wk7V2oY9yM4AZAB07dhzx4osv1ivGwsJC0tM1LQCoLSpFSjvsKqxgV2H95kP758YSdh+z3DkqhdPbNm5JnkhpB7e52Q6PrShm6d5ybsxKZmQn958J03fCoXZwqB2qTJw4cVljx9K78jfcWvs08DSAMeZ+YKdnOwG4DBgR4Nzdntd9xpg3cMa61UrUrLVzgDkA2dnZNicnp14x5ubmUt9zopXawhEJ7XC8pIzh971PcWnDJq49I3s4w05p06gYIqEdIoGb7fDCjqWwdy+DBg4kZ3BnV2KoTt8Jh9rBoXZoWm499dnBk2idgpOYjfEcOgdYb63d6ee8NCDOWlvg2Z4E/DosQYtEgKNFZRSXVpCUEMc5/es3H1rX1qkM6dY6NIFJWHnHqOnWp0jUc6vP/DVjTDugFLjRWnvYs/8KajxEYIzpAjxlrb0A6Ai84TxvQALwD2vte+ELW8RdlfOhtUtL4i9X++14lijXpkUSnVqmkJLYuNvYIhL53Lr1Od7P/mt97NuN88AB1trNwNCQBicSwSo8T2/GqSslpv3+O0PcDkFEwsSNpz5FpIG802zo6U0RkZigRE2kGam89alETUQkNihRE2lGqm59uhyIuOrmF1Yw4r73+XjjfrdDEZEQU6Im0oyoR00A8otKOXishLLy8M+DKSLh5f5MiSJykqKScm56YQV5+UW1jhWXlgN6mCDWedMzfQ1Eop4SNZEIs2LHYRas2xuwTK/MtDBFI5GockUZJewi0U+Jmkik8XSXZKQk8MIPR9c6bAz07ZgR5qAkkmhRdpHYoURNJEIN6tKKQV1buR1GzDl2ooytB4/V2t+mRRJdWqcCzlJe246Ws2Z3vs86TstM905Gu+tIEUeOl/gsl5oYT69MZ01Eay1r8476jatLq1TapCUBcPCYU5861ESinxI1EREPay0XzP6YbQeP1zp2zehT+M2lgwH4ek8B9y4qhkWf+KznnZvPZGAXJ8l+ZMEGXl7qc1U8hnZvzZs3jgOgwsKFs33XB/D7ywczbeQpALRpkQhAvDI1kainRE1ExMNavEla/84tTzrWuVWqdzs1KZ7uGXGkp6f7rKf60k6dW6XWqqtSz3YtTnrvrxxAq9Qk7/aVo04hJTGerFNa+y0vItFBiZqISDVTR3TDGPjDd/yvVnd6p5bcNy6VnByfq+Gd5LZz+3LbuX3rLBcfZ/j3LXXXB3DR0C5cNLRLUGVFpHlToiYi4hEXZ3hwqpYTFpHIoURNJAKs33OUvUdPALB2t/8B5SIiEluUqIm4bPWufKY8WnsQeUK8BoqHm7WW9XsKgMDjxUREwkWJmojL9h4t9m6P79MecMYrXT+up1shxSxr4fxHPsYY2PK7C90OR0REiZpIpPjW6R2Ye+1It8MQEZEIokXZRURERCKUEjURERGRCKVETURERCRCKVETERERiVBK1EREREQilJ76FAnS8r1lPDlnMRXWNmm9R46XNml90nDGwOs/Hut2GCIiXkrURIL03tZSNhw+GLL6u7ROCVndEhxjDMNPaeN2GCIiXkrURIJUVuG83nfJQHp3yGjSuhPiDVndWzdpnSIi0vwpURMJUoXnjueQbq0ZqqQqKlVUWO59aw3GwK8vGeR2OCIiephAJFiViVp8nNbgjGZ/W7yNvy3e5nYYIiKAEjWRoFU+RBBnlKiJiEh4KFETCZJniJp61EREJGyUqIkEqerWp7txiIhI7NA/OSJBKC4tZ88x3foUEZHwUqImEoQZf1vm3U5Ul5qIiISJpucQCcL2g8cAGHtaO7q1SXU5Ggml0zs17Rx5IiKNoURNJAhlngFqD1w2BKNbn1ErLs7w3q0T3A5DRMRL93BEglDuSdTi45WkiYhI+KhHTSQI3kRNvWmu2LS/kMWbfa+zOrpXO07LTAdg/Z6jLNt22Ge5OGO4ctQp3vdvf7mb/KLSWuXapSUxeVDnJohaRKTxlKiJBMGbqGkONVf88PmlbN5/zOexh6YO9SZqn206yK/+tdZnuaT4uJMStdkfbGTD3kKfZTOSE/jqV+c1MmoRkcZToiYShHKrRM1NA7u0YvP+Y1w4pDMtUxJPOtazfZp3u1/HjJOSseoSavzZXTC4MyNOPeGz7Lje7RoZsYhI01CiJhKE8nIlam569MphPHrlsDrLje3dnrG92wdV563n9G1sWCIiIadETaJO/vFSfvPOWg4dK2myOo+XlgNK1EREJLyUqEnUyd2wj1eW7WzyetMSISVBD0q74UDhCax1BvrHKVkWkRiiRE2izolSZ/n0cb3bcd3Ynk1W7+Gta0jQqgSuOO/PCzl4rISl/+8c2qcnux2OiEjYKFGTqFM58L9b6xacM6Bjk9Wbu29dk9UlIiISDHUPSNSpnEpDt8hERKS5U6ImUafCO5WGy4GIiIg0kv4pk6ijVQRERCRaKFGTqKNbnyIiEi30MIE0O8Wl5Ww7eNzv8b1HiwH1qImISPOnRE2anWlPfsaqnfl1louPV6IWLR6aOpSS8goyUvQrS0Rii37rSbOzybM4d5dWKaT7+Yc7NSmBCwd3DmdYEkITT+/gdggiIq5QoibN1n9um0BGjQW6RUREookSNRGJeH9duJmi0nJ+ML4nLZL0a0tEYod+44lIxHvio00cPFbCVWecokRNRGKKpucQERERiVBK1EREREQilBI1ERERkQilwR4S0d5bnce/V+85aV9RablL0YiIiISXEjWJaPe9vY5dR4pq7U9PTiA5Id6FiERERMLHlUTNGHML8EPAAH+11j5sjJnl2bffU+wX1tp3fZw7GXgEiAeestY+EJ6oxQ0l5RUA/OKC0+mQkeLdP6BLS5ISdOc+VrRNS8IY5xeGiEgsCXuiZowZhJOQjQJKgPeMMe94Dv/ZWvtQgHPjgceBc4GdwBfGmLestWtDHLa47NKsrnRomVJ3QYlK799+ltshiIi4wo0uif7AYmvtcWttGfAR8O0gzx0FfGOt3WytLQFeBC4JUZwiIiIirnIjUVsNTDDGtDPGtAAuALp7jv3EGPOlMWauMaaNj3O7Ajuqvd/p2SciIiISdYy1NvwXNeb7wI1AIbAWKAIeAA4AFrgP6Gytvb7GeVOB86y1P/C8/x4wylp7k49rzABmAHTs2HHEiy++WK8YCwsLSU9Pr+cni05utsUtHx4n/4Tl4ZxUWqe4OybNjXbYml/OhzvKqPD8Ne2YZpjSKwmA0grL82tK/J57VrcEerdxHrj4an8Zn+/x/bRsfBxcOzDZ+/7VDSXkn/D9e2Fg+3gGZRSTnp5OXmEF724p9Xv9y/skev/McneUsulIhc9ywXymj3eV0TrZcN+4VDKSImOkmn5HVFFbONQODrVDlYkTJy6z1mY3pg5XHiaw1j4NPA1gjLkf2Gmt3Vt53BjzV+BtH6fupKr3DaAbsNvPNeYAcwCys7NtTk5OvWLMzc2lvudEKzfbIunTBXDiBGPHjnV9jJob7fDD55fy0U7vXw1G9WhLTs4YAI6XlPHD+f/xe+7lZw4iZ5jT4fzNx5v5eNk6n+VSE+N59sYc7/tfL8tl8/5jPsv27dmd9PR95OTksGTzQT7+ZLHf6987bRy9Ozi/rN94cQUf7/L5VzXoz1RUbjg7Z3zELCGl3xFV1BYOtYND7dC03Hrqs4O1dp8x5hTgMmCMMaaztTbPU+TbOLdIa/oC6GOM6QnsAq4ArgpL0BJSBwpPUOxjfrTyivD3+EaSyja5flxPTu+UQfuMJO+xxPg4/nD5EL/nDjultXd7fJ9M/nB5os9y8XEn91D9bFI/CorLfJbt0zGd/M37AOiZmRbw+pkZVb10V4w8hXGntfdZLtjPdHrnjIhJ0kREwsWt33qvGWPaAaXAjdbaw8aYvxljsnBufW4FZgIYY7rgTMNxgbW2zBjzE+A/ONNzzLXWrnHlE0iTeXnpDn7+6pduhxGRKkcm5PTLZELfzJOOJcbH8d2R3X2cVVu/Thn065QRVNnzB3cOeDx3s/PaISMl6OuPOa0dY05rV2e5+nwmEZFY4Natz/E+9n3PT9ndOA8cVL5/F6g1v5o0X+vyjnq3u7ZOrXV8cNdWJ/XOxBKLk6mZyBiWJSIiYab7CBIx7pkygO+f2dPtMCLK7ef25X/GlNC/c0u3QxERERcoUROJYCNObet2CCIi4iKtwSMiIiISodSjJhLBXlm6g91Hirl8RFe6tWnhdjgiIhJmStREItjLS3fwxdbDjO7VVomaiEgM0q1PkQhWOT1HXJwe+xQRiUXqUZOgvfNlHp9uOtDk9S7berjJ64wWFZ5MTWmaiEhsUqImQfvpK6so8rF6QFNpnep75vxYVrkug+ZRExGJTUrUJGgnypwk7deXDCSuiTOHlqmJTB7YqUnrjAbWu4KWMjURkVikRE3q7eozTq21PqSEhnrURERimx4mEIlgrVITaZeWRFK8/qqKiMQi9aiJRLDnrx/ldggiIuIi/TddREREJEIpURMRERGJULr1KQAUl5ZzorTC57FjpZb846Xege0SPt97egkb9xby/PdH0bdjhtvhiIhImClRE77amc/UJxdR7CdRA+CD+eELSLz2F5xgz9FiysqVJouIxCIlasLavHyKSytIjDekJsbXOl5WVkZCgvNVOatfB03NEUaV86hpeg4RkdikRE28vj2sK3/4ztBa+3Nzc8nJyQl/QIL13HBu6gmGRUSkedDDBCIRrEI9aiIiMU2JmkgEs1qUXUQkpilRE4lgVUtIKVUTEYlFGqMmEsGuOeNUDh8voU2LRLdDERERFyhRE3HRjfOWs2Fvgc9j5w/uzO3n9g1zRCIizdi8qbDRx3RSUx6G7Ouc7aXPwNu3+q9jVn7V9pMTIG+V73LDp8PFs53t3StgTk4DAq6bErUYsHpXPm9/mecd71TTuj2+EwUJve2HjrNxX6HPY9kFxWGORkSkmfOVpDVzStRiwK/fXsvnWw7VWS4jRbfXwu2xq4Zxosz3RMOtUvXnISLSINV7xWrKvq6qd60uMxcGV67LMN/X/FXjxxcrUYsBRSXlAEwfcyqdW6f6LJOcEMfFQ7uEMywBTm2X5nYIIiLRI1CC1kwpUYshl4/oxpBurd0OQ4CtB44x6eGFnNq2Be/ffpbb4YiISITS9BwiLrBASVkFpeUB1lcVEZGYp0RNxAX+HuwQEZFGeHKC8xNFdOtTxEWayFZEpAn5m0qjGVOPmogL1J8mIiLBUI9alCmvqJ0CWKUFEUv9aSIiEogStSjy43nLePerPW6HISIiIk1EiVoUWbjhgHc7rkZXzant0jgtMz3MEYk/7dOSufeiAbTUJMMiIhKAErUo9NWsSVplIMK1apHIdeN6uh2GiIhEOCVqIiIiEh2GT3c7gianRE3EBflFpby/di8ZKQmcN7CT2+GIiESHi2e7HUGTU6Im4oJ9R4v56Sur6N0hXYmaiIj4pXnUREREJDrsXuH8RBH1qIm4QDPbiYiEwJwc53VWvqthNCUlas3Y51sOsXjzQe/7kjIt8N3caMJbEREJRIlaM3bD35dx6FjJSfsS4w2J8bqjHem0JruIiARDiVozVniiDIAf5ZxGgmeG26zurUlJjHczLKkHrckuIiKBKFGLArec3UfJmYiISBRSoibigj4d0ll/32T1qImISEBK1ERcEBdnSIlTL6iIiASmRE1ERESiw4xctyNockrURFyw9cAx/veVVZzatgV/mpbldjgiItGhyzC3I2hyStQkZhSXllNe4XtejPg4430gw1rL8ZLy2ueXWY6dKCMpIc47BUpJWQWl5f7nr0tLrvordrykzDstx6HjJSzbdphjnid3RUREfFGi1ozc+I/lfLBur/e9Jritn/99ZRXvfJnn89iYXu14YcZoAApOlDFk1nzflSz4D49eOYyLhnYBYO6nW3jg3+t9Fs1ITuCrX53nfX/+Ix+z7eDxRnwCEZEwmjcVNlb7XTgjt6rH6q2bYflzPk8bkX4a5Cyv2jGrlf9rTHkYsq9ztpc+A2/f6r9s9dUGnpwAeat8lxs+PaoWZ1ei1oz8+6s8anYIDT+lNckJmuA2GMnxcbRI8j2APznx5Db0Va68vJz4+HjvnHUACXHGb52pNfanJsafVNYA5w/qHGz4IiLhtdHPf1gjXYHv/5A3V0rUmqHVvzqPeM+8DimJcRjN8RCUP03L4k9BlGuZksjaX0+utT83N5ecnJyT9v1gfC9+ML5XUNd/79YJQZUTEYkovtbNvHi2316rZbm55NR1vi/Z11X1rtVl5sLgykUBJWrNUGpiPPFxSs6CVV5h+XLnEeLjDEO6tXY7HBERkaDpnplEvRNl5Xz7L4v47pOfuR2KiIhIvahHTaJe5ZOWBvVCiogELQrnJGuOlKhJ1Kt8/kJ3i0VE6iEK5yRrjnTrU6JehadLTQ9diIhIc+NKomaMucUYs9oYs8YYc6tn34PGmPXGmC+NMW8YY1r7OXerMeYrY8xKY8zScMbtlqPFpbzw+fZaU3NIcLy3PpWniYgE762bnR9xVdgTNWPMIOCHwChgKDDFGNMHeB8YZK0dAmwA7gpQzURrbZa1NjvkAUeApz/ewl2vfwVAUkKcRlrVk63sUXM5DhGRZmX5c34ntZXwcaNHrT+w2Fp73FpbBnwEfNtaO9/zHmAx0M2F2CLS0eJSAEb1aMvsK7KI02CreqnqUVO7iYhI8+LGwwSrgd8aY9oBRcAFQM1bmNcDL/k53wLzjTEWeNJaOydkkUaY8wZ1YrJmsq+3jJQE3rn5TOKUqImISDNjKm8LhfWixnwfuBEoBNYCRdba2zzH7gaygcusj+CMMV2stbuNMR1wbpfeZK2tNUWxMWYGMAOgY8eOI1588cV6xVhYWEh6enr9PliIzFt3gve3lXHl6Umc1yMx7NePpLZwk9rBoXZwqB2qqC0c0dYOObmXAJCb82a9zou2dmiMiRMnLmvsMC1Xpuew1j4NPA1gjLkf2OnZng5MAc72laR5zt3ted1njHkDZ6xbrUTN09M2ByA7O9vWXPqnLr6WC3LLRwVrYNtWevfuTc6ZPcN+/UhqCzepHRxqB4faoYrawhF17ZDrvDTnfz+jgVtPfXbwvJ4CXAa8YIyZDNwBXGytPe7nvDRjTEblNjAJ51aqiF/5RaX8/NVV/ObttW6HIiIiUi9uzaP2mjFmLfAv4EZr7WHgMSADeN8z9cYT4NzqNMa86zmvI/CJMWYV8DnwjrX2PRfil2akuLScl5fu5M1Vu90ORUSk+eg81PkRV7l163O8j329/ZTdjfPAAdbazThTeogErfImuh6WFRGph5m1RhWJC7SEVATavL+QGX9bxpHjzrQchSdKXY6oefOuTKCZ1EREpJlRohaBFm8+xDf7Ck/alxhv6N85w6WImrfKp1I0O4eIiDQ3StQi2KVZXfjFhf0BSE2MJyMl/FNzRAOtTCAi0gCzWnle892NI8YpUYtgqUnxdMhIcTuMZk8rE4iISHOlRE2iXlJCHEO7tSJTSa+IiDQzStQk6nVsmcKbPznT7TBERETqza151ERERESkDkrUJOpZaykrr6CiIvzr2oqIiDSGbn1GiMWbD7L9oLNy1tJth1yOJrps2n+Mc/70Eb3ap/Hfn+a4HY6IiDRnFRXwzftQuC8sl1OiFgF2HDrOFXMW19qfFB+eDs8V2w9zvKTc57GOLasG4B8tLuWrnf4f0x7avTXpyc5XauPeAvYVnPBZLj05gaHdWwNQUWH5bPNBv3X27pDujSEvv4jN+4/5LGeAsb3b+/xMu44UVRUSEZHgTHnY7Qgi09aF8I/vhu1yStQiQOUKBK1bJHJu/44AJCfGce24nmG5/i/eWM26vKM+j10z+hTOae1sb9pXyNVPLfFbzzs3n8nALs68O3/9eDMvL93ps9zQ7q1588ZxgLNqQKA6f3/5YKaNPAWABWv3cs+ba3yWi48zbLr/goCfKUFrSIlElnlTYeN8/8erz9/15ATIW+W73PDpcPFsZ3v3CpiT47/OGbnQZZiz/dbNsPw53+U6Dz15CaXKOcV8mfIwZF/nbC99Bt6+1X/Z5viZ5GTHq3UuZF1TR+G/NPpyStQiSLc2qTw4NfxLmWZ1b0WbFr4n0+3VPh3KnC9lRkoCY09r57eetKSqr1PvDul+y56Wme7dNsYErLP6PHIdW6b4LRtXY460mp/JGPhudne/1xERF/S7APK+hMI9bkciUn8Dvw2XPl5HoRAmasaY4UGcX2qt/arRUYirfnfZkIDHc3O3AdC7Qwb/+OHooOqcMeE0Zkw4rc5y8XEm6DonDezEpIGdgipb12cSkQiQfV3wvTbBLhDeZVjwM+lfPLuq16ouwdYZjZ9JXBWoR+0j4AsCj+zpCfRoyoAkfD7euJ+XvtjBmb3bc8WoU9wOR0RERGoIlKh9Ya39VqCTjTH/beJ4JIy2HTzO21/m0TI1kSvcDkZEYs/SZ5xXjYUS8ctvolZXkhZsGREREZ8qB90rURPxK+iHCYwxmcAtQCrwf9bab0IWlYiIiIjU66nPPwJ/ByzwAjAyJBHFgJeX7uCJ3E1UWGem/BNlFS5HJCIi0khFh+HFqxm1bzN8mep2NKFzojCslwv01Od7wG+ttR97diUBW3ESteTQhxa9Xl26k80Hak/c2rdjhgvRiIiINIEdX8C2T2kBUOR2MGHQYUBYLhOoR20acI8x5kfAPZ6fe3Fuff44DLFFLYvTk/bIFVkM6dYagDgD3du0cDEqERGRxnD+bTvSahCt/+d5l2MJsfhEaB2e2RICPUyQD/zUGNML+C2wC7jRs1+aQOdWqfRsn+ba9Tu2TGF0r7b0cjEGERGJLuXxydCu7nk0JTiBbn32An4ElAL/C5wGvGyMeRv4i7XW9+KQ0mycO6Aj5w7o6HYYIiIi4kegVb9fAN4DFgN/s9Z+bK09DzgKBFicTUREJAiz8jU7vkgdAo1RSwG2AGmAd/CUtfY5Y8zLoQ5MQq+krIITZeUkxseRkhjvdjgiIiJSQ6AetR8BDwK/AG6ofsBaGwvPc0S9l5fuYPCs+fz67bVuhyIiIiI+BHqYYBGwKIyxxITdR4r4Yutht8MQEXHfkxOc12AXJ5emUVYCm3Oh9HjT1pu3smnrEyDwwwRzrLUzAp0cTBk52dVPLfFuJ8QHWu9eRCTK5a1yO4LYtOT/4P1fhqx6a+ozl77UJVBrXmqMKQ5w3AATmzieqHeg4AQA07K7M6RrK5ejERGRmFO4z3ntMADa9W7auuMS2J50Bu2bttaYFihR+1kQ539cdxHx5e4p/UmIDzREUEREJISGXgnjbm7yao/m5jZ5nbEs0Bi158IZiIiIiIicTF06IiIiIhFKI/5i2Jm92zP7ymGc2lZrjIqIiESieiVqxpg4IN1aezRE8UgY9WifRg+t8ykibhk+3e0IRCJenYmaMeYfOBPelgPLgFbGmD9Zax8MdXAiIhLFLp7tdgQiES+YHrUB1tqjxpirgXeBO3ASNiVq9bDt4DHuev0rCkvK3A7Fa83ufD795gD9O7dkfJ9Mt8MRkYb64mlY/jxgXbn8iIJC+DrdlWtHkmbTDkd3ux2B1EMwiVqiMSYRuBR4zFpbaoxx57dBM/bBun0s2nQQgMyMZFIjYG3NFduPcP+767nqjFOUqIk0Z4v/Dw5udO3yGQCFrl0+YjS7dmjby+0IJAjBJGpPAluBVcBCY8ypgMao1VOFdXLbi4d24b5LB5EYhjnUFn1zgFU78/lRzmneff9Ysp3Dx0sAJ1ETkShgK5zXy5+GdqcFLhsCS5ctI3vEiPqfOCfHeZ2R25ThuKbB7eCG5JaufFek/upM1Ky1s4HqAwm2GWO0IkEDtU9PplVqYsivU1pewfXPfUFxacVJidqzi7awYe/J/+VLS3K/d09EmkDnLGgfYKb5eVNh43zfx6Y8DNnXOdtLn4G3b/Vfz6z8qu0nJ5Cdt8oZEFPT8OlV49B2r6hKzGrqMsz/tZqRwg35UfNZJHIE8zBBR+B+oIu19nxjzABgDPB0qIOThiuvsBSXVtTaf8XIUzhQeML7PjkhnitGdQ9naCLiFn9Jmpv6THI7ApGIFsytz2eBZ4C7Pe83AC+hRK1ZSE44+Rbr9Wf2dCkSEXFd9Z6wQLKvq+pdq8vMheTm5pKTkxO4XJdhwV9fRLyCGSjV3lr7MlABYK0tw5mqQ0RERERCKJhE7Zgxph2e576NMaMB/bdIREREJMSCufX5v8BbwGnGmE+BTOA7IY0qCh05Xup2CCJSl8L9sGdVUEXbHPoSvomQeRFLjwdX7skJzuvMhaGLRUSaVDBPfS4zxpwF9AMM8LW1VllHPT324TdA1TQd4dA2LanWGDURCWDuJDi0OaiiQwG+DGk09RdXxxPcecEloSISOYJ56nMVzsMDL1lrN4U+pOjULi2Jg8dKmDSwY1iul5IYz/J7zg3LtUSixtE857XnWXUmPYcOHaJt27ZhCCpImf2hTQ+3oxCRJhbMrc+LgWnAy8aYCpyk7WVr7faQRhal+nbMcDsEEanLlS9CUouARb4M5klHEZFGqvO+mLV2m7X2D9baEcBVwBBgS8gjizJac0tERETqK5geNYwxPYDv4vSslQM/D2FMUc2E6TrFpeV866FckhLiyP2ZFpIQERFpjoIZo7YESAReAaZaa4MbaSsnsWF8iKDS7vxiPUwgIiLSjAXTozbdWrs+5JGIiEhoDZ/udgQiUk9+EzVjzDXW2r8DFxhjLqh53Fr7p5BGFqWMCdfNTxGRGioXSBeRZiNQj1qa59XXY4oaGx/AvqPF3PPm6pMmuS0ojpCJMaV5+eA+2L7Y7ShOknXkCGxp7XYYoVFW7HYEIiIn8ZuoWWuf9GwusNZ+Wv2YMWZcSKNq5j78eh//WbO31v6WKQmkJdcxIaVIpRMF8PFDbkdRS2uI7kXk0jIhIdntKEJj9wrntcswd+MQkaAFM0btUWB4EPvEo7zCef3W6R344fhe3v2nZaaRnKBETYJkK6q2p7/tXhw1rFy5kqysLLfDCJ3M0+ue4b+5mpPjvM6K5kxbJLoEGqM2BhgLZBpjbq92qCUQpb/FmlbHlsmMOa2dK9eOjzPcOPE04uP01Gezl9wSeo53OwqvI9vKIyoeEZFoFqhHLQlI95SpPk7tKFqUPeIlxsfxs/NOdzsMERERaYRAY9Q+Aj4yxjxrrd0WxphEREREhMC3Ph+21t4KPGaMqfWUp7X24lAGJo1TXmH5eON+4oxhQt9Mt8MRERGRBgh06/Nvntcmf+zMGHML8EOcFZX+aq192BjTFmfB9x7AVuC71trDPs6dDDyCM07uKWvtA00dXzQoLa/g2me+IDkhjq9/c77b4YiIiEgDBLr1uczz+lHlPmNMG6C7tfbLhl7QGDMIJ0kbBZQA7xlj3vHs+8Ba+4Ax5k7gTuCOGufGA48D5wI7gS+MMW9Za9c2NJ7G2rS/8KT50gC2HTrmUjQSFQ58A0WHoKTQ7UhERMRlwaz1mQtc7Cm7EthvjPnIWnt7oPMC6A8sttYe99T/EfBt4BIgx1PmOSCXGokaTnL3TeV6o8aYFz3nuZKoLdywn/+Z+7nf41qFQOpt04fwt0tP3qfvkTSVGbluRyAi9RTMPGqtrLVHjTE/AJ6x1t5rjGlwjxqwGvitMaYdUARcACwFOlpr8wCstXnGmA4+zu0K7Kj2fidwRiNiaZQdh497t4ed0vqkY8kJcXxnRLcwRyTN3hHPczst2kPbns72gEvci0eiiya6FWl2gknUEowxnYHvAnc39oLW2nXGmN8D7wOFwCog2PWVfHUt+FzOyhgzA5gB0LFjR3Jzc+sVZ2FhYZ3nbNju3PLM6ZbAtQNKax0/unkVuZvrddkmU1LuNEtFRUW9P3tNwbRFLAhHO3TevYF+wO5Ww9jQ+yfOzhIggtpf3weH2qGK2sKhdnCoHZpWMInar4H/AJ9aa78wxvQCNjbmotbap4GnAYwx9+P0jO01xnT29KZ1Bvb5OHUn0L3a+27Abj/XmAPMAcjOzrY5OTn1ijE3N5e6ztm1ZBusXU3nLl3IyRlcr/qrKy4t5yf/WOH3+A/G92R0L2fi3Plr9vDy0p0+yyUnxPH41cO9dfL+e8TFxdX5OeoSTFvEgtzcXHJ2PQ4b5/svVH3G9ycnQN4q3+WGT69aIHv3iqoZ4z265L1Pl7z3nTczcqt6Qt66GZY/57vOzkNh5sJqsbTyH+eUhyH7Omd76TPw9q3+yzbRZzqJPlO1WFz8TE1EvyMcageH2qFp1ZmoWWtfAV6p9n4zcHljLmqM6WCt3WeMOQW4DBgD9ASmAw94Xt/0ceoXQB9jTE9gF3AFcFVjYnHLgrV7KS4r5+zTO2KxLFhXe23QSlOGdPZubz903G/Z1MSqBSMS4+Po0iqF+HiNb2pS/S6AvC+hcI/bkYg0TEGe2xGISD0Ya33eOawqYEw3nLU9x+HcZvwEuMVa67tbJ5iLGvMx0A4oBW631n7gGbP2MnAKsB2Yaq09ZIzpgjMNxwWecy8AHsaZnmOutfa3dV0vOzvbLl26tF4xBvM/gnlLtnH3G6u5ctQp/O6y+vWojb7/A/YcLeazu75FZnoy/13vqwPRMahrK7q0TgVg64FjbNhb4LNcfJzh7P4dve8//eYAp2Wm06lVSr1iq0n/O3KEpR2WPQv/ugWG/w9c/Ghor9VA+j441A5V1BYOtYND7VDFGLPMWpvdmDqCufX5DPAPYKrn/TWefec29KLW2loLBVprDwJn+9i/G+eBg8r37wLvNvTakSghPo5JAzsFVbZH+zR6tE8Lquy43u0bE5aIiIi4LJgVuzOttc9Ya8s8P88CmupeYtPSZ5wfERGRMAimR+2AMeYa4AXP+yuBg6ELqXnYvL+Qu99Y7XYYEm6Vg7krB3j7suE/sOw5sBX1rz9/R91lREQkZgSTqF0PPAb82fP+U8++mPbqsqoheh0ykl2MRCLOR7+HXcsaV0d6cLfCRUQkugXz1Od2nJUJpJryCuchjHG923HjxN4uRyMRpcIzLeBZd0DnrPqfn5AMPc5s0pBERKR5CmYJqV44i6CPxnnq8zPgtsplnGLd+D6ZJCUEM9RPYk6/8zUTvIiINEowtz7/gbMQ+rc976/AGa/m2tJN0WD+7ROwFjKSg/kjEBERkVgUTFeQsdb+rdpTn3/Hz7JNEryWKYm0Sk0kLk4T0oqIiIhvwXTnfGiMuRN4ESdBmwa8Y4xpC2CtPRTC+ERERERiVjCJ2jTP68wa+6/HSdx6NWlEMeLGfyzn8LESHr1yGO3S9dRos1F9PUUREZEQC+apz57hCKS5OXy8pFHnL9t6mD1Hiykpb8BcWxKZThTCke1QWux2JCIiEiU0kr0B9hec4OWlzjxqGmEmAFSUw19G15iwVt8OERFpHCVqDbD7SJF3u/oi6BIDnpzgvM5cePL+suKqJC2zP7TtBR0Hhjc2ERGJOkrUGmFIt1b07pDudhgSTnmrAh9PbAE3Lg5PLCIiEvXqnJ7DOK4xxvzS8/4UY8yo0IcmIiIiEtuCmUftL8AYnMXYAQpwJsAVERERkRAK5tbnGdba4caYFQDW2sPGmKQQxxX1vtW/A/nHS0lJiHc7FBEREYlQwSRqpcaYeDyrERhjMgHNKdFI9397sNshiIiISIQL5tbnbOANoIMx5rfAJ8D9IY1KRERERIKa8HaeMWYZcDbOxFCXWmvXhTyyCHa8pLzRdew4dJwKa+naOpWE+GDyZQlo5Quw6YOQXqL/3r3QxjP/82s/OPlgRVlIry0iIrGpzkTNGHMKcBz4V/V91trtoQwsklXOo1Z9PrX6mvrEZ+w5Wsxnd32Lzq1Smyq02PXO/0LpsZBe4qQZ8w5v8V0orX1IYxARkdgSzBi1d3DGpxkgBegJfA3E7GyeCfHOjPN9O2a4HIl4lXuW9LrkcYgPzbMua9etZUD/AYELdcsOybVFRCQ2BXPr86RR78aY4dReoD2mWOu8ZmZoMfWIM/i7kBCaRG3foVwGtG/lvOkyLCTXEBERqa7eKxNYa5cbY0aGIpjmosKTqWklxxg0J8d5nZXvahgiIhIbghmjdnu1t3HAcGB/yCJqBip71OJMVap21+tf8v7afT7LZ5/ahie+NwKAguJSJj70EQePnQh5nDFt3lTYOL/qffXE6skJ/peCGj4dLp7tbO9eUZWYATk+TxAREQmdYHrUqg/EKsMZs/ZaaMJpHip71Kp3qR0tLuNAoe/kK7+o1LttwVuua+tU2qZp7uCQqJ6kNbU+k0JXt4iISDUBEzXPRLfp1tqfhSmeZsGTpvGf1XsYsPo9PvrZRO7/9mDuvcj3QPOkatNvpCcl8PndZwPQOjWJpARNzRFSvm5RzlwY3Lldhp10fm5uLjk5OU0Tl4iISBD8JmrGmARrbZnn4QGpxnp61MoqLCfKnEUaWqUmAol1nhsXZ+iQkRLK8CJfcT4UH23iSm3dRURERJqZQD1qn+OMR1tpjHkLeAXwTlRlrX09xLFFLKucoOH2rIa/TqyaTkNERET8CmaMWlvgIPAtquZTs0DMJmoVStQa7sDXTpKWkAot2jVt3aflVE3NMeXhpq1bRETEBYEStQ6eJz5XU5WgVYrpVMXW+PhG83TUX7/JMPXZ0NWffV3o6hYREQmTQIlaPJCO7+nCYjpRU4+aiIiIhEOgRC3PWvvrsEXSnGiQWuRb+ozzqp41ERFpxgIlarqh50dlj9rQbq2ZNLAjaUn1XuBBQu3tW51XJWoiItKMBcowzg5bFM1M5fQc/Ttn8IPxvVyORkRERKKV39lWrbWHwhlIc1LZo2b0FIGIiIiEkO7Z1VNBcSkPzf8agG0Hj/HPFbuYPKgTKYnxLkcWgZb/DXYvP3nfoc3uxCIiItIMKVGrp/+s2cvxknIAFm8+xIdf72dZn3OUqNVUnA9v3YTfB4RT24Q1HBERkeZIiVo9nShzkrSkhDhSEuMoKi13OaIIVV4KWEhMg0k1Hh6OT4LTp7gSloiISHOiRK2BLh/ejf+s2eN2GJEvMQVG/sDtKERERJolJWqNYDWfWuSale92BCIiIo3m96lPCZ6e/hQREZFQUKImIiIiEqGUqEl0enKC8yMiItKMaYxaPZSVV1BUUvWU5yd3fAuAFkkxMjVHeSmUFAZXttjlMWJ5q9y9voiISBNQohakE2XlnPOnj9hxqMi7Ly05hpqvOB8ezYZj+9yOREREJGbEUKbROPsLTrDjUBHGQPv0ZM4+vYPbIYXX4a1OkmbiIDkj+PMGTw1ZSCIiItFOiVo9dWmVyqd3Orc8r3/2CwpPlPH09GwyUhJdjixMOg6EGz5xOwoREZGYoEStEZZtO0x+USnlFZpPTURERJqenvoUERERiVDqUZPoNHy62xGIiIg0mhI18W/eVNg4nxyA1X1PPjarlf/zpjwM2dc520ufgbdv9V+2+lJPT07wP63G8Olw8Wxne/cKmJPjv84ZuVVlRUREmjHd+hT/Ns53OwIREZGYph61IMXy+uvbu3+bU9qmwIENVTuDXfQ8+7qq3rW6zFwYXLkuw7TouoiIxAQlakHadcSZ6PZA4QnvvgsGd+J4STmJ8dHdMXnKjjdgh+dNcoBbniIiItKklKgFKc4YANKrrUbwu8uGuBVOWB1uPZg22d9xJrvtd4Hb4YiIiMQMJWr11Cszze0QwqfzMMhbwdGWp9PmzFvdjkZERCTmRPc9uxDbcuAYm/YXRu+Et6df6HYEIiIiMc2VHjVjzG3ADwALfAVcBzwH9PMUaQ0csdZm+Th3K1AAlANl1trs0Efs28WPfUJBcRmr7p1Eq9QYWUJKREREwibsiZoxpitwMzDAWltkjHkZuMJaO61amT8CgR7rm2itPRDiUKUgz+0IREREYppbY9QSgFRjTCnQAthdecAYY4DvAt9yKTaptPRptyMQERGJaWEfo2at3QU8BGwH8oB8a231mVXHA3uttRv9VQHMN8YsM8bMCG20dYimoWkVFVBRfvKPiIiIuMrYMM/kaoxpA7wGTAOOAK8Ar1pr/+45/n/AN9baP/o5v4u1drcxpgPwPnCTtbbWTKmeJG4GQMeOHUe8+OKL9YqzsLCQ9PR07/s3Npbw5qZS+rSO4+7RqQD8aMExisrgL2e3oEWiqVf9kSS9YDNZK/8fCeXHfB7f2PkSdvW7PsxRRZ6a34lYpXZwqB2qqC0cageH2qHKxIkTlzV2LL0btz7PAbZYa/cDGGNeB8YCfzfGJACXASP8nWyt3e153WeMeQMYBdRK1Ky1c4A5ANnZ2TYnJ6deQebm5lL9nMVF62HTJmxSC3JyzgIg4cP/QFkZZ44/k5YpzfhhgqWbYZknSTPVOlltBQDHM4dT3/aLRjW/E7FK7eBQO1RRWzjUDg61Q9NyI1HbDow2xrQAioCzgaWeY+cA6621O32daIxJA+KstQWe7UnAr8MQs9dlw7uG83LhNeJauOiRqveehdcPt81yJRwREZFYF/ZEzVq7xBjzKrAcKANW4On5Aq4AXqhe3hjTBXjKWnsB0BF4w3negATgH9ba98IVe01zrxtJeYWlRWK8WyGIiIhIFHPlqU9r7b3AvT72X+tj327gAs/2ZmBoqOML1sgebd0OQURERKKYlpAS/2bkOq8bAk1pJyIiIqGiJaQa4Q/vree+t9dSXBqlU1l0Geb8iIiIiCuUqDXC859t4+lPtlBaXuF2KCIiIhKFdOszCN/sK+SJjza5HUbTOX4IVv4DSouc97uW+S731s3Oa8vLwhOXiIiInESJWhDufWu1dzstKQqabMkT8NHva+9PqjFB4fLnnNccJWoiIiJuiIKsI/QKTzhj0M4f1IlvV5tHLdyrOjSZE4XOa8+zoNtIZzshBYZ/z72YREREpBYlavUwY0IvnysQeOZ1a376TIKxP3E7ChEREfFDDxOIiIiIRCj1qDVCj/ZpFJWU00z700RERCTCKVFrhHduHu92CCIiIhLFlKg1wM7Dx1m54widW6Uw4tQoXkaqc8Ss1iUiIhKTlKg1wJV/XcyOQ0VcOLhz9CVq86bCxvkwKx9mLnT25ea6GpKIiEis0sMEdThaXMqqHUdO2negoASAId1auRBRIxXug8WP+z++cX74YhEREZGAlKjV4es9Bd7tXu1PnhD2e2NODXc4jZe3qmq76wj34hAREZE6KVEL0rBTWtOqRe051JqtnmfBqWPcjkJEREQCUKIWpPjmOqmtP/FRlHSKiIhEKSVqIiIiIhFKT302wB2T+1FWYUmMV54rIiIioaNErQGuHdfT7RBCZ8rDbkcgIiIiHkrU5GTZ17kdgYiIiHgoUatDYXFZrX1vrtxFhbVcNKQLCZFw+3PvWtjyUXBl968PbSwiIiLSZJSo1eGD9XsBKDxRlbDd+dpXFJWWc97ATpGRqL10DRzaVL9zElJ871/6jPOqnjURERHXKVGrQ2piPAD9OmW4HEkAxfnO66DLIS2z7vJxCTDsGt/H3r7VeVWiJiIi4jolakEa2KWl2yHUbfLvIT2IRE1ERESahQi4bxfZrHU7AhEREYlVStSCZIiylQlEREQk4ilRq4M61ERERMQtStSCFG1LfYqIiEjk08MEdfA1Rm3ZPecAVU+EioiIiISCErUGaJEUIc32xdOw8EE4frDp6pyV33R1iYiISKNESMYhDbL6dSjIc7bb9ITU1q6GIyIiIk1LiVodrI/HCb739BJOlFbw/PdHkRIJtz+/+zfodz7EJ7odiYiIiDQhJWpBMtWeJli69TBFpeVURMoka6ltmi5Je3KC8zpzYdPUJyIiIg2mRK0OkZKLhU3eKrcjEBEREQ9NzxEkzc4hIiIi4aZETURERCRCKVETERERiVAao1aHLQeOAbBmd9X8Yi2S4ikqLQ9vIPOmwtWvVL1/Yjzs+dLZfm7KyWWHT4eLZzvbu1fAnBz/9c7IhS7DnO23bm6qaEVERKQJqEetDh9t2A/Amt1H2XHoOACPXjWMlMQ4EuLC2Hwb58OHv6t6X3QodNfqMyl0dYuIiEjQ1KNWh5YpCRwtLmP9ngLW5h2le9sWDOjckuevP4OkhDDnuR89ABPvcrZHXAf/vc/Z/uUhiPMzn1uXYcGvNnDx7KqeOBEREXGdetQaoHWLJEb1bOt2GI4zb/efpImIiEizpkRNREREJEIpURMRERGJUErURERERCKUEjURERGRCKWnPoP0w/G9OC0zzb0Apjzs3rVFRETEFUrUAiivsBwtLgPgJxN706pFonvBZF938vvFf3EnDhEREQkb3foMYOfh497ttOQImwLDWuc1o5O7cYiIiEjIKFELoDIXapuWyMcbD7CvoNi9YJY+4/xUMp4/uoGXuROPiIiIhJwStSAUlVRw3bNfsGL7EfeCePtW50dERERihhI1ERERkQilRE1EREQkQilRExEREYlQStREREREIpTmUQunYwdg26dVj5M2xJp/Oq9lJ5okJBEREYlcStQC2Hm4CICKxiRW1b1yLWz9uJF1TD/5fVyEze8mIiIiTUaJWgDlngTNAJ/d9S3atEhqXIXH9juvvXIguWXj6gLoOgJatG18PSIiIhKRXEnUjDG3AT8ALPAVcB1wJ/BDwJPN8Atr7bs+zp0MPALEA09Zax8IVZzWk6iN7NmWzq1Sm67iyQ9Ah/5NV5+IiIhEpbA/TGCM6QrcDGRbawfhJFxXeA7/2Vqb5fnxlaTFA48D5wMDgCuNMQNCFWvlDU9jTKguISIiIuKXW099JgCpxpgEoAWwO8jzRgHfWGs3W2tLgBeBS0IUozdT23HoOD98fikrth8O2aXq9OQE50dERERiRtgTNWvtLuAhYDuQB+Rba+d7Dv/EGPOlMWauMaaNj9O7Ajuqvd/p2ReaWD2Z2tHiUt5fu5d9BS4+aZm3yvkRERGRmBH2MWqeBOwSoCdwBHjFGHMN8H/AfTj9WPcBfwSur3m6jyp9PpJpjJkBzADo2LEjubm59YqzsLCQb778CoDSkhIAVq9eTfL+9fWqp7qRx46RBnz+xRccT9tbr3NzPK/1/RxNobCw0JXrRhq1g0Pt4FA7VFFbONQODrVD03LjYYJzgC3W2v0AxpjXgbHW2r9XFjDG/BV428e5O4Hu1d53w89tU2vtHGAOQHZ2ts3JyalXkLm5uQzs1R+WLyUpKQlKShg0aBA5AzvVq56TrEmD4zBq5Mj6P0yQ67zU93M0hdzcXFeuG2nUDg61g0PtUEVt4VA7ONQOTcuNRG07MNoY0wIoAs4GlhpjOltr8zxlvg2s9nHuF0AfY0xPYBfOQwhXhSpQW216jkZ556ew83M4tKXRMYmIiEjsCHuiZq1dYox5FVgOlAErcHq+njLGZOHcytwKzAQwxnTBmYbjAmttmTHmJ8B/cJ4WnWutXROyWD2vh4+XNrySosPwxV+r3iekQnrH+tUxb2rDry8iIiLNlivzqFlr7wXurbH7e37K7gYuqPb+XaDW1B2hULkgQVmFs9G1dQPmUqusJLklTH8LWnWv/yS1GZ2d1z6T6n99ERERaba0MkFAnglve7Thwe8MpUf7tIZXFRcPXYY17NyLZzs/IiIiElPcmketWajsDGubltS4JE1ERESkAZSoBfDxxgMAbNl/zN1Adq9wfkRERCSm6NZnAMdLygAorfA5VVv4zMlxXmfluxqGiIiIhJd61ALwrvXpahQiIiISq9SjFkDlPGoBM7XjhwLfljxR0KQxiYiISOxQohbAjsNFAFQEuvX57IWwb23dlcWpqUVERKR+lD0EkJoYD1Q9/enT0V3Oa4/xgZOxAZc0XWAiIiISE5SoBeJJ0DIzkusuO+1vkNomtPGIiIhITNHDBAF0aOkkaN3bNGBFAhEREZFGUo9aAF08S0ad1iHd3UBm5Lp7fREREXGFErUAvNNzGJcn6Gjo0lMiIiLSrOnWZwBHi0oBOHysxOVIREREJBYpUQvgm32FAHy5y+UVAd662fkRERGRmKJbnwH4XJlg7xp4/14odeZY40Rh6ANZ/pzzevHs0F9LREREIoYStUB8TaC26kX45v2T96W0hsS0sIQkIiIisUOJWgBVDxNU31nhvGZfDwMvc7bb94WEpHCGJiIiIjFAiVpDte0FPce7HYWIiIhEMT1MEEAwa7KLiIiIhIoSNREREZEIpUQtgD4dnRUJxvZu724gnYc6PyIiIhJTNEYtgOSEeADSk11uppkL3b2+iIiIuEI9agEUl5YDNcaolR53JRYRERGJPUrUAnhl2Q4AVuw4UrVz6VzntaI8/AGJiIhITFGiFkCbFs7caCfKqiVlrU9xXrtkhS+QWa2cHxEREYkpGqMWQFGJk6ClJyfWPtimR+MvMG8qbJwPUx6G7OucfUufgbdvbXzdIiIi0uypRy2AhHineY4Wl4bmAhvnB1+2z6TQxCAiIiIRSz1qQTi9U0ZoL1DZm1a5Xf29iIiIxCz1qImIiIhEKCVqIiIiIhFKiVoAqYnOhLentG3hciQiIiISi5SoBbDnaDEAg7p6psYozocj212MSERERGKJHiYIIM5Aha22hNQ3H1QdTGnd+AvMym98HSIiIhK11KMWQGJ8jeapXI2g81BIbR32eERERCS2KFELoKzcArBg3d6TD7Tr40I0IiIiEmuUqLnpyQnOj4iIiIgPGqPmprxVbkcgIiIiEUw9aiIiIiIRSomaiIiISIRSohZAubVVbyoqoCDPvWBEREQk5miMmh9f7CnzbhsMvPUTWDnPs8O4FJWIiIjEEvWo+ZF3rMK7PWVIZziwwXnTrg8MvcKlqERERCSWqEfNj8q7nj+Z2Jt26clVBy79C3Qf1TQXGT69aeoRERGRqKRErQ4hvct58ewQVi4iIiLNnW591uG/6/exZPNBt8MQERGRGKREzY/K5z3X7D7Kpv3HQnOR3SucHxEREREfdOvTTXNynNdZ+a6GISIiIpFJPWp+VJ9CTURERMQNStSCtfMLtyMQERGRGKNErb5S27odgYiIiMQIJWrBMvHOa5seroYhIiIisUOJmh+VQ9Q6ZCTTNi3R1VhEREQkNumpTz8qE7WrzziVyYM6w2v1rGDeVNg4v/b+4dOrJrrtOxk2vNeYMEVERCSKKVGrQ4NXJvCVpNWUcyfYirrLiYiISExSouaPp0ut8EQZxaXlpNT3/M5DndeZC/2X6TIMrn6lIdGJiIhIDFCiVoc5CzczqkdbzqnviYESNBEREZEg6GECP6rPd5tw4kiNPSIiIiKhp0StDmPi1jDhrTEaSyYiIiJh58qtT2PMbcAPcLqpvgKuA+4DLgJKgE3AddbaIz7O3QoUAOVAmbU2O5Sx9jM7iLPlkJgGgy6D+CCbbFYrz6vW8RQREZGGCXuPmjGmK3AzkG2tHQTEA1cA7wODrLVDgA3AXQGqmWitzQplkla51qepvOU5/HtwyWOhupyIiIhILW49TJAApBpjSoEWwG5rbfX5LBYD33ElshoaOjuHiIhIfZWWlrJz506Ki4vdDqXBWrVqxbp169wOI6xSUlLo1q0biYlNP0F+2BM1a+0uY8xDwHagCJhfI0kDuB54yV8VwHxjjAWetNbOCUmctfYoZRMRkdDauXMnGRkZ9OjRA9PgiTzdVVBQQEZGhtthhI21loMHD7Jz50569uzZ5PWHPVEzxrQBLgF6AkeAV4wx11hr/+45fjdQBszzU8U4a+1uY0wH4H1jzHprba25MIwxM4AZAB07diQ3N7decZaUlACGnq0MHHf+8nxTjzpyPK/1vW4kKiwsjIrP0VhqB4fawaF2qKK2cDRFO7Rq1Yp27dpRWFjYNEG5oLy8nIKCArfDCKukpCSOHDkSkr8Hbtz6PAfYYq3dD2CMeR0YC/zdGDMdmAKcba31OR+GtXa353WfMeYNYBRQK1Hz9LTNAcjOzrY5OTn1CvLlr+cDpfTs3B42Qbfu3elWnzpynZf6XjcS5ebmRsXnaCy1g0Pt4FA7VFFbOJqiHdatW0fLli2bJiCXxFqPWqWUlBSGDRvW5PW6MT3HdmC0MaaFcfp1zwbWGWMmA3cAF1trj/s60RiTZozJqNwGJgGrQxFkZZZY1fHcPLugRUREgnXw4EGysrLIysqiU6dOdO3a1fveudPk39KlS7n55pvrdb0ePXowePBgsrKyGDx4MG+++WZjwq9l1qxZPPTQQwD88pe/ZMGCBU1afzi4MUZtiTHmVWA5zi3OFTg9X2uAZJzbmQCLrbU3GGO6AE9Zay8AOgJveI4nAP+w1oZkVfPVB8oB2HqgkHENqWDKw00ZjoiISMi1a9eOlStXAk6Sk56ezk9/+lPv8bKyMhISfKcO2dnZZGdn1/u254cffkj79u35+uuvmTRpEpdcckmD4w/k17/+dUjqDTVXJry11t5rrT3dWjvIWvs9a+0Ja21va213z7QbWdbaGzxld3uSNKy1m621Qz0/A621vw1VjHnHnAlutxw45uyo76DO7OucHxERkWbs2muv5fbbb2fixInccccdfP7554wdO5Zhw4YxduxYvv76a8C59TtlyhTASfKuv/56cnJy6NWrF7Nnz67zOkePHqVNmzbe95deeikjRoxg4MCBzJnjPDdYXl7Otddey6BBgxg8eDB//vOfAdi0aROTJ09mxIgRjB8/nvXr1/v8HK+++irg9OTde++9DB8+nMGDB3vLHzt2jOuvv56RI0cybNiwJu/hawit9elHSjwUVlSbR01ERCSMetz5Tkjq3frAhfU+Z8OGDSxYsID4+HiOHj3KwoULSUhIYMGCBfziF7/gtddeq3XO+vXr+fDDDykoKKBfv3786Ec/8jl9xcSJE7HWsnnzZl5++WXv/rlz59K2bVuKiooYOXIkl19+OVu3bmXXrl2sXu2Mejpy5AgAM2bM4IknnqBPnz4sWbKEH//4x/z3v/8N+Jnat2/P8uXL+ctf/sJDDz3EU089xW9/+1u+9a1vMXfuXI4cOcKoUaM455xzSEtLq3ebNRUlan5UjVFrYKK29BnnVb1qIiLSzE2dOpX4+HgA8vPzmT59Ohs3bsQYQ2lpqc9zLrzwQpKTk0lOTqZDhw7s3buXbt261SpXeetz06ZNnH322eTk5JCens7s2bN54403ANixYwcbN26kX79+bN68mZtuuokLL7yQSZMmUVhYyKJFi5g6daq3zhMnTtT5mS677DIARowYweuvvw7A/Pnzeeutt7zj2oqLi9m+fTv9+/evR2s1LSVqwarvrc+3b3VelaiJiEgDNKTnK1Sq9yjdc889TJw4kTfeeIOtW7f6fdI1OTnZux0fH09ZWVnAa5x22ml07NiRtWvXcvz4cRYsWMBnn31GixYtyMnJobi4mDZt2rBq1Sr+85//8Pjjj/Pyyy/z8MMP07p1a+/YumBVxlc9Nmstr732Gv369atXXaGkRdnroGc9RUREquTn59O1a1cAnn322Sard9++fWzZsoVTTz2V/Px82rRpQ4sWLVi/fj2LFy8G4MCBA1RUVHD55Zdz3333sXz5clq2bEnPnj155ZVXACfZWrVqVYNiOO+883j00UepnCFsxYoVTfPhGkGJWh2Mj4k6REREYtXPf/5z7rrrLsaNG0d5eXmj65s4cSJZWVlMnDiRBx54gI4dOzJ58mTKysoYMmQI99xzD6NHjwZg165d5OTkkJWVxbXXXsvvfvc7AObNm8fTTz/N0KFDGThwYIMfArjnnnsoLS1lyJAhDBo0iHvuuafRn6+xdOuzDmnJ8VBB/W99ioiINGOzZs3yuX/MmDFs2LDB+/6+++4DnAnec3JyKCgoqHVu5eD/mrZu3epzf3JyMv/+9799Hlu+fHmtfT179uS992rP1lU9juq9f9Wvm52d7V1RIDU1lSeffNLndd2iHjU/KtdF+MG4U90NRERERGKWErWALKkf3+/ZVo+aiIiIhJcStQASqXbvvffZ7gUiIiIiMUlj1PyoPntaRVwicT0n1K+CWflNGo+IiIjEHvWoiYiIiEQoJWoiIiIiEUq3Putj3lTYOB9m5EKXYc6+t26G5c/5Lt95KMxcGLbwREREGuPgwYOcfbYzJnvPnj3Ex8eTmZkJwOeff05SUlLA83NzcykrK+Occ86pdezZZ5/lZz/7GV27dqW0tJT+/fvz/PPP06JFiyaLPz09ncLCQnbv3s3NN9/sXYS9OVOPWn1snF+/8ukdQxOHiIhICLRr146VK1eycuVKbrjhBm677Tbv+7qSNHAStSVLlvg9Pm3aNFauXMmaNWtISkripZdeasrwvbp06RIVSRooUfPLBlqLvbI3DeDi2c6DA75+rn4l5HGKiIiE0rJlyzjrrLMYMWIE5513Hnl5eQDMnj2bAQMGMGTIEK644gq2bt3KE088weOPP05WVhYff/yx3zrLyso4duwYbdq0AeBf//oXZ5xxBsOGDeOcc85h7969AHz00UdkZWWRlZXFsGHDKCgoAODBBx9k5MiRDBkyhHvvvbdW/Vu3bmXQoEGA05N32WWXMXnyZPr06cPPf/5zb7n58+czZswYhg8fztSpUyksLGyaRmtCuvXpR3E5dOSo22GIiEgM63HnO36P3f/twVx1xikA/GPJdn7xxld+yzZ0gXdrLTfddBNvvvkmmZmZvPTSS9x9993MnTuXBx54gC1btpCcnMyRI0do3bo1N9xwA4mJidx9990+63vppZf45JNPyMvLo2/fvlx00UUAnHnmmSxevBhjDE899RR/+MMf+OMf/8hDDz3E448/zrhx4ygsLCQlJYX58+ezceNGPv/8c6y1XHzxxSxcuJAJE/zPzrBy5UpWrFhBcnIy/fr146abbiI1NZXf/OY3LFiwgLS0NH7/+9/zpz/9iV/+8pcNaqtQUaLmQ+GJMgC+G58LQFxFqXvBiIiIuOTEiROsXr2ac889F4Dy8nI6d+4MwJAhQ7j66qu59NJLufTSS4Oqb9q0aTz22GNYa7nxxht58MEHufPOO9m5cyfTpk0jLy+PkpISevbsCcC4ceO4/fbbufrqq7nsssvo1q0b8+fPZ/78+Qwb5tzdKiwsZOPGjQETtbPPPptWrVoBMGDAALZt28aRI0dYu3Yt48aNA6CkpIQxY8Y0qJ1CSYmaDyVlFQAkG0+C1neyi9GIiEisCrYn7KozTvH2rjUlay0DBw7ks88+q3XsnXfeYeHChbz11lvcd999rFmzJuh6jTFcdNFFPProo9x5553cdNNN3H777Vx88cXk5uZ61+i88847ufDCC3n33XcZPXo0CxYswFrLXXfdxcyZM4O+XnJysnc7Pj6esrIyrLWce+65vPDCC0HX4waNUQsgJcHTPN2y3Q1ERETEBcnJyezfv9+bqJWWlrJmzRoqKirYsWMHEydO5A9/+ANHjhyhsLCQjIwM7ziyunzyySecdtppAOTn59O1a1cAnnuuaiaFTZs2MXjwYO644w6ys7NZv3495513HnPnzvWOJ9u1axf79u2r92cbPXo0n376Kd988w0Ax48fP2mx+UihHrUAKixgoKiknFRwpuUQERGJEXFxcbz66qvcfPPN5OfnU1ZWxq233krfvn255ppryM/Px1rLbbfdRuvWrbnooou47LLLeO+993j00UcZP378SfVVjlGrqKigW7duPPvsswDMmjWLqVOn0rVrV0aPHs2WLVsAePjhh/nwww+Jj49nwIABnH/++SQnJ7Nu3Trvbcr09HT+/ve/06FDh3p9tszMTJ599lmuvPJKTpw4AcBvfvMb+vbt28hWa1rGBny8MTpkZ2fbpUuXBl3+0LESht/3Pj9PfJkfx/+THVn/S/dLI2twYTjl5uaSk5PjdhiuUzs41A4OtUMVtYWjKdph3bp19O/fv2kCcklBQQEZGRluhxF2vv7sjDHLrLWNui2nW58iIiIiEUqJWn28dbPzIyIiIhIGStSCYTyvy5/zv1yUiIiISBNToubDkeMlAIwxqwEwMTCOT0RERCKPEjUfDhQ6idou2w6AuPJiN8MRERGRGKVELYCEOOee5/G2zfsJHBEREWmelKgFkJHiTDN3arsWLkciIiISegcPHvQugt6pUye6du3qfV9SUhLw3KVLl3LzzcE/cPfII49w6623et/PnDmTc845x/v+0UcfDVjfE088wfPPPx/wGs8++yw/+clPfB67//77g441mPpCRRPeBmCM06OWGKd8VkREol+7du1YuXIl4ExCm56ezk9/+lPv8bKyMhISfKcO2dnZZGdnB70ywdixY5k3b573/cqVK6moqKC8vJz4+HgWLVoUcA3RG264Iajr+HP//ffzi1/8olF1hIMykProPNT5ERERiRHXXnstt99+OxMnTuSOO+7g888/Z+zYsQwbNoyxY8fy9ddfA86Ev1OmTAGcJO/6668nJyeHXr16MXv27Fr1Dhs2jA0bNlBUVER+fj4tWrQgKyuLr776CoBFixYxduxYNm3axOTJkxkxYgTjx49n/fr13ms89NBDAHzxxRcMGTKEMWPG8LOf/YxBgwZ5r7N7924mT55Mnz59+PnPfw44a4gWFRWRlZXF1VdfDcDf//53Ro0aRVZWFjNnzqS8vByAZ555hr59+3LWWWfx6aefhqKJA1KPWgDHS8oA2HO0mE4AMxe6Go+IiMSQWa1CVG9+vU/ZsGEDCxYsID4+nqNHj7Jw4UISEhJYsGABv/jFL3jttddqnbN+/Xo+/PBDCgoK6NevHz/60Y9ITEz0Hk9ISCArK4svvviCoqIizjjjDPr06cOiRYvo0KED1lq6d+/O2WefzRNPPEGfPn1YsmQJP/7xj/nvf/970rWuu+465syZw9ixY7nzzjtPOrZy5UpWrFhBcnIy/fr146abbuKBBx7gscce8/Yerlu3jpdeeolPP/2UxMREfvzjHzNv3jzOPfdc7r33XpYtW0arVq2YOHEiw4YNq3f7NYYStQDKyi2YqoRNREQkFk2dOpX4+HjAWUB9+vTpbNy4EWMMpaWlPs+58MILSU5OJjk5mQ4dOrB37166det2Uplx48axaNEiioqKGDNmDH369OH+++8nMzOTsWPHUlhYyKJFi5g6dar3nMp1OSsdOXKEgoICxo4dC8BVV13F22+/7T1+9tln06qVk/QOGDCAbdu20b1795Pq+OCDD1i2bBkjR44EoKioiA4dOrBkyRJycnLIzMwEYNq0aWFfuF2Jmg/lFc68aV044HIkIiISsxrQ8xUqaWlp3u177rmHiRMn8sYbb7B161a/65smJyd7t+Pj4ykrq93pMXbsWJ588kmKi4u58cYbyczMZO3atWRmZjJu3DgqKipo3bq1t+fLl7rWLA8mDmst06dP53e/+91J+//5z396x6u7RWPUfDhQ6GTr8Th/mHGlRc6BWa1C1xUtIiLSDOTn59O1a1fAeQqyMcaOHcvixYvZv38/HTp0wBhDZmYmb775JmPHjqVly5b07NmTV155BXASqlWrVp1UR5s2bcjIyGDx4sUAvPjii0FdOzEx0dsbePbZZ/Pqq6+yb98+AA4dOsS2bds444wzyM3N5eDBg5SWlnrjCCclaj6kJjrdu4dtBgBlqe3cDEdERCRi/PznP+euu+5i3Lhx3gH3DdWmTRsyMzMZOHCgd9+YMWPYt28fQ4c6D+/NmzePp59+mqFDhzJw4EDefPPNWvU8/fTTzJgxgzFjxmCt9d7qDGTGjBkMGTKEq6++mgEDBvCb3/yGSZMmMWTIEM4991zy8vLo3Lkzs2bNYsyYMZxzzjkMHz68UZ+3IUxdXYbRIDs72y5dujTo8gvW7uUHzy/lhZTfMYav2Hje8/QZc0lVb1oEdUeHQ25urt+u7ViidnCoHRxqhypqC0dTtMO6devo3795T7JeUFBARkZGWK9ZWFhIeno6AA888AB5eXk88sgjYY3B15+dMWaZtTa7MfVqjJoPNVNXg7v3p0VERMS/d955h9/97neUlZVx6qmnNvqWbCRRohZAYpyBCmiRrGYSERGJVNOmTWPatGluhxESGqMWQIskZ6xal1apMG9qHaVFREREmpa6inyoHLdXWl5RtbPfBZD3JXQe4lJUIiIiEmuUqAVwrLgU4gFjIPs650dEREQkTHTrMyh6mEBERETCT4maD5VPfZ6Uni19xvkRERGJUgcPHiQrK4usrCw6depE165dve9LSkrqPD83N5clS5bU2m+tpX379hw+fBiAvLw8jDF88skn3jKZmZkcPHjQb92VS0QF0qNHDw4cqL2qUG5uLosWLarz/GDrCyclasEwBt6+1fkRERGJUu3atWPlypWsXLmSG264gdtuu837Pikpqc7z/SVqxhjOOOMMPvvsMwAWLVrEsGHDvMnT119/Tfv27WnXzv8E8w1JtKrH1Zjz3aRELQBjon8yYBERkUCWLVvGWWedxYgRIzjvvPPIy8sDYPbs2QwYMIAhQ4ZwxRVXsHXrVp544gkef/xxsrKy+Pjjj0+qp3IBdnCSrttvv/2kxK2yx+zBBx9k5MiRDBkyhHvvvdd7fuWEthUVFfz4xz9m4MCBTJkyhQsuuIBXX33VW+7RRx9l+PDhDB48mPXr13vj+vOf/+yNa//+/Vx++eWMHDmSkSNH8umnnwJOj+KkSZMYNmwYM2fOrHMd0XDQwwQ+1P5z0Rg1ERFxQaD1pac8XPWQ29JnAt/1aeCKOtZabrrpJt58800yMzN56aWXuPvuu5k7dy4PPPAAW7ZsITk5mSNHjtC6dWtuuOEGEhMTufvuu2vVNXbsWH79618D8Pnnn/OrX/2Khx9+GHAStXHjxjF//nw2btzI559/jrWWiy++mIULFzJhwgRvPa+//jpbt27lq6++Yt++ffTv35/rr7/ee7x9+/YsX76cv/zlLzz00EM89dRT3HDDDaSnp/PTn/4UgKuuuorbbruNM888k+3bt3Peeeexbt06fvWrX3HmmWfyy1/+knfeeYc5c+Y0qN2akhK1AEytNQpERERix4kTJ1i9ejXnnnsuAOXl5XTu3BnAu07mpZdeyqWXXlpnXaNGjWLFihUcO3aM0tJS0tPT6dWrF9988w2LFi3if//3f3nqqaeYP38+w4YNA5yloTZu3HhSovbJJ58wdepU4uLi6NSpExMnTjzpOpdddhkAI0aM4PXXX/cZy4IFC1i7dq33/dGjRykoKGDhwoXecy688ELatGkTZEuFjhK1AFqlJsIJnDFqIiIi4RZsT1iIppCy1jJw4EDvLcrq3nnnHRYuXMhbb73Ffffdx5o1awLW1aJFC3r37s3cuXO9i5uPHj2ad999l3379tGvXz+stdx1113MnDkzYEyBJCcnAxAfH09ZWZnPMhUVFXz22WekpqbWOmYi7N98jVHzyfkSxEXWn5WIiEhYJScns3//fm+iVlpaypo1a6ioqGDHjh1MnDiRP/zhDxw5coTCwkIyMjIoKCjwW9+4ceN4+OGHGTNmDABjxozhkUceYfTo0RhjOO+885g7dy6FhYUA7Nq1i3379p1Ux5lnnslrr71GRUUFe/fuJTc3t87PUTOuSZMm8dhjj3nfr1y5EoAJEyYwb948AP797397n1J1kxK1AIyPLRERkVgRFxfHq6++yh133MHQoUPJyspi0aJFlJeXc8011zB48GCGDRvGbbfdRuvWrbnooot4++23fT5MAE6itnnzZm+iNnz4cHbu3Ol9kGDSpElcddVVjBkzhsGDB/Od73ynVuJ3+eWX061bNwYNGsTMmTM544wzaNUqwFg+4KKLLuKNN97wxjV79myWLl3KkCFDGDBgAE888QQA9957LwsXLmT48OHMnz+fU045pSmasVF06zOAwuLSqjcNHIgpIiLSHM2aNcu7vXDhwlrHq8+BVqlv37589tlnZGRk+Kxz6tSpJ926TE5O5sSJEyeVueWWW7jllltqnVvZyxYXF8dDDz1Eeno6Bw8eZNSoUQwePBiArVu3estnZ2d7e9v69u3Ll19+eVJ9L730Uq1rtGvXjvnz53vf//nPf/b5OcJJiZoPld+hsooKp88xwu5Xi4iIxLIpU6Zw5MgRSkpKuOeee+jUqZPbIYWMErWgKFETERGJFMGMS4sWGqMWwEnp2ZMTnB8RERGRMFGPmg+157s1kLfKjVBERCTGWGsjbooICSyUKxioRy0ATXgrIiLhlJKSwsGDByNi6SIJjrWWgwcPkpKSEpL61aMWFP3PRkREQq9bt27s3LmT/fv3ux1KgxUXF4csaYlUKSkpdOvWLSR1u5KoGWNuA36Ac5fxK+A6oAXwEtAD2Ap811pba6Y5Y8xk4BEgHnjKWvtAU8dX+R+ZpHjj4z6oiIhIaCQmJtKzZ0+3w2iU3Nxc7xJQ0nhhv/VpjOkK3AxkW2sH4SRcVwB3Ah9Ya/sAH3je1zw3HngcOB8YAFxpjBkQqljTkhMqLxyqS4iIiIj45dYYtQQg1RiTgNOTthu4BHjOc/w54FIf540CvrHWbrbWlgAves4LCY1RExERETeF/dantXaXMeYhYDtQBMy31s43xnS01uZ5yuQZYzr4OL0rsKPa+53AGaGKtaKiMlEzMHx6qC4jIiIi4lPYEzVjTBucXrCewBHgFWPMNcGe7mOfz24vY8wMYIbnbaEx5ut6htq+NxwA4Fejqu1+tJ7VRIX2VLZFbFM7ONQODrVDFbWFQ+3gUDtU6dfYCtx4mOAcYIu1dj+AMeZ1YCyw1xjT2dOb1hnY5+PcnUD3au+74dw2rcVaOweY09AgjTFLrbXZDT0/mqgtHGoHh9rBoXaoorZwqB0caocqxpilja3DjTFq24HRxpgWxpnR72xgHfAWUHl/cTrwpo9zvwD6GGN6GmOScB5CeCsMMYuIiIiEnRtj1JYYY14FlgNlwAqcnq904GVjzPdxkrmpAMaYLjjTcFxgrS0zxvwE+A/O06JzrbVrwv0ZRERERMLBlXnUrLX3AvfW2H0Cp3etZtndwAXV3r8LvBvSAB0Nvm0ahdQWDrWDQ+3gUDtUUVs41A4OtUOVRreF0TIVIiIiIpFJa32KiIiIRKiYT9SMMZONMV8bY74xxvhaDcEYY2Z7jn9pjBnuRpyhZIzpboz50Bizzhizxhhzi48yOcaYfGPMSs/PL92INRyMMVuNMV95PmetJ3Zi5DvRr9qf9UpjzFFjzK01ykTld8IYM9cYs88Ys7ravrbGmPeNMRs9r238nBvw90lz46ctHjTGrPd8998wxrT2c27Av0fNiZ92mGWM2VXt+3+Bn3Oj5jvhpx1eqtYGW40xK/2cG03fB5//Zobs94S1NmZ/cB5I2AT0ApKAVcCAGmUuAP6NM4fbaGCJ23GHoB06A8M92xnABh/tkAO87XasYWqPrUD7AMej/jtR4/PGA3uAU2PhOwFMAIYDq6vt+wNwp2f7TuD3ftop4O+T5vbjpy0mAQme7d/7agvPsYB/j5rTj592mAX8tI7zouo74asdahz/I/DLGPg++Pw3M1S/J2K9Ry2YJakuAZ63jsVAa+PM8xY1rLV51trlnu0CnOlSurobVUSL+u9EDWcDm6y129wOJBystQuBQzV2R9wSd+Hgqy2stfOttWWet4tx5rOMan6+E8GIqu9EoHbwTLf1XeCFsAblggD/Zobk90SsJ2q+lqSqmaAEUyZqGGN6AMOAJT4OjzHGrDLG/NsYMzC8kYWVBeYbY5YZZ4WLmmLqO4EzX6G/X76x8p04aYk7INgl7qL5ewFwPU7vsi91/T2KBj/x3AKe6+c2Vyx9J8YDe621G/0cj8rvQ41/M0PyeyLWE7VglqQKetmq5s4Ykw68BtxqrT1a4/BynFtfQ3HW0fpnmMMLp3HW2uHA+cCNxpgJNY7H0nciCbgYeMXH4Vj6TgQjZr4XAMaYu3Hmwpznp0hdf4+au/8DTgOygDyc2341xdJ34koC96ZF3fehjn8z/Z7mY1/A70SsJ2rBLEkV9LJVzZkxJhHnCzfPWvt6zePW2qPW2kLP9rtAojGmfZjDDAvrzN2HtXYf8AZOV3V1MfGd8DgfWG6t3VvzQCx9J/AscQdgmmCJu+bOGDMdmAJcbT0Db2oK4u9Rs2at3WutLbfWVgB/xffni4nvhDEmAbgMeMlfmWj7Pvj5NzMkvydiPVELZkmqt4D/8TzpNxrIr+zajBaesQVPA+ustX/yU6aTpxzGmFE4352D4YsyPIwxacaYjMptnIHTq2sUi/rvRDV+/5ccK98JDy1x52GMmQzcAVxsrT3up0wwf4+atRrjUr+N788XE98JnDW811trd/o6GG3fhwD/Zobm94TbT0+4/YPzBN8GnKcw7vbsuwG4wbNtgMc9x78Cst2OOQRtcCZO1+uXwErPzwU12uEnwBqcJ1QWA2PdjjtEbdHL8xlXeT5vTH4nPJ+zBU7i1aravqj/TuAkpnlAKc7/fr8PtAM+ADZ6Xtt6ynYB3q12bq3fJ835x09bfIMzxqbyd8UTNdvC39+j5vrjpx3+5vn7/yXOP7Sdo/074asdPPufrfy9UK1sNH8f/P2bGZLfE1qZQERERCRCxfqtTxEREZGIpURNREREJEIpURMRERGJUErURERERCKUEjURERGRCKVETUTCwhhTboxZWe2nR4CyhWEMzS9jTBdjzKue7SxjzAXVjl1sjLkzRNfNMcbkG2Pe9bzv51l6Z5UxZoxnX4IxZoExpkW18+YZYw4ZY74TirhEJPw0PYeIhIUxptBam97UZcPFGHMtzpx5PwnDtXKAn1prp3je/wlnTc2twAPW2suNMTcBR621z9U491ngbWvtq6GOU0RCTz1qIuIKY0y6MeYDY8xyY8xXxphLfJTpbIxZ6OmBW22MGe/ZP8kY85nn3Fc8a+7VPDfXGPOwMWaR59xRnv1tjTH/9CymvdgYM8Sz/6xqvX0rjDEZxpgennOTgF8D0zzHpxljrjXGPGaMaWWM2WqMifPU08IYs8MYk2iMOc0Y856nN+xjY8zpnjJTPfWuMsYsDKK5SoFUnEmIS40xrYGLgOcb0PQi0owkuB2AiMSMVGPMSs/2FmAq8G1r7VHjrBG62Bjzlj25m/8q4D/W2t8aY+KBFp6y/w84x1p7zBhzB3A7TiJVU5q1dqxxFoCeCwwCfgWssNZeaoz5Fk6ykwX8FLjRWvupJ/ErrqzEWltijPkl1XrUPD1sWGvzjTGrgLOAD3ESqP9Ya0uNMXNwZmzfaIw5A/gL8C3gl8B51tpdnqSrLo974kwGZnrO/63VLRGRqKdETUTCpcham1X5xjiLGt/vSaIqgK5AR2BPtXO+AOZ6yv7TWrvSGHMWMAD41FlyjyTgMz/XfAHAWrvQGNPSkxSdCVzu2f9fY0w7Y0wr4FPgT8aYecDr1tqdnvqD8RIwDSdRuwL4iyfZGwu8Uq2eZM/rp8CzxpiXgdepg7V2O5ADYIzpjbMkzXpjzN88n/8ea+2GYIMVkeZDiZqIuOVqIBMY4el92gqkVC/gSbAmABcCfzPGPAgcBt631l4ZxDVq9jhZnLVaa5Wz1j5gjHkHZx2+xcaYc6jWq1aHt4DfGWPaAiOA/wJpwJHqyWm1i93g6WG7EFhpjMmy1ga7oP1vcXoUbwbm4YxbuxenPUUkymiMmoi4pRWwz5OkTQROrVnAGHOqp8xfgaeB4TgLwI/z9CxVjgnr6+ca0zxlzgTyrbX5wEI8SY1n0P4Bz+3X06y1X1lrfw8sBU6vUVcBkOHrItbaQuBz4BGcgfzl1tqjwBZjzFTPtYwxZqhn+zRr7RJr7S+BA0D3uhrLc95ZwC5r7Uac8WoVQLlnW0SikHrURMQt84B/GWOWAiuB9T7K5AA/M8aUAoXA/1hr93vGh71gjKm8lfj/AF+3/g4bYxYBLYHrPftmAc8YY74EjgPTPftv9SSM5cBanKcsO1er60PgTs84u9/5uNZLwCuemCtdDfyfMeb/AYnAi8Aq4EFjTB+c3r0PPPsCMs790/8HfNezaw5OGyYAP6rrfBFpnjQ9h4hEJWNMLs4UF0vdjqW+ak7PUc9zn0XTc4hEDd36FBGJPCXAIOOZ8DZYngchziL4sXUiEuHUoyYiIiISodSjJiIiIhKhlKiJiIiIRCglaiIiIiIRSomaiIiISIRSoiYiIiISoZSoiYiIiESo/w/gZPQh14IzugAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\n",
    "plot_roc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\n",
    "\n",
    "plot_roc(\"Train Weighted\", train_labels, train_predictions_weighted, color=colors[1])\n",
    "plot_roc(\"Test Weighted\", test_labels, test_predictions_weighted, color=colors[1], linestyle='--')\n",
    "\n",
    "\n",
    "plt.legend(loc='lower right')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plotting AUPRC "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x176a9d610>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_prc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\n",
    "plot_prc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\n",
    "\n",
    "plot_prc(\"Train Weighted\", train_labels, train_predictions_weighted, color=colors[1])\n",
    "plot_prc(\"Test Weighted\", test_labels, test_predictions_weighted, color=colors[1], linestyle='--')\n",
    "\n",
    "\n",
    "plt.legend(loc='lower right')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Another way of dealing with an imbalanced dataset it using **oversampling**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "pos_features = train_features[bool_train_labels]\n",
    "neg_features = train_features[~bool_train_labels]\n",
    "\n",
    "pos_labels = train_labels[bool_train_labels]\n",
    "neg_labels = train_labels[~bool_train_labels]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(181965, 29)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ids = np.arange(len(pos_features))\n",
    "choices = np.random.choice(ids, len(neg_features))\n",
    "\n",
    "res_pos_features = pos_features[choices]\n",
    "res_pos_labels = pos_labels[choices]\n",
    "\n",
    "res_pos_features.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(363930, 29)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resampled_features = np.concatenate([res_pos_features, neg_features], axis=0)\n",
    "resampled_labels = np.concatenate([res_pos_labels, neg_labels], axis=0)\n",
    "\n",
    "order = np.arange(len(resampled_labels))\n",
    "np.random.shuffle(order)\n",
    "resampled_features = resampled_features[order]\n",
    "resampled_labels = resampled_labels[order]\n",
    "\n",
    "resampled_features.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# easiest way is to use td.data \n",
    "BUFFER_SIZE = 100000\n",
    "\n",
    "def make_ds(features, labels):\n",
    "  ds = tf.data.Dataset.from_tensor_slices((features, labels))#.cache()\n",
    "  ds = ds.shuffle(BUFFER_SIZE).repeat()\n",
    "  return ds\n",
    "\n",
    "pos_ds = make_ds(pos_features, pos_labels)\n",
    "neg_ds = make_ds(neg_features, neg_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Features:\n",
      " [-5.          5.         -5.          2.35104447 -3.52476591  4.36097658\n",
      " -5.         -5.         -0.35826554 -4.64158739  4.31457878 -4.61978138\n",
      " -1.91911153 -5.         -0.24895403 -5.         -5.         -3.76234308\n",
      "  0.06124262 -4.54528794  5.         -5.          5.         -1.05886965\n",
      "  0.50587419 -0.22615014  3.1688201   2.90940077 -1.4570092 ]\n",
