5_Credit_Card__Fraud_Detection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eQj-t94GKDf7"
   },
   "source": [
    "##  Credit Card Fraud Detection (Unsupervised Anomaly Detection)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data\n",
    "Source: https://dacon.io/competitions/official/235930/data\n",
    "#### Train Dataset\n",
    "* Unlabeled card data\n",
    "* Columns:\n",
    "    * `id`: card transaction ID\n",
    "    * `V1`, `V2`, ... ,`V30` : card features\n",
    "\n",
    "#### Validation Dataset\n",
    "* Labeled card data\n",
    "* Used for performance measure only\n",
    "* Columns:\n",
    "    * `id`: card transaction ID\n",
    "    * `V1`, `V2`, ... ,`V30`: card features\n",
    "    * `class`: class indicating transcation type (0: normal, 1: fraud)\n",
    "\n",
    "#### Test Dataset\n",
    "* Unlabeled card data\n",
    "* Used for final score measurement\n",
    "* Columns:\n",
    "    * `id`: card transaction ID\n",
    "    * `V1`, `V2`, ... ,`V30` : card features\n",
    "    \n",
    "### Score Measures\n",
    "* Evaluated with macro f1-score\n",
    "* Public score: measured with 30% of the test data\n",
    "* Private score: measured with the remaining 70% of the test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "DG4kIYY-KJEt"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import tensorflow as tf\n",
    "import pyod\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import font_manager, rc\n",
    "\n",
    "from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "from sklearn.preprocessing import Normalizer, MinMaxScaler\n",
    "from sklearn.pipeline import Pipeline\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.neighbors import LocalOutlierFactor\n",
    "from sklearn.metrics import f1_score, classification_report, confusion_matrix, precision_recall_fscore_support\n",
    "from sklearn.ensemble import IsolationForest\n",
    "from sklearn.covariance import EllipticEnvelope\n",
    "from pyod.models.auto_encoder import AutoEncoder\n",
    "import time\n",
    "\n",
    "pd.set_option('mode.chained_assignment',  None) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Loading & EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "LInSs9c4KNWX"
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('train.csv')\n",
    "test = pd.read_csv('test.csv')\n",
    "val = pd.read_csv('val.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 300
    },
    "id": "iVGrjDkQKRlO",
    "outputId": "9dc77cf6-8388-4104-d1be-193df071f999"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-4b448e5b-42cd-4f5c-ae7c-9f33ee1edb5c\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>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>V29</th>\n",
       "      <th>V30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3</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>4.983721</td>\n",
       "      <td>-0.994972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4</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>1.418291</td>\n",
       "      <td>-0.994972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>-0.425966</td>\n",
       "      <td>0.960523</td>\n",
       "      <td>1.141109</td>\n",
       "      <td>-0.168252</td>\n",
       "      <td>0.420987</td>\n",
       "      <td>-0.029728</td>\n",
       "      <td>0.476201</td>\n",
       "      <td>0.260314</td>\n",
       "      <td>-0.568671</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.208254</td>\n",
       "      <td>-0.559825</td>\n",
       "      <td>-0.026398</td>\n",
       "      <td>-0.371427</td>\n",
       "      <td>-0.232794</td>\n",
       "      <td>0.105915</td>\n",
       "      <td>0.253844</td>\n",
       "      <td>0.081080</td>\n",
       "      <td>-0.256131</td>\n",
       "      <td>-0.994960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>-0.644269</td>\n",
       "      <td>1.417964</td>\n",
       "      <td>1.074380</td>\n",
       "      <td>-0.492199</td>\n",
       "      <td>0.948934</td>\n",
       "      <td>0.428118</td>\n",
       "      <td>1.120631</td>\n",
       "      <td>-3.807864</td>\n",
       "      <td>0.615375</td>\n",
       "      <td>...</td>\n",
       "      <td>1.943465</td>\n",
       "      <td>-1.015455</td>\n",
       "      <td>0.057504</td>\n",
       "      <td>-0.649709</td>\n",
       "      <td>-0.415267</td>\n",
       "      <td>-0.051634</td>\n",
       "      <td>-1.206921</td>\n",
       "      <td>-1.085339</td>\n",
       "      <td>0.262698</td>\n",
       "      <td>-0.994901</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9</td>\n",
       "      <td>-0.894286</td>\n",
       "      <td>0.286157</td>\n",
       "      <td>-0.113192</td>\n",
       "      <td>-0.271526</td>\n",
       "      <td>2.669599</td>\n",
       "      <td>3.721818</td>\n",
       "      <td>0.370145</td>\n",
       "      <td>0.851084</td>\n",
       "      <td>-0.392048</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.073425</td>\n",
       "      <td>-0.268092</td>\n",
       "      <td>-0.204233</td>\n",
       "      <td>1.011592</td>\n",
       "      <td>0.373205</td>\n",
       "      <td>-0.384157</td>\n",
       "      <td>0.011747</td>\n",
       "      <td>0.142404</td>\n",
       "      <td>0.994900</td>\n",
       "      <td>-0.994901</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4b448e5b-42cd-4f5c-ae7c-9f33ee1edb5c')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-4b448e5b-42cd-4f5c-ae7c-9f33ee1edb5c button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-4b448e5b-42cd-4f5c-ae7c-9f33ee1edb5c');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "   ID        V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0   3 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
       "1   4 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
       "2   6 -0.425966  0.960523  1.141109 -0.168252  0.420987 -0.029728  0.476201   \n",
       "3   8 -0.644269  1.417964  1.074380 -0.492199  0.948934  0.428118  1.120631   \n",
       "4   9 -0.894286  0.286157 -0.113192 -0.271526  2.669599  3.721818  0.370145   \n",
       "\n",
       "         V8        V9  ...       V21       V22       V23       V24       V25  \\\n",
       "0  0.247676 -1.514654  ...  0.247998  0.771679  0.909412 -0.689281 -0.327642   \n",
       "1  0.377436 -1.387024  ... -0.108300  0.005274 -0.190321 -1.175575  0.647376   \n",
       "2  0.260314 -0.568671  ... -0.208254 -0.559825 -0.026398 -0.371427 -0.232794   \n",
       "3 -3.807864  0.615375  ...  1.943465 -1.015455  0.057504 -0.649709 -0.415267   \n",
       "4  0.851084 -0.392048  ... -0.073425 -0.268092 -0.204233  1.011592  0.373205   \n",
       "\n",
       "        V26       V27       V28       V29       V30  \n",
       "0 -0.139097 -0.055353 -0.059752  4.983721 -0.994972  \n",
       "1 -0.221929  0.062723  0.061458  1.418291 -0.994972  \n",
       "2  0.105915  0.253844  0.081080 -0.256131 -0.994960  \n",
       "3 -0.051634 -1.206921 -1.085339  0.262698 -0.994901  \n",
       "4 -0.384157  0.011747  0.142404  0.994900 -0.994901  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 300
    },
    "id": "ccvVJgqdYKeY",
    "outputId": "37d1d2ec-7d5b-4684-f76f-90a58ecd99b0"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-998589c6-cad6-4180-a5e5-6ae5ec85d640\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>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>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>V29</th>\n",
       "      <th>V30</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>-0.338262</td>\n",
       "      <td>1.119593</td>\n",
       "      <td>1.044367</td>\n",
       "      <td>-0.222187</td>\n",
       "      <td>0.499361</td>\n",
       "      <td>-0.246761</td>\n",
       "      <td>0.651583</td>\n",
       "      <td>0.069539</td>\n",
       "      <td>-0.736727</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.633753</td>\n",
       "      <td>-0.120794</td>\n",
       "      <td>-0.385050</td>\n",
       "      <td>-0.069733</td>\n",
       "      <td>0.094199</td>\n",
       "      <td>0.246219</td>\n",
       "      <td>0.083076</td>\n",
       "      <td>-0.255991</td>\n",
       "      <td>-0.994878</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>22</td>\n",
       "      <td>0.962496</td>\n",
       "      <td>0.328461</td>\n",
       "      <td>-0.171479</td>\n",
       "      <td>2.109204</td>\n",
       "      <td>1.129566</td>\n",
       "      <td>1.696038</td>\n",
       "      <td>0.107712</td>\n",
       "      <td>0.521502</td>\n",
       "      <td>-1.191311</td>\n",
       "      <td>...</td>\n",
       "      <td>0.402492</td>\n",
       "      <td>-0.048508</td>\n",
       "      <td>-1.371866</td>\n",
       "      <td>0.390814</td>\n",
       "      <td>0.199964</td>\n",
       "      <td>0.016371</td>\n",
       "      <td>-0.014605</td>\n",
       "      <td>0.168937</td>\n",
       "      <td>-0.994784</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>63</td>\n",
       "      <td>1.145524</td>\n",
       "      <td>0.575068</td>\n",
       "      <td>0.194008</td>\n",
       "      <td>2.598192</td>\n",
       "      <td>-0.092210</td>\n",
       "      <td>-1.044430</td>\n",
       "      <td>0.531588</td>\n",
       "      <td>-0.241888</td>\n",
       "      <td>-0.896287</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.119703</td>\n",
       "      <td>-0.076510</td>\n",
       "      <td>0.691320</td>\n",
       "      <td>0.633984</td>\n",
       "      <td>0.048741</td>\n",
       "      <td>-0.053192</td>\n",
       "      <td>0.016251</td>\n",
       "      <td>0.169496</td>\n",
       "      <td>-0.994502</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>69</td>\n",
       "      <td>0.927060</td>\n",
       "      <td>-0.323684</td>\n",
       "      <td>0.387585</td>\n",
       "      <td>0.544474</td>\n",
       "      <td>0.246787</td>\n",
       "      <td>1.650358</td>\n",
       "      <td>-0.427576</td>\n",
       "      <td>0.615371</td>\n",
       "      <td>0.226278</td>\n",
       "      <td>...</td>\n",
       "      <td>0.079359</td>\n",
       "      <td>0.096632</td>\n",
       "      <td>-0.992569</td>\n",
       "      <td>0.085096</td>\n",
       "      <td>0.377447</td>\n",
       "      <td>0.036096</td>\n",
       "      <td>-0.005960</td>\n",
       "      <td>0.331307</td>\n",
       "      <td>-0.994467</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>83</td>\n",
       "      <td>-3.005237</td>\n",
       "      <td>2.600138</td>\n",
       "      <td>1.483691</td>\n",
       "      <td>-2.418473</td>\n",
       "      <td>0.306326</td>\n",
       "      <td>-0.824575</td>\n",
       "      <td>2.065426</td>\n",
       "      <td>-1.829347</td>\n",
       "      <td>4.009259</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.181268</td>\n",
       "      <td>-0.163747</td>\n",
       "      <td>0.515821</td>\n",
       "      <td>0.136318</td>\n",
       "      <td>0.460054</td>\n",
       "      <td>-0.251259</td>\n",
       "      <td>-1.105751</td>\n",
       "      <td>-0.287012</td>\n",
       "      <td>-0.994373</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 32 columns</p>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-998589c6-cad6-4180-a5e5-6ae5ec85d640')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-998589c6-cad6-4180-a5e5-6ae5ec85d640 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-998589c6-cad6-4180-a5e5-6ae5ec85d640');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "   ID        V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0  10 -0.338262  1.119593  1.044367 -0.222187  0.499361 -0.246761  0.651583   \n",