      "\n",
      "Label:  1\n"
     ]
    }
   ],
   "source": [
    "for features, label in pos_ds.take(1):\n",
    "  print(\"Features:\\n\", features.numpy())\n",
    "  print()\n",
    "  print(\"Label: \", label.numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /Users/johnnydevriese/miniforge3/envs/pytorch_m1/lib/python3.8/site-packages/tensorflow/python/data/experimental/ops/interleave_ops.py:260: RandomDataset.__init__ (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.random(...)`.\n"
     ]
    }
   ],
   "source": [
    "resampled_ds = tf.data.experimental.sample_from_datasets([pos_ds, neg_ds], weights=[0.5, 0.5])\n",
    "resampled_ds = resampled_ds.batch(BATCH_SIZE).prefetch(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.50048828125\n"
     ]
    }
   ],
   "source": [
    "for features, label in resampled_ds.take(1):\n",
    "  print(label.numpy().mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "278.0"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resampled_steps_per_epoch = np.ceil(2.0*neg/BATCH_SIZE)\n",
    "resampled_steps_per_epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# resampled_model = make_model()\n",
    "# resampled_model.load_weights(initial_weights)\n",
    "\n",
    "# # Reset the bias to zero, since this dataset is balanced.\n",
    "# output_layer = resampled_model.layers[-1] \n",
    "# output_layer.bias.assign([0])\n",
    "\n",
    "val_ds = tf.data.Dataset.from_tensor_slices((val_features, val_labels)).cache()\n",
    "val_ds = val_ds.batch(BATCH_SIZE).prefetch(2) \n",
    "\n",
    "# resampled_history = resampled_model.fit(\n",
    "#     resampled_ds,\n",
    "#     epochs=EPOCHS,\n",
    "#     steps_per_epoch=resampled_steps_per_epoch,\n",
    "#     callbacks=[early_stopping],\n",
    "#     validation_data=val_ds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "20/20 [==============================] - 1s 36ms/step - loss: 1.3586 - true positive: 8496.0000 - false positive: 9455.0000 - true negative: 67781.0000 - false negative: 12190.0000 - accuracy: 0.7790 - precision: 0.4733 - recall: 0.4107 - auc: 0.7840 - prc: 0.5494 - val_loss: 0.7087 - val_true positive: 60.0000 - val_false positive: 20152.0000 - val_true negative: 25335.0000 - val_false negative: 22.0000 - val_accuracy: 0.5573 - val_precision: 0.0030 - val_recall: 0.7317 - val_auc: 0.6980 - val_prc: 0.0374\n",
      "Epoch 2/1000\n",
      "20/20 [==============================] - 0s 17ms/step - loss: 0.9108 - true positive: 11764.0000 - false positive: 8948.0000 - true negative: 11256.0000 - false negative: 8992.0000 - accuracy: 0.5620 - precision: 0.5680 - recall: 0.5668 - auc: 0.5971 - prc: 0.7123 - val_loss: 0.6555 - val_true positive: 70.0000 - val_false positive: 17179.0000 - val_true negative: 28308.0000 - val_false negative: 12.0000 - val_accuracy: 0.6227 - val_precision: 0.0041 - val_recall: 0.8537 - val_auc: 0.8597 - val_prc: 0.3722\n",
      "Epoch 3/1000\n",
      "20/20 [==============================] - 0s 18ms/step - loss: 0.6840 - true positive: 13926.0000 - false positive: 8060.0000 - true negative: 12429.0000 - false negative: 6545.0000 - accuracy: 0.6434 - precision: 0.6334 - recall: 0.6803 - auc: 0.7110 - prc: 0.7834 - val_loss: 0.5823 - val_true positive: 71.0000 - val_false positive: 12970.0000 - val_true negative: 32517.0000 - val_false negative: 11.0000 - val_accuracy: 0.7151 - val_precision: 0.0054 - val_recall: 0.8659 - val_auc: 0.9054 - val_prc: 0.5553\n",
      "Epoch 4/1000\n",
      "20/20 [==============================] - 0s 18ms/step - loss: 0.5660 - true positive: 14970.0000 - false positive: 6611.0000 - true negative: 13729.0000 - false negative: 5650.0000 - accuracy: 0.7007 - precision: 0.6937 - recall: 0.7260 - auc: 0.7887 - prc: 0.8392 - val_loss: 0.5113 - val_true positive: 72.0000 - val_false positive: 9034.0000 - val_true negative: 36453.0000 - val_false negative: 10.0000 - val_accuracy: 0.8015 - val_precision: 0.0079 - val_recall: 0.8780 - val_auc: 0.9324 - val_prc: 0.6380\n",
      "Epoch 5/1000\n",
      "20/20 [==============================] - 1s 66ms/step - loss: 0.4874 - true positive: 15877.0000 - false positive: 5857.0000 - true negative: 14765.0000 - false negative: 4461.0000 - accuracy: 0.7481 - precision: 0.7305 - recall: 0.7807 - auc: 0.8447 - prc: 0.8785 - val_loss: 0.4485 - val_true positive: 74.0000 - val_false positive: 5967.0000 - val_true negative: 39520.0000 - val_false negative: 8.0000 - val_accuracy: 0.8689 - val_precision: 0.0122 - val_recall: 0.9024 - val_auc: 0.9488 - val_prc: 0.6923\n",
      "Epoch 6/1000\n",
      "20/20 [==============================] - 1s 35ms/step - loss: 0.4238 - true positive: 16870.0000 - false positive: 4773.0000 - true negative: 15704.0000 - false negative: 3613.0000 - accuracy: 0.7953 - precision: 0.7795 - recall: 0.8236 - auc: 0.8867 - prc: 0.9126 - val_loss: 0.3937 - val_true positive: 75.0000 - val_false positive: 3796.0000 - val_true negative: 41691.0000 - val_false negative: 7.0000 - val_accuracy: 0.9165 - val_precision: 0.0194 - val_recall: 0.9146 - val_auc: 0.9575 - val_prc: 0.7283\n",
      "Epoch 7/1000\n",
      "20/20 [==============================] - 1s 28ms/step - loss: 0.3744 - true positive: 17198.0000 - false positive: 3856.0000 - true negative: 16829.0000 - false negative: 3077.0000 - accuracy: 0.8307 - precision: 0.8169 - recall: 0.8482 - auc: 0.9139 - prc: 0.9333 - val_loss: 0.3477 - val_true positive: 75.0000 - val_false positive: 2407.0000 - val_true negative: 43080.0000 - val_false negative: 7.0000 - val_accuracy: 0.9470 - val_precision: 0.0302 - val_recall: 0.9146 - val_auc: 0.9637 - val_prc: 0.7496\n",
      "Epoch 8/1000\n",
      "20/20 [==============================] - 0s 18ms/step - loss: 0.3471 - true positive: 17332.0000 - false positive: 3376.0000 - true negative: 17224.0000 - false negative: 3028.0000 - accuracy: 0.8437 - precision: 0.8370 - recall: 0.8513 - auc: 0.9241 - prc: 0.9427 - val_loss: 0.3105 - val_true positive: 76.0000 - val_false positive: 1638.0000 - val_true negative: 43849.0000 - val_false negative: 6.0000 - val_accuracy: 0.9639 - val_precision: 0.0443 - val_recall: 0.9268 - val_auc: 0.9686 - val_prc: 0.7636\n",
      "Epoch 9/1000\n",
      "20/20 [==============================] - 0s 18ms/step - loss: 0.3205 - true positive: 17501.0000 - false positive: 2818.0000 - true negative: 17708.0000 - false negative: 2933.0000 - accuracy: 0.8596 - precision: 0.8613 - recall: 0.8565 - auc: 0.9342 - prc: 0.9508 - val_loss: 0.2794 - val_true positive: 76.0000 - val_false positive: 1174.0000 - val_true negative: 44313.0000 - val_false negative: 6.0000 - val_accuracy: 0.9741 - val_precision: 0.0608 - val_recall: 0.9268 - val_auc: 0.9724 - val_prc: 0.7736\n",
      "Epoch 10/1000\n",
      " 5/20 [======>.......................] - ETA: 0s - loss: 0.3055 - true positive: 4383.0000 - false positive: 633.0000 - true negative: 4494.0000 - false negative: 730.0000 - accuracy: 0.8669 - precision: 0.8738 - recall: 0.8572 - auc: 0.9398 - prc: 0.9546"
     ]
    }
   ],
   "source": [
    "resampled_model = make_model()\n",
    "resampled_model.load_weights(initial_weights)\n",
    "\n",
    "# Reset the bias to zero, since this dataset is balanced.\n",
    "output_layer = resampled_model.layers[-1] \n",
    "output_layer.bias.assign([0])\n",
    "\n",
    "resampled_history = resampled_model.fit(\n",
    "    resampled_ds,\n",
    "    # These are not real epochs\n",
    "    steps_per_epoch=20,\n",
    "    epochs=10*EPOCHS,\n",
    "    callbacks=[early_stopping],\n",
    "    validation_data=(val_ds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'plot_metrics' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/4k/y4ljh2217c57vl68z1zkl0440000gn/T/ipykernel_71935/4170836928.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mplot_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresampled_history\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'plot_metrics' is not defined"
     ]
    }
   ],
   "source": [
    "plot_metrics(resampled_history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using XGBoost instead of a neural net\n",
    "\n",
    "We know that using a neural net might not be the best model for this use case. What about the using boosted decision tree instead? And also what about using **Synthetic Minority Over-sampling Technique (SMOTE)** for handling our imbalanced dataset? \n",
    "\n",
    "source: https://blog.dominodatalab.com/credit-card-fraud-detection-using-xgboost-smote-and-threshold-moving#:~:text=%20Credit%20Card%20Fraud%20Detection%20using%20XGBoost%2C%20SMOTE%2C,will%20now%20train%20an%20XGBoost%20classifier%2C...%20More%20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>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>V9</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>0.0</td>\n",
       "      <td>-1.359807</td>\n",
       "      <td>-0.072781</td>\n",
       "      <td>2.536347</td>\n",
       "      <td>1.378155</td>\n",
       "      <td>-0.338321</td>\n",
       "      <td>0.462388</td>\n",
       "      <td>0.239599</td>\n",
       "      <td>0.098698</td>\n",
       "      <td>0.363787</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018307</td>\n",