       "1  22  0.962496  0.328461 -0.171479  2.109204  1.129566  1.696038  0.107712   \n",
       "2  63  1.145524  0.575068  0.194008  2.598192 -0.092210 -1.044430  0.531588   \n",
       "3  69  0.927060 -0.323684  0.387585  0.544474  0.246787  1.650358 -0.427576   \n",
       "4  83 -3.005237  2.600138  1.483691 -2.418473  0.306326 -0.824575  2.065426   \n",
       "\n",
       "         V8        V9  ...       V22       V23       V24       V25       V26  \\\n",
       "0  0.069539 -0.736727  ... -0.633753 -0.120794 -0.385050 -0.069733  0.094199   \n",
       "1  0.521502 -1.191311  ...  0.402492 -0.048508 -1.371866  0.390814  0.199964   \n",
       "2 -0.241888 -0.896287  ... -0.119703 -0.076510  0.691320  0.633984  0.048741   \n",
       "3  0.615371  0.226278  ...  0.079359  0.096632 -0.992569  0.085096  0.377447   \n",
       "4 -1.829347  4.009259  ... -0.181268 -0.163747  0.515821  0.136318  0.460054   \n",
       "\n",
       "        V27       V28       V29       V30  Class  \n",
       "0  0.246219  0.083076 -0.255991 -0.994878      0  \n",
       "1  0.016371 -0.014605  0.168937 -0.994784      0  \n",
       "2 -0.053192  0.016251  0.169496 -0.994502      0  \n",
       "3  0.036096 -0.005960  0.331307 -0.994467      0  \n",
       "4 -0.251259 -1.105751 -0.287012 -0.994373      0  \n",
       "\n",
       "[5 rows x 32 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 300
    },
    "id": "pFMjxQRQKSXJ",
    "outputId": "7f236b63-362d-4f0b-eff0-7ea09fc299c8"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-ef5f9c17-de70-4873-8731-2adef3fd92e7\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>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>V29</th>\n",
       "      <th>V30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AAAA0x1</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>1.783274</td>\n",
       "      <td>-0.994983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AAAA0x2</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>-0.269825</td>\n",
       "      <td>-0.994983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AAAA0x5</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>0.670579</td>\n",
       "      <td>-0.994960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AAAA0x7</td>\n",
       "      <td>1.229658</td>\n",
       "      <td>0.141004</td>\n",
       "      <td>0.045371</td>\n",
       "      <td>1.202613</td>\n",
       "      <td>0.191881</td>\n",
       "      <td>0.272708</td>\n",
       "      <td>-0.005159</td>\n",
       "      <td>0.081213</td>\n",
       "      <td>0.464960</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.167716</td>\n",
       "      <td>-0.270710</td>\n",
       "      <td>-0.154104</td>\n",
       "      <td>-0.780055</td>\n",
       "      <td>0.750137</td>\n",
       "      <td>-0.257237</td>\n",
       "      <td>0.034507</td>\n",
       "      <td>0.005168</td>\n",
       "      <td>-0.237686</td>\n",
       "      <td>-0.994937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AAAA0xc</td>\n",
       "      <td>0.384978</td>\n",
       "      <td>0.616109</td>\n",
       "      <td>-0.874300</td>\n",
       "      <td>-0.094019</td>\n",
       "      <td>2.924584</td>\n",
       "      <td>3.317027</td>\n",
       "      <td>0.470455</td>\n",
       "      <td>0.538247</td>\n",
       "      <td>-0.558895</td>\n",
       "      <td>...</td>\n",
       "      <td>0.049924</td>\n",
       "      <td>0.238422</td>\n",
       "      <td>0.009130</td>\n",
       "      <td>0.996710</td>\n",
       "      <td>-0.767315</td>\n",
       "      <td>-0.492208</td>\n",
       "      <td>0.042472</td>\n",
       "      <td>-0.054337</td>\n",
       "      <td>-0.167819</td>\n",
       "      <td>-0.994866</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ef5f9c17-de70-4873-8731-2adef3fd92e7')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-ef5f9c17-de70-4873-8731-2adef3fd92e7 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-ef5f9c17-de70-4873-8731-2adef3fd92e7');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "        ID        V1        V2        V3        V4        V5        V6  \\\n",
       "0  AAAA0x1 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388   \n",
       "1  AAAA0x2  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361   \n",
       "2  AAAA0x5 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921   \n",
       "3  AAAA0x7  1.229658  0.141004  0.045371  1.202613  0.191881  0.272708   \n",
       "4  AAAA0xc  0.384978  0.616109 -0.874300 -0.094019  2.924584  3.317027   \n",
       "\n",
       "         V7        V8        V9  ...       V21       V22       V23       V24  \\\n",
       "0  0.239599  0.098698  0.363787  ... -0.018307  0.277838 -0.110474  0.066928   \n",
       "1 -0.078803  0.085102 -0.255425  ... -0.225775 -0.638672  0.101288 -0.339846   \n",
       "2  0.592941 -0.270533  0.817739  ... -0.009431  0.798278 -0.137458  0.141267   \n",
       "3 -0.005159  0.081213  0.464960  ... -0.167716 -0.270710 -0.154104 -0.780055   \n",
       "4  0.470455  0.538247 -0.558895  ...  0.049924  0.238422  0.009130  0.996710   \n",
       "\n",
       "        V25       V26       V27       V28       V29       V30  \n",
       "0  0.128539 -0.189115  0.133558 -0.021053  1.783274 -0.994983  \n",
       "1  0.167170  0.125895 -0.008983  0.014724 -0.269825 -0.994983  \n",
       "2 -0.206010  0.502292  0.219422  0.215153  0.670579 -0.994960  \n",
       "3  0.750137 -0.257237  0.034507  0.005168 -0.237686 -0.994937  \n",
       "4 -0.767315 -0.492208  0.042472 -0.054337 -0.167819 -0.994866  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "UtGkqygDKURE",