       "      <td>0.277838</td>\n",
       "      <td>-0.110474</td>\n",
       "      <td>0.066928</td>\n",
       "      <td>0.128539</td>\n",
       "      <td>-0.189115</td>\n",
       "      <td>0.133558</td>\n",
       "      <td>-0.021053</td>\n",
       "      <td>149.62</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.191857</td>\n",
       "      <td>0.266151</td>\n",
       "      <td>0.166480</td>\n",
       "      <td>0.448154</td>\n",
       "      <td>0.060018</td>\n",
       "      <td>-0.082361</td>\n",
       "      <td>-0.078803</td>\n",
       "      <td>0.085102</td>\n",
       "      <td>-0.255425</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.225775</td>\n",
       "      <td>-0.638672</td>\n",
       "      <td>0.101288</td>\n",
       "      <td>-0.339846</td>\n",
       "      <td>0.167170</td>\n",
       "      <td>0.125895</td>\n",
       "      <td>-0.008983</td>\n",
       "      <td>0.014724</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.358354</td>\n",
       "      <td>-1.340163</td>\n",
       "      <td>1.773209</td>\n",
       "      <td>0.379780</td>\n",
       "      <td>-0.503198</td>\n",
       "      <td>1.800499</td>\n",
       "      <td>0.791461</td>\n",
       "      <td>0.247676</td>\n",
       "      <td>-1.514654</td>\n",
       "      <td>...</td>\n",
       "      <td>0.247998</td>\n",
       "      <td>0.771679</td>\n",
       "      <td>0.909412</td>\n",
       "      <td>-0.689281</td>\n",
       "      <td>-0.327642</td>\n",
       "      <td>-0.139097</td>\n",
       "      <td>-0.055353</td>\n",
       "      <td>-0.059752</td>\n",
       "      <td>378.66</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.966272</td>\n",
       "      <td>-0.185226</td>\n",
       "      <td>1.792993</td>\n",
       "      <td>-0.863291</td>\n",
       "      <td>-0.010309</td>\n",
       "      <td>1.247203</td>\n",
       "      <td>0.237609</td>\n",
       "      <td>0.377436</td>\n",
       "      <td>-1.387024</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.108300</td>\n",
       "      <td>0.005274</td>\n",
       "      <td>-0.190321</td>\n",
       "      <td>-1.175575</td>\n",
       "      <td>0.647376</td>\n",
       "      <td>-0.221929</td>\n",
       "      <td>0.062723</td>\n",
       "      <td>0.061458</td>\n",
       "      <td>123.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.158233</td>\n",
       "      <td>0.877737</td>\n",
       "      <td>1.548718</td>\n",
       "      <td>0.403034</td>\n",
       "      <td>-0.407193</td>\n",
       "      <td>0.095921</td>\n",
       "      <td>0.592941</td>\n",
       "      <td>-0.270533</td>\n",
       "      <td>0.817739</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.009431</td>\n",
       "      <td>0.798278</td>\n",
       "      <td>-0.137458</td>\n",
       "      <td>0.141267</td>\n",
       "      <td>-0.206010</td>\n",
       "      <td>0.502292</td>\n",
       "      <td>0.219422</td>\n",
       "      <td>0.215153</td>\n",
       "      <td>69.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Time        V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0   0.0 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388  0.239599   \n",
       "1   0.0  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361 -0.078803   \n",
       "2   1.0 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
       "3   1.0 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
       "4   2.0 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921  0.592941   \n",
       "\n",
       "         V8        V9  ...       V21       V22       V23       V24       V25  \\\n",
       "0  0.098698  0.363787  ... -0.018307  0.277838 -0.110474  0.066928  0.128539   \n",
       "1  0.085102 -0.255425  ... -0.225775 -0.638672  0.101288 -0.339846  0.167170   \n",
       "2  0.247676 -1.514654  ...  0.247998  0.771679  0.909412 -0.689281 -0.327642   \n",
       "3  0.377436 -1.387024  ... -0.108300  0.005274 -0.190321 -1.175575  0.647376   \n",
       "4 -0.270533  0.817739  ... -0.009431  0.798278 -0.137458  0.141267 -0.206010   \n",
       "\n",
       "        V26       V27       V28  Amount  Class  \n",
       "0 -0.189115  0.133558 -0.021053  149.62      0  \n",
       "1  0.125895 -0.008983  0.014724    2.69      0  \n",
       "2 -0.139097 -0.055353 -0.059752  378.66      0  \n",
       "3 -0.221929  0.062723  0.061458  123.50      0  \n",
       "4  0.502292  0.219422  0.215153   69.99      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import random\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from imblearn.over_sampling import SMOTE # conda install imbalanced-learn\n",
    "from xgboost import XGBClassifier\n",
    "from xgboost import Booster\n",
    "from xgboost import DMatrix\n",
    "from sklearn import metrics\n",
    "from datetime import datetime\n",
    "dataDF = pd.read_csv(\"./creditcard.csv\")\n",
    "dataDF.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'Class distribution of the target variable'}>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "value_counts = dataDF[\"Class\"].value_counts()\n",
    "value_counts.plot(kind=\"bar\", title=\"Class distribution of the target variable\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>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>V9</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>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>...</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "      <td>284807.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>94813.860</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>88.350</td>\n",
       "      <td>0.002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>47488.146</td>\n",
       "      <td>1.959</td>\n",
       "      <td>1.651</td>\n",
       "      <td>1.516</td>\n",
       "      <td>1.416</td>\n",
       "      <td>1.380</td>\n",
       "      <td>1.332</td>\n",
       "      <td>1.237</td>\n",
       "      <td>1.194</td>\n",
       "      <td>1.099</td>\n",
       "      <td>...</td>\n",
       "      <td>0.735</td>\n",
       "      <td>0.726</td>\n",
       "      <td>0.624</td>\n",
       "      <td>0.606</td>\n",
       "      <td>0.521</td>\n",
       "      <td>0.482</td>\n",
       "      <td>0.404</td>\n",
       "      <td>0.330</td>\n",
       "      <td>250.120</td>\n",
       "      <td>0.042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000</td>\n",
       "      <td>-56.408</td>\n",
       "      <td>-72.716</td>\n",
       "      <td>-48.326</td>\n",
       "      <td>-5.683</td>\n",
       "      <td>-113.743</td>\n",
       "      <td>-26.161</td>\n",
       "      <td>-43.557</td>\n",
       "      <td>-73.217</td>\n",
       "      <td>-13.434</td>\n",
       "      <td>...</td>\n",
       "      <td>-34.830</td>\n",
       "      <td>-10.933</td>\n",
       "      <td>-44.808</td>\n",
       "      <td>-2.837</td>\n",
       "      <td>-10.295</td>\n",
       "      <td>-2.605</td>\n",
       "      <td>-22.566</td>\n",
       "      <td>-15.430</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>54201.500</td>\n",
       "      <td>-0.920</td>\n",
       "      <td>-0.599</td>\n",
       "      <td>-0.890</td>\n",
       "      <td>-0.849</td>\n",
       "      <td>-0.692</td>\n",
       "      <td>-0.768</td>\n",
       "      <td>-0.554</td>\n",
       "      <td>-0.209</td>\n",
       "      <td>-0.643</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.228</td>\n",
       "      <td>-0.542</td>\n",
       "      <td>-0.162</td>\n",
       "      <td>-0.355</td>\n",
       "      <td>-0.317</td>\n",
       "      <td>-0.327</td>\n",
       "      <td>-0.071</td>\n",
       "      <td>-0.053</td>\n",
       "      <td>5.600</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>84692.000</td>\n",
       "      <td>0.018</td>\n",
       "      <td>0.065</td>\n",
       "      <td>0.180</td>\n",
       "      <td>-0.020</td>\n",
       "      <td>-0.054</td>\n",
       "      <td>-0.274</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.022</td>\n",
       "      <td>-0.051</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.029</td>\n",
       "      <td>0.007</td>\n",
       "      <td>-0.011</td>\n",
       "      <td>0.041</td>\n",
       "      <td>0.017</td>\n",
       "      <td>-0.052</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.011</td>\n",
       "      <td>22.000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>139320.500</td>\n",
       "      <td>1.316</td>\n",
       "      <td>0.804</td>\n",
       "      <td>1.027</td>\n",
       "      <td>0.743</td>\n",
       "      <td>0.612</td>\n",
       "      <td>0.399</td>\n",
       "      <td>0.570</td>\n",
       "      <td>0.327</td>\n",
       "      <td>0.597</td>\n",
       "      <td>...</td>\n",
       "      <td>0.186</td>\n",
       "      <td>0.529</td>\n",
       "      <td>0.148</td>\n",
       "      <td>0.440</td>\n",
       "      <td>0.351</td>\n",
       "      <td>0.241</td>\n",
       "      <td>0.091</td>\n",
       "      <td>0.078</td>\n",
       "      <td>77.165</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>172792.000</td>\n",
       "      <td>2.455</td>\n",