    "outputId": "a184ec8e-6411-4858-ca77-f4759e57e7e8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 113842 entries, 0 to 113841\n",
      "Data columns (total 31 columns):\n",
      " #   Column  Non-Null Count   Dtype  \n",
      "---  ------  --------------   -----  \n",
      " 0   ID      113842 non-null  int64  \n",
      " 1   V1      113842 non-null  float64\n",
      " 2   V2      113842 non-null  float64\n",
      " 3   V3      113842 non-null  float64\n",
      " 4   V4      113842 non-null  float64\n",
      " 5   V5      113842 non-null  float64\n",
      " 6   V6      113842 non-null  float64\n",
      " 7   V7      113842 non-null  float64\n",
      " 8   V8      113842 non-null  float64\n",
      " 9   V9      113842 non-null  float64\n",
      " 10  V10     113842 non-null  float64\n",
      " 11  V11     113842 non-null  float64\n",
      " 12  V12     113842 non-null  float64\n",
      " 13  V13     113842 non-null  float64\n",
      " 14  V14     113842 non-null  float64\n",
      " 15  V15     113842 non-null  float64\n",
      " 16  V16     113842 non-null  float64\n",
      " 17  V17     113842 non-null  float64\n",
      " 18  V18     113842 non-null  float64\n",
      " 19  V19     113842 non-null  float64\n",
      " 20  V20     113842 non-null  float64\n",
      " 21  V21     113842 non-null  float64\n",
      " 22  V22     113842 non-null  float64\n",
      " 23  V23     113842 non-null  float64\n",
      " 24  V24     113842 non-null  float64\n",
      " 25  V25     113842 non-null  float64\n",
      " 26  V26     113842 non-null  float64\n",
      " 27  V27     113842 non-null  float64\n",
      " 28  V28     113842 non-null  float64\n",
      " 29  V29     113842 non-null  float64\n",
      " 30  V30     113842 non-null  float64\n",
      "dtypes: float64(30), int64(1)\n",
      "memory usage: 26.9 MB\n"
     ]
    }
   ],
   "source": [
    "train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "QESG30-DZftq",
    "outputId": "72a66db5-60b0-439c-c2ee-519280fd1421"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ID     0\n",
       "V1     0\n",
       "V2     0\n",
       "V3     0\n",
       "V4     0\n",
       "V5     0\n",
       "V6     0\n",
       "V7     0\n",
       "V8     0\n",
       "V9     0\n",
       "V10    0\n",
       "V11    0\n",
       "V12    0\n",
       "V13    0\n",
       "V14    0\n",
       "V15    0\n",
       "V16    0\n",
       "V17    0\n",
       "V18    0\n",
       "V19    0\n",
       "V20    0\n",
       "V21    0\n",
       "V22    0\n",
       "V23    0\n",
       "V24    0\n",
       "V25    0\n",
       "V26    0\n",
       "V27    0\n",
       "V28    0\n",
       "V29    0\n",
       "V30    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "znwYGmZ3OIZn"
   },
   "outputs": [],
   "source": [
    "train_df = train.drop(['ID'], axis=1)\n",
    "test_df = test.drop(['ID'], axis=1)\n",
    "val_df = val.drop(['ID', 'Class'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 394
    },
    "id": "1iOq43aRBzC5",
    "outputId": "f9e8a761-303e-499d-d353-ec49ee4ecf33"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-286a490a-f189-41d5-aaf9-322c11461014\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>V29</th>\n",
       "      <th>V30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>1.138420e+05</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "      <td>113842.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.000197</td>\n",
       "      <td>0.001289</td>\n",
       "      <td>0.009717</td>\n",
       "      <td>-0.004169</td>\n",
       "      <td>0.000475</td>\n",
       "      <td>0.005141</td>\n",
       "      <td>0.005769</td>\n",
       "      <td>-0.002451</td>\n",
       "      <td>-0.002107</td>\n",
       "      <td>0.005209</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.001242</td>\n",
       "      <td>4.088347e-07</td>\n",
       "      <td>-0.001317</td>\n",
       "      <td>-0.000884</td>\n",
       "      <td>0.001680</td>\n",
       "      <td>-0.000293</td>\n",
       "      <td>-0.000234</td>\n",
       "      <td>-0.000508</td>\n",
       "      <td>0.927253</td>\n",
       "      <td>0.116232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.951060</td>\n",
       "      <td>1.651064</td>\n",
       "      <td>1.496916</td>\n",
       "      <td>1.412633</td>\n",
       "      <td>1.367533</td>\n",
       "      <td>1.330583</td>\n",
       "      <td>1.204111</td>\n",
       "      <td>1.185504</td>\n",
       "      <td>1.095415</td>\n",
       "      <td>1.071337</td>\n",
       "      <td>...</td>\n",
       "      <td>0.722001</td>\n",
       "      <td>7.238291e-01</td>\n",
       "      <td>0.636061</td>\n",
       "      <td>0.605854</td>\n",
       "      <td>0.520069</td>\n",
       "      <td>0.480979</td>\n",
       "      <td>0.399505</td>\n",
       "      <td>0.356130</td>\n",
       "      <td>3.412933</td>\n",
       "      <td>0.558161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-56.407510</td>\n",
       "      <td>-72.715728</td>\n",
       "      <td>-32.454198</td>\n",
       "      <td>-5.600607</td>\n",
       "      <td>-42.147898</td>\n",
       "      <td>-26.160506</td>\n",
       "      <td>-41.506796</td>\n",
       "      <td>-50.943369</td>\n",
       "      <td>-13.434066</td>\n",
       "      <td>-24.403185</td>\n",
       "      <td>...</td>\n",
       "      <td>-22.757540</td>\n",
       "      <td>-8.887017e+00</td>\n",
       "      <td>-44.807735</td>\n",
       "      <td>-2.824849</td>\n",
       "      <td>-10.295397</td>\n",
       "      <td>-1.855355</td>\n",
       "      <td>-9.895244</td>\n",
       "      <td>-9.617915</td>\n",
       "      <td>-0.307413</td>\n",
       "      <td>-0.994972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-0.923479</td>\n",
       "      <td>-0.595602</td>\n",
       "      <td>-0.883877</td>\n",
       "      <td>-0.853728</td>\n",
       "      <td>-0.689853</td>\n",
       "      <td>-0.766094</td>\n",
       "      <td>-0.552071</td>\n",
       "      <td>-0.209492</td>\n",
       "      <td>-0.647477</td>\n",
       "      <td>-0.533477</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.229710</td>\n",
       "      <td>-5.402665e-01</td>\n",
       "      <td>-0.162180</td>\n",
       "      <td>-0.355582</td>\n",
       "      <td>-0.315470</td>\n",
       "      <td>-0.326160</td>\n",
       "      <td>-0.070847</td>\n",
       "      <td>-0.053249</td>\n",
       "      <td>-0.230560</td>\n",
       "      <td>-0.360304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.012074</td>\n",
       "      <td>0.066390</td>\n",
       "      <td>0.183868</td>\n",
       "      <td>-0.019359</td>\n",
       "      <td>-0.054060</td>\n",
       "      <td>-0.272436</td>\n",
       "      <td>0.039036</td>\n",
       "      <td>0.020970</td>\n",
       "      <td>-0.052157</td>\n",
       "      <td>-0.090810</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.030281</td>\n",
       "      <td>8.345807e-03</td>\n",
       "      <td>-0.012261</td>\n",
       "      <td>0.040573</td>\n",
       "      <td>0.018278</td>\n",
       "      <td>-0.052815</td>\n",
       "      <td>0.001502</td>\n",
       "      <td>0.011158</td>\n",
       "      <td>-0.000699</td>\n",
       "      <td>-0.002590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.315373</td>\n",
       "      <td>0.801687</td>\n",
       "      <td>1.037120</td>\n",
       "      <td>0.742208</td>\n",
       "      <td>0.614214</td>\n",
       "      <td>0.405285</td>\n",
       "      <td>0.568750</td>\n",
       "      <td>0.328303</td>\n",
       "      <td>0.590705</td>\n",
       "      <td>0.455287</td>\n",
       "      <td>...</td>\n",
       "      <td>0.186001</td>\n",
       "      <td>5.287508e-01</td>\n",
       "      <td>0.147474</td>\n",
       "      <td>0.438225</td>\n",
       "      <td>0.353989</td>\n",
       "      <td>0.240838</td>\n",
       "      <td>0.091279</td>\n",
       "      <td>0.077851</td>\n",
       "      <td>0.768532</td>\n",
       "      <td>0.640653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2.454930</td>\n",
       "      <td>21.467203</td>\n",
       "      <td>4.187811</td>\n",
       "      <td>16.491217</td>\n",
       "      <td>34.801666</td>\n",
       "      <td>23.917837</td>\n",
       "      <td>44.054461</td>\n",
       "      <td>20.007208</td>\n",
       "      <td>10.392889</td>\n",
       "      <td>15.331742</td>\n",
       "      <td>...</td>\n",
       "      <td>27.202839</td>\n",
       "      <td>8.361985e+00</td>\n",
       "      <td>22.528412</td>\n",
       "      <td>4.022866</td>\n",
       "      <td>7.519589</td>\n",
       "      <td>3.119295</td>\n",
       "      <td>11.135740</td>\n",
       "      <td>33.847808</td>\n",
       "      <td>180.101027</td>\n",
       "      <td>1.034951</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 30 columns</p>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-286a490a-f189-41d5-aaf9-322c11461014')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-286a490a-f189-41d5-aaf9-322c11461014 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-286a490a-f189-41d5-aaf9-322c11461014');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                  V1             V2             V3             V4  \\\n",
       "count  113842.000000  113842.000000  113842.000000  113842.000000   \n",
       "mean        0.000197       0.001289       0.009717      -0.004169   \n",
       "std         1.951060       1.651064       1.496916       1.412633   \n",
       "min       -56.407510     -72.715728     -32.454198      -5.600607   \n",
       "25%        -0.923479      -0.595602      -0.883877      -0.853728   \n",
       "50%         0.012074       0.066390       0.183868      -0.019359   \n",
       "75%         1.315373       0.801687       1.037120       0.742208   \n",
       "max         2.454930      21.467203       4.187811      16.491217   \n",
       "\n",
       "                  V5             V6             V7             V8  \\\n",
       "count  113842.000000  113842.000000  113842.000000  113842.000000   \n",
       "mean        0.000475       0.005141       0.005769      -0.002451   \n",
       "std         1.367533       1.330583       1.204111       1.185504   \n",
       "min       -42.147898     -26.160506     -41.506796     -50.943369   \n",
       "25%        -0.689853      -0.766094      -0.552071      -0.209492   \n",
       "50%        -0.054060      -0.272436       0.039036       0.020970   \n",
       "75%         0.614214       0.405285       0.568750       0.328303   \n",
       "max        34.801666      23.917837      44.054461      20.007208   \n",
       "\n",
       "                  V9            V10  ...            V21           V22  \\\n",
       "count  113842.000000  113842.000000  ...  113842.000000  1.138420e+05   \n",
       "mean       -0.002107       0.005209  ...      -0.001242  4.088347e-07   \n",
       "std         1.095415       1.071337  ...       0.722001  7.238291e-01   \n",
       "min       -13.434066     -24.403185  ...     -22.757540 -8.887017e+00   \n",
       "25%        -0.647477      -0.533477  ...      -0.229710 -5.402665e-01   \n",
       "50%        -0.052157      -0.090810  ...      -0.030281  8.345807e-03   \n",
       "75%         0.590705       0.455287  ...       0.186001  5.287508e-01   \n",
       "max        10.392889      15.331742  ...      27.202839  8.361985e+00   \n",
       "\n",