       "      <td>22.058</td>\n",
       "      <td>9.383</td>\n",
       "      <td>16.875</td>\n",
       "      <td>34.802</td>\n",
       "      <td>73.302</td>\n",
       "      <td>120.589</td>\n",
       "      <td>20.007</td>\n",
       "      <td>15.595</td>\n",
       "      <td>...</td>\n",
       "      <td>27.203</td>\n",
       "      <td>10.503</td>\n",
       "      <td>22.528</td>\n",
       "      <td>4.585</td>\n",
       "      <td>7.520</td>\n",
       "      <td>3.517</td>\n",
       "      <td>31.612</td>\n",
       "      <td>33.848</td>\n",
       "      <td>25691.160</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Time         V1         V2         V3         V4         V5  \\\n",
       "count 284807.000 284807.000 284807.000 284807.000 284807.000 284807.000   \n",
       "mean   94813.860      0.000      0.000     -0.000      0.000     -0.000   \n",
       "std    47488.146      1.959      1.651      1.516      1.416      1.380   \n",
       "min        0.000    -56.408    -72.716    -48.326     -5.683   -113.743   \n",
       "25%    54201.500     -0.920     -0.599     -0.890     -0.849     -0.692   \n",
       "50%    84692.000      0.018      0.065      0.180     -0.020     -0.054   \n",
       "75%   139320.500      1.316      0.804      1.027      0.743      0.612   \n",
       "max   172792.000      2.455     22.058      9.383     16.875     34.802   \n",
       "\n",
       "              V6         V7         V8         V9  ...        V21        V22  \\\n",
       "count 284807.000 284807.000 284807.000 284807.000  ... 284807.000 284807.000   \n",
       "mean       0.000     -0.000     -0.000     -0.000  ...      0.000      0.000   \n",
       "std        1.332      1.237      1.194      1.099  ...      0.735      0.726   \n",
       "min      -26.161    -43.557    -73.217    -13.434  ...    -34.830    -10.933   \n",
       "25%       -0.768     -0.554     -0.209     -0.643  ...     -0.228     -0.542   \n",
       "50%       -0.274      0.040      0.022     -0.051  ...     -0.029      0.007   \n",
       "75%        0.399      0.570      0.327      0.597  ...      0.186      0.529   \n",
       "max       73.302    120.589     20.007     15.595  ...     27.203     10.503   \n",
       "\n",
       "             V23        V24        V25        V26        V27        V28  \\\n",
       "count 284807.000 284807.000 284807.000 284807.000 284807.000 284807.000   \n",
       "mean       0.000      0.000      0.000      0.000     -0.000     -0.000   \n",
       "std        0.624      0.606      0.521      0.482      0.404      0.330   \n",
       "min      -44.808     -2.837    -10.295     -2.605    -22.566    -15.430   \n",
       "25%       -0.162     -0.355     -0.317     -0.327     -0.071     -0.053   \n",
       "50%       -0.011      0.041      0.017     -0.052      0.001      0.011   \n",
       "75%        0.148      0.440      0.351      0.241      0.091      0.078   \n",
       "max       22.528      4.585      7.520      3.517     31.612     33.848   \n",
       "\n",
       "          Amount      Class  \n",
       "count 284807.000 284807.000  \n",
       "mean      88.350      0.002  \n",
       "std      250.120      0.042  \n",
       "min        0.000      0.000  \n",
       "25%        5.600      0.000  \n",
       "50%       22.000      0.000  \n",
       "75%       77.165      0.000  \n",
       "max    25691.160      1.000  \n",
       "\n",
       "[8 rows x 31 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.set_option(\"display.float_format\", lambda x: \"%.3f\" % x)\n",
    "dataDF.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x864 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = dataDF.drop(\"Class\", axis=1).hist(figsize=(10,12),bins=100)\n",
    "# We hide the axes' labels to make the plot neater and more compact\n",
    "for axis in ax.flatten():\n",
    "    axis.set_xticklabels([])\n",
    "    axis.set_yticklabels([])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataDF[\"Hour\"] = dataDF[\"Time\"].apply(datetime.fromtimestamp).dt.hour"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Engineering\n",
    "\n",
    "We noticed during the exploratory analysis that the Amount column is not zero mean centered. Let's fix this, and also center the Hour attribute, which we'll be using instead of Time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert timestamp to hours \n",
    "dataDF[\"Hour\"] = dataDF[\"Time\"].apply(datetime.fromtimestamp).dt.hour"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fraudulent transactions are 0.17% of the training set.\n",
      "Fraudulent transactions are 0.17% of the test set.\n"
     ]
    }
   ],
   "source": [
    "trainDF, testDF = train_test_split(dataDF, test_size=0.2, random_state=1234, stratify=dataDF[[\"Class\"]])\n",
    "tr_value_counts = trainDF[\"Class\"].value_counts()\n",
    "print(\"Fraudulent transactions are %.2f%% of the training set.\" % (tr_value_counts[1] * 100 / len(trainDF)))\n",
    "tst_value_counts = testDF[\"Class\"].value_counts()\n",
    "print(\"Fraudulent transactions are %.2f%% of the test set.\" % (tst_value_counts[1] * 100 / len(testDF)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We noticed during the exploratory analysis that the Amount column is not zero mean centered. Let's fix this, and also center the Hour attribute, which we'll be using instead of Time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "trainDF_norm = trainDF.copy()\n",
    "trainDF_norm[\"Amount\"] = trainDF[\"Amount\"].subtract(trainDF[\"Amount\"].mean())\n",
    "trainDF_norm[\"Hour\"] = trainDF[\"Hour\"].subtract(trainDF[\"Hour\"].mean())\n",
    "testDF_norm = testDF.copy()\n",
    "testDF_norm[\"Amount\"] = testDF[\"Amount\"].subtract(testDF[\"Amount\"].mean())\n",
    "testDF_norm[\"Hour\"] = testDF[\"Hour\"].subtract(testDF[\"Hour\"].mean())\n",
    "trainDF = trainDF_norm\n",
    "testDF = testDF_norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop because using \"Hour\" instead\n",
    "trainDF = trainDF.drop([\"Time\"], axis=1)\n",
    "testDF = testDF.drop([\"Time\"], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>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>V9</th>\n",
       "      <th>V10</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>Hour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>135152</th>\n",
       "      <td>1.082</td>\n",
       "      <td>-0.075</td>\n",
       "      <td>1.395</td>\n",
       "      <td>1.373</td>\n",
       "      <td>-0.958</td>\n",
       "      <td>0.077</td>\n",
       "      <td>-0.625</td>\n",
       "      <td>0.205</td>\n",
       "      <td>0.861</td>\n",
       "      <td>-0.248</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.012</td>\n",
       "      <td>0.204</td>\n",
       "      <td>0.063</td>\n",
       "      <td>0.399</td>\n",
       "      <td>0.298</td>\n",
       "      <td>-0.400</td>\n",
       "      <td>0.089</td>\n",
       "      <td>0.039</td>\n",
       "      <td>-76.750</td>\n",
       "      <td>5.578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103706</th>\n",
       "      <td>1.013</td>\n",
       "      <td>0.188</td>\n",
       "      <td>1.615</td>\n",
       "      <td>2.594</td>\n",
       "      <td>-0.609</td>\n",
       "      <td>0.839</td>\n",
       "      <td>-0.683</td>\n",
       "      <td>0.400</td>\n",
       "      <td>-0.260</td>\n",
       "      <td>0.597</td>\n",
       "      <td>...</td>\n",
       "      <td>0.017</td>\n",
       "      <td>0.218</td>\n",
       "      <td>0.074</td>\n",
       "      <td>0.230</td>\n",
       "      <td>0.185</td>\n",
       "      <td>-0.046</td>\n",
       "      <td>0.059</td>\n",
       "      <td>0.027</td>\n",
       "      <td>-81.450</td>\n",
       "      <td>2.578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>231651</th>\n",
       "      <td>-0.701</td>\n",
       "      <td>0.090</td>\n",
       "      <td>1.540</td>\n",
       "      <td>-3.114</td>\n",
       "      <td>0.458</td>\n",
       "      <td>0.431</td>\n",
       "      <td>0.183</td>\n",
       "      <td>0.195</td>\n",
       "      <td>-1.230</td>\n",
       "      <td>-0.503</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.069</td>\n",
       "      <td>-0.367</td>\n",
       "      <td>-0.449</td>\n",
       "      <td>-1.461</td>\n",
       "      <td>0.628</td>\n",
       "      <td>-0.475</td>\n",
       "      <td>0.004</td>\n",
       "      <td>0.027</td>\n",
       "      <td>-76.250</td>\n",
       "      <td>-0.422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199939</th>\n",
       "      <td>-0.430</td>\n",
       "      <td>-0.595</td>\n",
       "      <td>0.676</td>\n",
       "      <td>-2.603</td>\n",
       "      <td>1.499</td>\n",
       "      <td>4.231</td>\n",
       "      <td>-1.079</td>\n",
       "      <td>1.239</td>\n",
       "      <td>-0.798</td>\n",
       "      <td>-0.003</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.015</td>\n",
       "      <td>0.214</td>\n",
       "      <td>0.090</td>\n",
       "      <td>0.694</td>\n",
       "      <td>-0.698</td>\n",
       "      <td>0.592</td>\n",
       "      <td>0.158</td>\n",
       "      <td>0.166</td>\n",