       "                 V23            V24            V25            V26  \\\n",
       "count  113842.000000  113842.000000  113842.000000  113842.000000   \n",
       "mean       -0.001317      -0.000884       0.001680      -0.000293   \n",
       "std         0.636061       0.605854       0.520069       0.480979   \n",
       "min       -44.807735      -2.824849     -10.295397      -1.855355   \n",
       "25%        -0.162180      -0.355582      -0.315470      -0.326160   \n",
       "50%        -0.012261       0.040573       0.018278      -0.052815   \n",
       "75%         0.147474       0.438225       0.353989       0.240838   \n",
       "max        22.528412       4.022866       7.519589       3.119295   \n",
       "\n",
       "                 V27            V28            V29            V30  \n",
       "count  113842.000000  113842.000000  113842.000000  113842.000000  \n",
       "mean       -0.000234      -0.000508       0.927253       0.116232  \n",
       "std         0.399505       0.356130       3.412933       0.558161  \n",
       "min        -9.895244      -9.617915      -0.307413      -0.994972  \n",
       "25%        -0.070847      -0.053249      -0.230560      -0.360304  \n",
       "50%         0.001502       0.011158      -0.000699      -0.002590  \n",
       "75%         0.091279       0.077851       0.768532       0.640653  \n",
       "max        11.135740      33.847808     180.101027       1.034951  \n",
       "\n",
       "[8 rows x 30 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "u7NpnLwdcHYM",
    "outputId": "778cd88f-83bd-4e3f-c301-f613c8f1ce7b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "정상거래: 28432, 사기거래: 30\n",
      "비율: 0.001055\n"
     ]
    }
   ],
   "source": [
    "val_norm, val_fraud = val['Class'].value_counts()\n",
    "val_contamination = val_fraud/val_norm\n",
    "print(\"정상거래: %d, 사기거래: %d\" %(val_norm, val_fraud))\n",
    "print(\"비율: %5f\" %(val_contamination))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Since the true labels in the training data are unknown, we simply assume that it contains the same proportion of fraud transactions as the validation data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 668
    },
    "id": "R-WKcOLH-5cw",
    "outputId": "4d0d98e7-e0eb-4ef2-8052-8b055865ed48"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3kAAANeCAYAAACvWcdbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdf5xdVX3v/9e7CUEuivyepJn8oJLijXhFyIOkX9o+JlIhUG6Dj4dFqCVRqbn9ChSvpE1ie794BS3pHaUElTaV1FCVQP1Fbm4kzUXmWh9tEIJIBORmjMTMNAE0CETUMPTz/WOvk+xMzpk5kzk/9jnn/Xw85jHnrL3Pmc8+Z2Vnr70+ay1FBGZmZmZmZtYefqXZAZiZmZmZmVntuJFnZmZmZmbWRtzIMzMzMzMzayNu5JmZmZmZmbURN/LMzMzMzMzaiBt5ZmZmZmZmbcSNPDMzMzMzszbiRl4LkHSfpI+WKV8oaY+kt0t6QNILkp5uQohmFVVRf/9U0vckvSTph5L+tBlxmg1XZd3dIelFSf8m6RZJE5sRq9lwVdTfien5JElPShpofJRmh6ui7t4k6RVJ+3I/v9aMWIvMjbzWsBb4Q0kaVn4l8AXgBWAN4ItjK6LR6q+ARcAJwALgGkmXNzZEs7JGq7tfAc6OiOOAM4G3AH/S2BDNKhqx/kbEUHr+p8BzDY3MbGSjnXuHgLsj4rW5nx0Nj7Lg3MhrDV8DTgJ+q1Qg6QTgEuDOiPh2RPwD4ApuRTRa/f2riHgkIoYi4ingXuC85oRqdojR6u4PIuKnpU3AvwOnNzxKs/JGrL/p+WnAHwJ/2YwAzSoYte7a6NzIawER8XPgHrLejpLLgO9HxHebE5VZdcZSf9Ndu98CHm9chGblVVN3Jf2BpBeBH5P15P1twwM1K6PKc+9twIeBnzc4PLOKqqy7/1nSXkmPS/p/Gx5kC3Ajr3WsBd4p6TXp+aJUZtYKqq2/HyE7L/19g+IyG82IdTcivpjSNX8d+BvgmcaHaFZRxfor6R3AhIj4arOCMxvBSOfee4D/CJwCvB/4/yRd0fgQi82NvBYREd8iu1N8qaQ3AOcCX2xuVGbVqab+SrqG7CT+uxHxy8ZHaXa4as+9EbGdrAf6M42N0KyySvVX0rHAX+ExpFZQI517I+KJiPi3iHg1Iv4FuBV4Z/OiLSbPAtZa7iS7CD4D2BQRvmNsraRi/ZX0PmA58NsR4RnerGiqPfdOBN7QsKjMqnNY/ZV0FjAT+Oc0t8Uk4PWS9gDzIuLpJsVqllftuTfIxkVbjnvyWsudwO+QdU0fSBeS9CupO/uo7KleI2lSk2I0q6RS/X038HHg7Z4dywqqUt39I0mnpsezgRXA/U2J0KyycvX3e8A04Kz080dkqcZnAbuaEKNZOZXOvQslnaDMuWQ90vc2KcbCUkQ0OwYbA0l9ZIP7J5dS2iT1AA8M2/X/RERPQ4MzG0WF+vtDoBvIp2h+PiL+uPERmpVXoe7+PXAx8FqyKej/EfhvEfGLZsVpVk65+jtsew/Zebe7waGZjajCufcu4ALgaGAA+ExErGpakAXlRp6ZmZmZmVkbcbqmmZmZmZlZG3Ejz8zMzMzMrI24kWdmZmZmZtZG3MgzMzMzMzNrIy27Tt7JJ58cM2fObHYYFf3sZz/j2GOPbXYYo2qFOMca49atW38cEafUMaRxKVrdbYU6UG9F+QyKXnehePV3LIryPRdBPT6LotffItbddq+TrXJ8Ra+7UIz6W/Tvs1Pjq1R/W7aRN3PmTB5++OFmh1FRX18fPT09zQ5jVK0Q51hjlLSzftGMX9HqbivUgXorymdQ9LoLxau/Y1GU77kI6vFZFL3+FrHutnudbJXjK3rdhWLU36J/n50aX6X663RNMzMzMzOzNuJGnpmZmZmZWRtxI8/MzMzMzKyNuJFnZmZmZmbWRtzIs8Ka0j0dSWzdupUp3dObHY61Idcx61SPPbYNSa73BVA6D/n7KA5Jx0v6kqTvS3pS0m9IOlHSZknb0+8T0r6StEpSv6THJJ2de5/Faf/tkhbnys+RtC29ZpUkNeM428mU7un+v3wYN/KssPYM7mLGsg1Mmnw6ewZ3NTsca0OuY9apXnllPzOWbXC9L4DSecjfR6HcCtwXEW8E3gI8CSwH7o+IWcD96TnARcCs9LMEuB1A0onADcBc4FzghlLDMO3z/tzrFjTgmNransFd/r98GDfyzMzMzMwASa8Hfhu4AyAi9kfET4GFwNq021rg0vR4IXBnZLYAx0uaAlwIbI6IvRHxPLAZWJC2HRcRWyIigDtz72VWM27kmZmZdaoJRzlV0OxQpwHPAX8v6TuSPivpWKArInanffYAXenxVCDffTSQykYqHyhTblZTLbsYupmZmY3Tq68wY9kGAHauvKTJwZgVwkTgbODaiHhQ0q0cTM0EICJCUtQ7EElLyFJA6erqoq+vr95/ckT79u1regyV9Pb20nVM9ruoMTb683Mjz8zMzMwsMwAMRMSD6fmXyBp5z0iaEhG7U8rls2n7IDAt9/ruVDYI9Awr70vl3WX2P0xErAZWA8yZMyd6enrK7dYwfX19NDuGSubPn8+qz3+NpUuXkmXBFk+jPz+na5qZmZmZARGxB9gl6YxUdD7wBLAeKM2QuRi4Nz1eDyxKs2zOA15IaZ2bgAsknZAmXLkA2JS2vShpXppVc1Huvcxqxj15ZmZmZmYHXQt8QdIkYAfwXrKOkXskXQXsBC5L+24ELgb6gZfTvkTEXkk3Ag+l/T4aEXvT4w8AnwOOAb6efsxqyo08MzMzM7MkIh4F5pTZdH6ZfQO4usL7rAHWlCl/GDhznGGajcjpmtbWvKCpmZmZmXWaURt5kl4j6duSvivpcUn/PZWfJunBdHF7d+rSRtLR6Xl/2j4z914rUvlTki7MlS9IZf2Slg+PwWwcvKCpmZmZmXWUanryfgm8LSLeApxFtpDjPGAlcEtEnA48D1yV9r8KeD6V35L2Q9Js4HLgTWQXwp+RNEHSBODTZBfYs4Er0r5m4+IFTc3MrEimdE/3uoRm1hCjjslLF6/70tOj0k8AbwP+IJWvBT5C1quxMD2GbNrZT6UUtoXAuoj4JfBDSf1kvSIA/RGxA0DSurTvE+M5MDMOXdD0LcBW4DqasKBp0da6ySvyujf11tvby6TJQ4VfW8fM2sOewV1el9DMGqKqiVdSb9tW4HSyXrcfAD+NiKG0S/7i9sAFcUQMSXoBOCmVb8m9bf41wy+g5475SKwjTOmezp7BXUyeOo3dAz8abffCLGhatLVu8oq87k29zZ8/nxnLNnD9m4cKt7aOpGlkvcNdZDfWVkfErSl9+G5gJvA0cFlEPJ9upt1KNsvby8B7IuKR9F6Lgb9Ib31TRKxN5edwcIa3jcB1UaQPwVqWpDXAJcCzEXFmKvsIWXr7c2m3D0fExrRtBVkm0KvAn0TEplS+gKxeTwA+GxE3p/LTgHVk1xdbgSsjYn9jjs7MrPiqauRFxKvAWZKOB74KvLGuUVVQ5N6Q4Vqld6TIcZbrZVn6X69j0uTT