       "      <td>-42.350</td>\n",
       "      <td>-3.422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103404</th>\n",
       "      <td>1.296</td>\n",
       "      <td>1.011</td>\n",
       "      <td>-3.192</td>\n",
       "      <td>0.472</td>\n",
       "      <td>3.350</td>\n",
       "      <td>2.433</td>\n",
       "      <td>0.190</td>\n",
       "      <td>0.622</td>\n",
       "      <td>-0.557</td>\n",
       "      <td>-1.487</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.263</td>\n",
       "      <td>-0.825</td>\n",
       "      <td>-0.245</td>\n",
       "      <td>0.675</td>\n",
       "      <td>1.011</td>\n",
       "      <td>-0.279</td>\n",
       "      <td>0.041</td>\n",
       "      <td>0.091</td>\n",
       "      <td>-83.260</td>\n",
       "      <td>2.578</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           V1     V2     V3     V4     V5    V6     V7    V8     V9    V10  \\\n",
       "135152  1.082 -0.075  1.395  1.373 -0.958 0.077 -0.625 0.205  0.861 -0.248   \n",
       "103706  1.013  0.188  1.615  2.594 -0.609 0.839 -0.683 0.400 -0.260  0.597   \n",
       "231651 -0.701  0.090  1.540 -3.114  0.458 0.431  0.183 0.195 -1.230 -0.503   \n",
       "199939 -0.430 -0.595  0.676 -2.603  1.499 4.231 -1.079 1.239 -0.798 -0.003   \n",
       "103404  1.296  1.011 -3.192  0.472  3.350 2.433  0.190 0.622 -0.557 -1.487   \n",
       "\n",
       "        ...    V21    V22    V23    V24    V25    V26   V27   V28  Amount  \\\n",
       "135152  ... -0.012  0.204  0.063  0.399  0.298 -0.400 0.089 0.039 -76.750   \n",
       "103706  ...  0.017  0.218  0.074  0.230  0.185 -0.046 0.059 0.027 -81.450   \n",
       "231651  ... -0.069 -0.367 -0.449 -1.461  0.628 -0.475 0.004 0.027 -76.250   \n",
       "199939  ... -0.015  0.214  0.090  0.694 -0.698  0.592 0.158 0.166 -42.350   \n",
       "103404  ... -0.263 -0.825 -0.245  0.675  1.011 -0.279 0.041 0.091 -83.260   \n",
       "\n",
       "         Hour  \n",
       "135152  5.578  \n",
       "103706  2.578  \n",
       "231651 -0.422  \n",
       "199939 -3.422  \n",
       "103404  2.578  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = trainDF.iloc[:, trainDF.columns != \"Class\"]\n",
    "y_train = trainDF.iloc[:, trainDF.columns == \"Class\"]\n",
    "X_test = testDF.iloc[:, testDF.columns != \"Class\"]\n",
    "y_test = testDF.iloc[:, testDF.columns == \"Class\"]\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fraudulent transactions are 50.00% of the test set.\n"
     ]
    }
   ],
   "source": [
    "X_train_smote, y_train_smote = SMOTE(random_state=1234).fit_resample(X_train, y_train)\n",
    "smote_value_counts = y_train_smote[\"Class\"].value_counts()\n",
    "print(\"Fraudulent transactions are %.2f%% of the test set.\" % (smote_value_counts[0] * 100 / len(y_train_smote)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# creating our xgboost search with stratified k folds cross validation \n",
    "\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.StratifiedKFold.html\n",
    "\n",
    "This cross-validation object is a variation of KFold that returns stratified folds.\n",
    "The folds are made by preserving the percentage of samples for each class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def xgboost_search(X, y, search_verbose=1):\n",
    "    params = {\n",
    "        \"gamma\":[0.5, 1, 1.5, 2, 5],\n",
    "        \"max_depth\":[3,4,5,6],\n",
    "        \"min_child_weight\": [100],\n",
    "        \"subsample\": [0.6, 0.8, 1.0],\n",
    "        \"colsample_bytree\": [0.6, 0.8, 1.0],\n",
    "        \"learning_rate\": [0.1, 0.01, 0.001]\n",
    "    }\n",
    "    xgb = XGBClassifier(objective=\"binary:logistic\", eval_metric=\"auc\", use_label_encoder=False)\n",
    "    skf = StratifiedKFold(n_splits=3, shuffle=True, random_state=1234)\n",
    "    grid_search = GridSearchCV(estimator=xgb, param_grid=params, scoring=\"roc_auc\", n_jobs=1, cv=skf.split(X,y), verbose=search_verbose)\n",
    "    grid_search.fit(X, y)\n",
    "    print(\"Best estimator: \")\n",
    "    print(grid_search.best_estimator_)\n",
    "    print(\"Parameters: \", grid_search.best_params_)\n",
    "    print(\"Highest AUC: %.2f\" % grid_search.best_score_)\n",
    "    return grid_search.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 540 candidates, totalling 1620 fits\n",
      "Best estimator: \n",
      "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
      "              colsample_bynode=1, colsample_bytree=0.6,\n",
      "              enable_categorical=False, eval_metric='auc', gamma=2, gpu_id=-1,\n",
      "              importance_type=None, interaction_constraints='',\n",
      "              learning_rate=0.1, max_delta_step=0, max_depth=3,\n",
      "              min_child_weight=100, missing=nan, monotone_constraints='()',\n",
      "              n_estimators=100, n_jobs=8, num_parallel_tree=1, predictor='auto',\n",
      "              random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=1,\n",
      "              subsample=1.0, tree_method='exact', use_label_encoder=False,\n",
      "              validate_parameters=1, verbosity=None)\n",
      "Parameters:  {'colsample_bytree': 0.6, 'gamma': 2, 'learning_rate': 0.1, 'max_depth': 3, 'min_child_weight': 100, 'subsample': 1.0}\n",
      "Highest AUC: 0.98\n"
     ]
    }
   ],
   "source": [
    "rows = random.sample(np.arange(0,len(X_train_smote.index)).tolist(), 5000)\n",
    "model_params = xgboost_search(X_train_smote.iloc[rows,], y_train_smote.iloc[rows,])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=0.6,\n",
       "              enable_categorical=False, eval_metric='auc', gamma=2, gpu_id=-1,\n",
       "              importance_type=None, interaction_constraints='',\n",
       "              learning_rate=0.1, max_delta_step=0, max_depth=3,\n",
       "              min_child_weight=100, missing=nan, monotone_constraints='()',\n",
       "              n_estimators=100, n_jobs=8, num_parallel_tree=1, predictor='auto',\n",
       "              random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=1,\n",
       "              subsample=1.0, tree_method='exact', use_label_encoder=False,\n",
       "              validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = XGBClassifier(objective=\"binary:logistic\", eval_metric=\"auc\", use_label_encoder=False)\n",
    "model.set_params(**model_params)\n",
    "model.fit(X_train_smote, y_train_smote)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Evalutation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict_proba(X_test)[:,1]\n",
    "fp_r, tp_r, t = metrics.roc_curve(y_test, y_pred)\n",
    "auc = metrics.auc(fp_r, tp_r)\n",
    "plt.figure(figsize=(8, 6))\n",
    "plt.plot(fp_r, tp_r, label=\"AUC = %.2f\" % auc)\n",
    "plt.plot([0,1],[0,1],\"r--\")\n",
    "plt.ylabel(\"TP rate\")\n",
    "plt.xlabel(\"FP rate\")\n",
    "plt.legend(loc=4)\n",
    "plt.title(\"ROC Curve\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Threshold value is: 0.83\n"
     ]
    }
   ],
   "source": [
    "t_opt_idx = np.argmax(tp_r - fp_r)\n",
    "t_opt = t[t_opt_idx]\n",
    "print(\"Threshold value is: %.2f\" % t_opt)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Selecting the optimal threshold value can be performed in a number of ways. Looking at the ROC curve, we can intuitively see that the best performance (misclassification costs aside) would be yield by the threshold that puts us in the top left section of the curve (i.e. TP rate is high, FP rate is low). With this criterion in mind, we can define a distance metric to the top left corner of the curve and find a threshold that minimises it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0.98, 'Impact of threshold adjustment on the error matrix')"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict_proba(X_test)[:,1]\n",
    "fig, axes = plt.subplots(3,3, figsize=(10,10))\n",
    "for t, ax in enumerate(axes.flat):\n",
    "    threshold = (t+1)/10\n",
    "    y_pred_int = (y_pred > threshold).astype(int)\n",
    "    c_matrix = metrics.confusion_matrix(y_test, y_pred_int)\n",
    "    sns.heatmap(c_matrix, annot=True, cmap=\"Blues\", fmt=\"d\", ax=ax, cbar=False)\n",
    "    ax.title.set_text(\"T=%.1f\" % threshold)\n",
    "plt.subplots_adjust(hspace=0.5, wspace=0.5)\n",
    "plt.suptitle(\"Impact of threshold adjustment on the error matrix\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save_model(\"smote_fraud.xgb\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comparing false positive rates -- because that's what we care about since it's fraud detection \n",
    "\n",
    "* neural net - 25 false positives \n",
    "* weighted neural net - 21 false positives \n",
    "* xgboost with SMOTE - 7 false positives \n",
    "\n",
    "XGBoost + KstratifiedFolds Cross validation + SMOTE + grid search is 3 times better than a neural net! "
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "b7e818f66e33c31ac0526ee7f8556503ff93918b8b22809241939dc19e90de0b"
  },
  "kernelspec": {
   "display_name": "Python 3.8.12 64-bit ('pytorch_m1': conda)",
   "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": "3.8.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}