2b+nv5q4C7OgqdkRGAKuj4hHJL0O2CppM/AesjGlN6cxzMuBZRw6pnQuWWbF3NyY0jlkjcWtktan1OPSmNIHyRp5C/BU3lYbnwM+RXajIu+WiOjNFwwbzvGrwP+W9Otp86eBt5Odzx9KdfcJDg4ZWSfpb8gaiLfX62DMzFrNmJZQiIifSnoA+A2y8UoTU29e/uK2dKE8IGki8HrgJ1S+gGaE8uF/v7C9IcO1Su9IkeMs18tSKtu5cvRel4jYI2mXpDMi4ikOLmj6BNlCpjdz+IKm16SU4bmkBU0lbQI+npts5QJgRVoD58U0RvVBsgVNb6vtp2CdKqUU706PX5L0JFn2w0IO3nRYS3bDYRm5MaXAljSz7JS07+bS+kypobhAUh9pTGkqL40pdSPPxi0ivpmfeG0UYxrOkf4tVBoyYmZmVNHIk3QK8Epq4B1DdkdtJfAA8E6ydInhF8qLgX9N27+RUuLWA1+U9EmyO3WzgG8DAmal1ItBsrt5pRO32Xh5QVNreeli+a1kNxM8pvQIFTlzodG6u7u5fvIQ+1PGBMD+xoxLvUbSIuBhsp7q5xn7cI6TqDxk5BBFq7u9wz7vffv2HVbW7Bhryf/mzJqnmp68KcDaNC7vV4B7ImKDpCeAdZJuAr5DmsEw/f6HdCduL1mjjYh4XNI9ZL0oQ8DVKQ0USdcAm8hy7tdExOM1O0LraF7Q1FqdpNcCXwY+GBEv5pdi9JjSsSly5kKjfeITn+C2597IzpVLcxOB1H1c6u3AjWRpwzcCnwDeV88/WLS6W8pGgezzfuCBB1i6tKHfQUP535xZ81Qzu+ZjZHeQh5fv4GA6Rb78F8DvV3ivjwEfK1O+kawXxczMEklHkTXwvhARX0nFHlNqLSkinik9lvR3wIb0dKzDOX5C5SEjZmZGdevkmZlZg6XZMu8AnoyIT+Y2lVLi4fBU+UXKzCONKSXLkrhA0glpXOkFwKa07UVJ89LfWpR7L7OaSzclSt4BfC89Xg9cLunoNHSjNJzjIdJwjpRyfzmwPmVdlIaMwKH/DszMjDFOvGJmZg1zHnAlsE3So6nsw2QTBnlMqRWapLvIepBPljRANsNrj6SzyNI1nwb+CxzxcI5llB8y0jomHMXWrVubHYWZtSk38szMCigivkU2MVU5HlNqhRYRV5QprtgQG+twjkpDRlrKq68wafLpzY7CzNqU0zXNzMzMzMzaiBt5ZmZmZmZmbcSNPDMzMzMzszbiRp6ZmVkHmNI9HUnk11o0M7P25IlXzMzMOsCewV0HFt2G7zc1FjMzqy/35JmZmZmZmbURN/LMzMzMzMzaiBt5ZmZmZmZmbcSNPGsNE47yZAFmZmZmZlVwI89aw6uv5CYMMDMzMzOzStzIMzMzMzMzayNu5JmZmZmZmbURN/LMzMzM6qi0EL21DkkTJH1H0ob0/DRJD0rql3S3pEmp/Oj0vD9tn5l7jxWp/ClJF+bKF6SyfknLG31s1hncyDMzMzOro0MXorcWcR3wZO75SuCWiDgdeB64KpVfBTyfym9J+yFpNnA58CZgAfCZ1HCcAHwauAiYDVyR9jWrKTfyzMzMzMwSSd3A7wKfTc8FvA34UtplLXBperwwPSdtPz/tvxBYFxG/jIgfAv3AuemnPyJ2RMR+YF3a16ymJjY7ADMzMzPjwHJBk6dOY/fAj5odTSf7a+DPgNel5ycBP42IofR8AJiaHk8FdgFExJCkF9L+U4EtuffMv2bXsPK55YKQtARYAtDV1UVfX9+RH1EN7Nu3r+kxVNLb20vXMdnvosbY6M/PjTwzMzNzA6MI0nJBO1de0uxIOpakS4BnI2KrpJ5mxhIRq4HVAHPmzImenqaGQ19fH82OoZL58+ez6vNfY+nSpUREs8Mpq9Gfn9M1re158LSZWRVSA2PP4K7R9zVrX+cBvyfpabJUyrcBtwLHSyp1jnQDg+nxIDANIG1/PfCTfPmw11QqN6upURt5kqZJekDSE5Iel3RdKj9R0mZJ29PvE1K5JK1KF72PSTo7916L0/7bJS3OlZ8jaVt6zSp5CiqrLQ+eNjMzs1FFxIqI6I6ImWT/938jIt4NPAC8M+22GLg3PV6fnpO2fyOyrqT1wOXpBvJpwCzg28BDwKx0w3lS+hvrG3Bo1mGq6ckbAq6PiNnAPODqdCG7HLg/ImYB96fnkF3wzko/S4DbIWsUAjeQ5R2fC9xQahimfd6fe92C8R+amQdPm5mZWU0sAz4kqZ9szN0dqfwO4KRU/iHS9XBEPA7cAzwB3AdcHRGvpnF91wCbyG5A35P2NaupURt5EbE7Ih5Jj18iq5BTOfSCePiF8p2R2ULWvT0FuBDYHBF7I+J5YDOwIG07LiK2pDsfd+bey2y8SoOn/z09r3rwNJAfPD18kPTUEcrNzDqapDWSnpX0vVyZM4CspUREX0Rckh7viIhzI+L0iPj9iPhlKv9Fen562r4j9/qPRcQbIuKMiPh6rnxjRPx62vaxxh+ZdYIxTbySxii9FXgQ6IqI3WnTHqArPR7rBfHU9Hh4ebm/X6hZhkZS5BmI8oocZ29vL5MmDx2YLWnS5CH2536PFneRBk8Xue4WuQ7U2/A61qmfg1kdfA74FNmN25JSBtDNaQzzcrLekXwG0Fyy7J65uQygOUAAWyWtTzeKSxlADwIbyTKAvo6ZmQFjaORJei3wZeCDEfFi/qZZRISkuk9lU7RZhkZS5BmI8ooc5/z585mxbAPXv3mIpUuXphnHDv6uYvak0uDpi4HXAMeRGzydeuvKDZ4eqHLwNCOUH6LIdbfIdaDehtexos7IZTYeU7qnN3wylYj4Zn7yqmQh0JMerwX6yBp5BzKAgC2SShlAPaQMIABJpQygPlIGUCovZQC1TyPPM52a2ThV1ciTdBRZA+8LEfGVVPyMpCkRsTudjJ9N5SPNJtQzrLwvlXeX2d9sXCJiBbACIPXkLY2Id0v6R7LB0esoP3j6X8kNnpa0HviipE8Cv8rBwdMiDZ4mq7OXA3/QoMMzM6vKnsFdRZmWv2MzgIZnowDsz63rlS+bNHmI/Sv/kkmTT2f/nv6WzjDo5EwRs2YbtZGX8tzvAJ6MiE/mNpUuiG/m8AvlayStI0u7eCE1BDcBH89NtnIBsCIi9kp6UdI8srSLRcBtNTg2s0qWAesk3QR8h0MHT/9DGjy9l6zRRkQ8Lqk0eHqINHgaQFJp8PQEYI0HT5uZja7TMoBKWQOlbBSAnSuXHljXK192BFkrhdXJmSJmzVZNT955wJXANkmPprIPkzXu7pF0FbATuCxt2whcTDYD4cvAewFSY+5GsqljAT5aSsEAPkCWv38MWbpF+6RcWCFERB9ZzzFpUPS5Zfb5BfD7FV7/MeCwwdERsZGszpuZ2cicAWRm1iCjNvIi4ltkaWnlnF9m/wCurvBea4A1ZcofBs4cLRYzMzNrWc4AMjNrkDHNrmlmZmY2Gkl3kfXCnSxpgGyWTGcAmZk1iBt5ZmZmVlMRcUWFTc4AMjNrgFEXQzczMzMzM7PW4UaemZmZmZlZG3Ejz8zMzMzMrI24kWdmZmZWY1O6pyOJbLlhM7PGciPPzMzMrMb2DO5ixrINBxY6N7PaK91MscO5kWdmVkCS1kh6VtL3cmUnStosaXv6fUIql6RVkvolPSbp7NxrFqf9t0tanCs/R9K29JpV8v+SZmbWYko3U+xwbuRZ4fiujBmQrQG2YFjZcuD+iJgF3J+eA1wEzEo/S4DbIWsUkq1PNhc4F7ght7D07cD7c68b/rfMzMysRbmRZ4XjuzJmEBHfBPYOK14IrE2P1wKX5srvjMwW4HhJU4ALgc0RsTcingc2AwvStuMiYktao+zO3HuZmZlZi3Mjz8ysdXRFxO70eA/QlR5PBXbl9htIZSOVD5QpNzMzszYwsdkBmJnZ2EVESIpG/C1JS8jSQOnq6qKvr68Rf7bm9u3b17Kxj0dvby+TJg+xP/0GOHViN9cPK9uf268TPyczs3biRp6ZGcCEo5DE5KnT2D3wo2ZHU8kzkqZExO6UcvlsKh8EpuX2605lg0DPsPK+VN5dZv+yImI1sBpgzpw50dPTU2nXQuvr66NVYx+P+fPnM2PZBnauXHogFf7aUwa47bk3HlJWerxz5VKyLF4zM2tVTtc0MwN49RVmLNvAnsFdo+/bPOuB0gyZi4F7c+WL0iyb84AXUlrnJuACSSekCVcuADalbS9Kmpdm1VyUey8zMzNrce7JMzMrIEl3kfXCnSxpgGyWzJuBeyRdBewELku7bwQuBvqBl4H3AkTEXkk3Ag+l/T4aEaXJXD5ANoPnMcDX04+ZmZm1ATfyzMwKKCKuqLDp/DL7BnB1hfdZA6wpU/4wcOZ4YjQzM7NicrqmmZmZmZlZG3Ejz8zMzMzMrI24kWdmZmZmBkiaJukBSU9IelzSdan8REmbJW1Pv09I5ZK0SlK/pMcknZ17r8Vp/+2SFufKz5G0Lb1mVZoAy6ym3MgzMzNrM1O6pyMJXzuajdkQcH1EzAbmAVdLmg0sB+6PiFnA/ek5wEXArPSzBLgdskYh2YRZc4FzgRtKDcO0z/tzr1vQgOOyDuNGnpmZWZvZM7iLGcs2HFgDz8yqExG7I+KR9Pgl4ElgKrAQWJt2Wwtcmh4vBO6MzBbg+LSO6YXA5ojYGxHPA5uBBWnbcRGxJU2adWfuvcxqZtTZNSWtAS4Bno2IM1PZicDdwEzgaeCyiHg+dTffSjaV98vAe0r/UFI39V+kt70pItam8nM4OI33RuC68CqsVgOSppGdPLuAAFZHxK2uv2ZmZjYaSTOBtwIPAl1pjVGAPWTXFpA1APMLrA6kspHKB8qUl/v7S8h6B+nq6qKvr++Ij6UW9u3b1/QYhuvt7WXS5CH29/bSdUz2vGgxljT686tmCYXPAZ8iu1guKXVZ3yxpeXq+jEO7rOeSdUfPzXVZzyG72N4qaX26s1Hqsn6Q7CJ5AV6vyWqjlHLxiKTXkdW7zcB7cP01M2sKSU8DLwGvAkMRMaeWN9/MakHSa4EvAx+MiBfzqc8REZLqfkM3IlYDqwHmzJkTPT099f6TI+rr66PZMQw3f/58ZizbwM6VS1n1+a+xdOlSinqvvdGf36jpmhHxTWDvsGJ3WVvhOeXCzKyw5kfEWRExJz2v5XinpiqNh7TWJekosgbeFyLiK6n4mfT/Pun3s6l8EJiWe3l3KhupvLtMuVlNHeli6A3vsobidVuPpIhd2uUUMc581/ukyUMHut/zZfvH2B3vlIvKilgHGqW3Qh3r1M/DrIkWAj3p8VqgjyzD4sDNN2CLpNLNtx7SzTeAlKWxALirsWEfrjQecufKS5odih2B1Ht8B/BkRHwyt2k9sBi4Of2+N1d+jaR1ZDcdXoiI3ZI2AR/P3Xy4AFgREXslvShpHtk1ySLgtrofmHWcI23kHdCoLuv0twrVbT2SInZpl1PEOPNd7zOWbeD6Nw+xdOnSQ8p2rqy+O94pFyMrYh1olFJdG17HiprqYdYmAvindO7923R+rNXNt0M04wbb8JuSwCE3KPNl+RtM5fZr9ZtOLXoT8TzgSmCbpEdT2YfJGnf3SLoK2AlclrZtJEsn7idLKX4vQGrM3Qg8lPb7aOmmBPABDo7n/zoe5mF1cKSNvGckTUl3Kqrtsu4ZVt6Hu6ytzkZKuXD9NTNrit+MiEFJpwKbJX0/v7GWN9+acYNt+I1K4JAblPmy0hiiSvu1+k2nVryJGBHfAirl255fZv8Arq7wXmuANWXKHwbOHEeYZqM60iUUSl3WcHiX9aK0MOQ8Upc1sAm4QNIJqdv6AmBT2vaipHmpe3xR7r3MxqWKlAtw/TUza6iIGEy/nwW+SjamrlbjnczMjCoaeZLuAv4VOEPSQOqmvhl4u6TtwO+k55B1We8g67L+O7LuaFL3dKnL+iEO77L+bHrND3CXtdVOKeXibZIeTT8X4/prZtYUko5Nsx0j6Viym2bfo0Y33xp4KGZWRBOOQhKSmNI9vdnRNNWo6ZoRcUWFTe6ytkJzyoWZWeF0AV9NY6MnAl+MiPskPUTtxjuZWad69ZVc+nNnT3407olXzMzMrI2kO+EAk6dOY/fAj2r21hGxA3hLmfKfUKObb2Zm5kaemZmZ5flOuJlZyzvSiVfMzMzMzMysgNzIMzMzMysiTyJhZkfI6ZpmZmZmReTUWTM7Qu7JMzMzMzMzayNu5JlZx5nSPf3A7IFmZuNVOqf4vGJmReFGnpl1nD2Duw6kQJmZjVfpnOLzipkVhRt5ZmZmZmZmbcSNPDMzMzMzszbiRp6ZmVmb8HhTMzMDN/KsIDxo3cxs/Dze1MzMwI08KwgPWjczMzMzqw038szMzMyOQDPSY/OZL1O6pzf0b5tZ63Ajz8zMzOwINCM9Np/5smdwV0P/tpm1DjfyzMzMzMzM2ogbeWZmZmZFN+Gow1NDU5nTNs1sODfyzMzMzIru1VcOTw1NZU7bNLPh3MgzMzMzaxOliVncu2fW2dzIs6bywr1mZuNT13VGnQ54mEKu65q+J0kHJmZx7551vA4/f7mRZ03lhXvNzManruuMOh3wMIVc1zV9T4fElGv4depFrnW4Dj9/FaaRJ2mBpKck9Uta3ux4rH4KeRd0HFx320yHXRi5/lqrct0dRa7h16kXuUXm+nvk2u06sl4K0ciTNAH4NHARMBu4QtLs5kZl9VLIu6BHyHW3tVSVHtxBF0auv62tk9Pdm1F3W/rz7vC0taLxuXd82uk6sp4K0cgDzgX6I2JHROwH1gELmxyT1UD+bsvEo49p3f8gK3PdLbh8HRxzenD7Xxi5/raIcufSDk93b1jdLX32Lf15d3jaWgH53HsEWvpGSxMoIpodA5LeCSyIiD9Kz68E5kbENcP2WwIsSU/PAJ5qaKBjczLw42YHUYVWiHOsMc6IiFPqFUxem9TdVqgD9VaUz6BhdRfapv6ORVG+5yKox2fhc+/YtXudbJXj87m3OkX/Pjs1vrL1d2Id/lDdRMRqYHWz46iGpIcjYk6z4xhNK8TZCjGOpsh1tx0+3/HyZzCyItffsfD3fFCnfBZFr7vt/j20+/HVW9Hqb9G/T8d3qKKkaw4C03LPu1OZWdG57lorc/21VuW6a63M9dfqriiNvIeAWZJOkzQJuBxY3+SYzKrhumutzPXXWpXrrrUy11+ru0Kka0bEkKRrgE3ABGBNRDze5LDGqzDd66NohTgLG2Ob1N3Cfr4N1JGfQZvU37HoyO+5gpb+LNqo7rb091CFdj++I9LC9bfo36fjyynExCtmZmZmZmZWG0VJ1zQzMzMzM7MacCPPzMzMzMysjbiRVyeSrpcUkk5OzyVplaR+SY9JOruJsf0PSd9PcXxV0vG5bStSjE9JurBZMebiWZBi6Ze0vNnxtAtJH5E0KOnR9HNxbluh6kA9uX51hlY65zWC630xtNv3IGmNpGclfS9XdqKkzZK2p98nNDNGOzKVzqGSZkr6ee5a4m+aFF+h/i1JmibpAUlPSHpc0nWpvOK1Vz24kVcHkqYBFwA/yhVfBFISOy4AACAASURBVMxKP0uA25sQWslm4MyI+E/A/wVWAEiaTTbD05uABcBnJE1oVpDpb3+a7LObDVyRYrTauCUizko/G6F4daCeXL86Skuc8xrB9b4Y2vR7+BzZv6O85cD9ETELuD89t9ZT9hya/CB3LfHHjQ6soP+WhoDrI2I2MA+4OhfTYdde9eJGXn3cAvwZkJ/VZiFwZ2S2AMdLmtKM4CLinyJiKD3dQrY+SynGdRHxy4j4IdAPnNuMGJNzgf6I2BER+4F1KUarn6LVgXpy/eoQLXTOawTX+2Jou+8hIr4J7B1WvBBYmx6vBS5taFBWEyOcQ4ugcP+WImJ3RDySHr8EPAlMbXQcbuTVmKSFwGBEfHfYpqnArtzzAZrwhZfxPuDr6XHRYixaPO3mmpR6sSaXQtNJn3knHasdVORzXiN04jEXUad8D10RsTs93gN0NTMYq4n8ORTgNEnfkfR/JP1WE+Ip9L8lSTOBtwIPpqJy1151UYh18lqNpP8NTC6z6c+BD5OlajbVSDFGxL1pnz8n61L+QiNjs8YYpZ7eDtxI1tt8I/AJshO3WUvyOc+s2CIiJHndroI6wnPobmB6RPxE0jnA1yS9KSJebEjQBSfptcCXgQ9GxIuSGnrt5UbeEYiI3ylXLunNwGnAdyVB1p39iKRzgUFgWm737lTW0BhLJL0HuAQ4Pw4ultjQGKtQtHhaymh1oETS3wEb0tNO+sw76VjbXpuc8xqhE4+5iDrle3hG0pSI2J2GqDzb7ICsvCM5h0bEL4FfpsdbJf0A+HXg4fpGe4hC/luSdBRZA+8LEfEVgIh4Jrc9f+1VF07XrKGI2BYRp0bEzIiYSdZlfHZE7AHWA4uUmQe8kEthaChJC8jGDP5eRLyc27QeuFzS0ZJOI5sk5tvNiDF5CJgl6TRJk8gmSFjfxHjaxrDxoO8ASrOhFa0O1JPrV4dooXNeI7jeF0OnfA/rgcXp8WLg3ibGYkeo0jlU0imlyaok/RrZOXRHg8Mr3L8lZT09dwBPRsQnc+WVrr3qwj15jbMRuJhsYP/LwHubGMungKOBzanHcUtE/HFEPC7pHuAJsu74qyPi1WYFGRFDkq4BNgETgDUR8Xiz4mkzfyXpLLKUgaeB/wJQtDpQT65fHaUlznmN4HpfDO34PUi6C+gBTpY0ANwA3AzcI+kqYCdwWfMitHEoew4Ffhv4qKRXgH8H/jgihk++U1cF/bd0HnAlsE3So6nsw2Qzfx527VUvOpi1YmZmZmZmZq3O6ZpmZmZmZmZtxI08MzMzMzOzNuJGXguQdJ+kj5YpXyhpj6SJks6W9E1J+yQ9I+m6ZsRqNlwV9ffrqd6WfvZL2taMWM3yqqi7R0v6m3TO3Svpf0oqzPpM1tmqqL8nS1or6dn085EmhGlWTV19u6QHJL0g6eky+81M21+W9H1JVc0u3u7cyGsNa4E/TLP15F1JtlbJ8cB9wN8CJwGnA//U0AjNKhux/kbERRHx2tIP8C/APzY8SrPDjXbuvQ74DeA/Ab8KPA/c1tAIzSobrf7+D+A/ADOBc4ErJTVzUjjrXKPV1ReANcCfVnj9XcB3yK6B/xz4kqRT6hRry/DEKy1A0jHAHuA/R8Q3U9kJZItQzgXeBUyLiCubF6VZeaPV34j4bm7fmcAPgDdExNMND9Ysp4pz7x8DL0XEn6Vtvwt8MiLOaFLIZgdUUX/vBy6KiIfStg+n57/VpJCtQ1V7nZB66D6blikrvfbXgW3AyRHxUir7Z7KbyH/T0AMpGPfktYCI+DlwD7AoV3wZ8P1U8ecBeyX9S0q5+J+SpjcjVrPhqqi/eYuAf3YDz4qgirp7B3CepF+V9B+AdwNfb3ykZoer8tyb7zkRcGaDwjM7YIzXCcO9CdhRauAl303lHc2NvNaxFninpNek54tSGUA32SKj1wHTgR+SdV2bFcVI9TdvEfC5RgVlVoWR6u52YBcwCLwI/EfgsHElZk00Uv29D1gu6XWSTgfeR5a+adYM1V4nDPdasnTOvBeA19UwtpbkRl6LiIhvAT8GLpX0BrL8+S+mzT8HvhoRD0XEL4D/Dvw/kl7fnGjNDjVK/QVA0m8Ck4EvNT5Cs/JGqbufJlsg+CTgWOAruCfPCmSU+vsnZNcP24F7yW4ODzQjTrNqrhMq2AccN6zsOOClMvt2lInNDsDG5E6yOxtnAJsi4plU/hiQH1zpgZZWRJXqb8li4CsRsa/hkZmNrFLdPQv484jYCyDpNuCjkk6OiB83J1Szw5Stv6nevru0k6SPA99uSoRmmdGuE8p5HPg1Sa/LpWy+heoaiG3NPXmt5U7gd4D3c2gX9t8D75B0lqSjgP8GfCsihndfmzVTpfpbGnR9GU7VtGKqVHcfAhZJen06934A+Dc38KxgytZfSW+QdJKkCZIuApYANzUpRjOoXFd/JaVxHpU91WskTQKIiP8LPArckMrfQTbj8ZcbHn3BuJHXQtJkFP9Clha0Plf+DeDDwP8CniVbQuEPmhCiWUWV6m9yKfBT4IEGh2U2qhHq7lLgF2Tpbs8BFwPvaHR8ZiMZof6eQzYr4UvAXwLvjojHGx6gWTJCXf1tstTijWRzT/ycQ5cKuxyYQ7aMzc3AOyPiuQaEXGheQsHMzMzMzKyNuCfPzMzMzMysjbiRZ2ZmZmZm1kbcyDMzMzMzM2sjbuSZmZmZmZm1kZZdJ+/kk0+OmTNn1vQ9f/azn3HsscfW9D0bzccAW7du/XFEnFLDkGqqHnW3GkWrG47ncEWvu9CY+luE76JIcUBxYhkpjqLX33rX3aJ8R7XQTscCox9P0esu1Lb+ttv3W612Pe5K9bdlG3kzZ87k4Ycfrul79vX10dPTU9P3bDQfA0jaWbtoaq8edbcaRasbjudwRa+70Jj6W4TvokhxQHFiGSmOotffetfdonxHtdBOxwKjH0/R6y7Utv622/dbrXY97kr11+maZmZmZmZmbcSNPDMzMzMzszbiRp6ZmZmZmVkbcSOvDUzpno4kJPHYY9uaHY7ZqKZ0T2fr1q1IYkr39GaHY1ZYpfO7z+1Wb76WsHZVqttbt27tqGuOlp14xQ7aM7iLGcs2APDKK99vcjRmo9szuItJk09nxrIN7Fx5SbPDMSus0vnd53arN19LWLsq1e1Jk4fYM7ir2eE0jHvyzMzMzCxHzrQwa3Fu5JmZmZlZTjBj2YaO6vUwazdu5JmZmZmZmbURN/LMzMzMzMzaiBt5ZmZmZmZmbcSNPDOzgpJ0vKQvSfq+pCcl/YakEyVtlrQ9/T4h7StJqyT1S3pM0tm591mc9t8uaXGu/BxJ29JrVklSM47TzMzMasuNPDOz4roVuC8i3gi8BXgSWA7cHxGzgPvTc4CLgFnpZwlwO4CkE4EbgLnAucANpYZh2uf9udctaMAxmZmZWZ25kWdmVkCSXg/8NnAHQETsj4ifAguBtWm3tcCl6fFC4M7IbAGOlzQFuBDYHBF7I+J5YDOwIG07LiK2REQAd+bey8zMzFqYF0M3Myum04DngL+X9BZgK3Ad0BURu9M+e4Cu9HgqkJ/vfCCVjVQ+UKb8MJKWkPUO0tXVRV9f3xEfVDX27dtX97/RKnH09vYyafIQp07sbnosUIzPxMzMRudGnplZMU0EzgaujYgHJd3KwdRMACIiJEW9A4mI1cBqgDlz5kRPT09d/15fXx/1/hutEsf8+fOZsWwD154ywLve9a6mxgLVfSaSppH1DHcBAayOiFtT6vDdwEzgaeCyiHg+jQW9FbgYeBl4T0Q8kt5rMfAX6a1vioi1qfwc4HPAMcBG4LrUI21mZjhd08ysqAaAgYh4MD3/Elmj75mUakn6/WzaPghMy72+O5WNVN5dptxsvIaA6yNiNjAPuFrSbDye1MysYdzIMzMroIjYA+ySdEYqOh94AlgPlGbIXAzcmx6vBxalWTbnAS+ktM5NwAWSTkgXyBcAm9K2FyXNSz0pi3LvZXbEImJ3qScuIl4imzBoKh5PambWME7XNDMrrmuBL0iaBOwA3kt2c+4eSVcBO4HL0r4bydLd+slS3t4LEBF7Jd0IPJT2+2hE7E2PP8DBlLevpx+zmpE0E3gr8CBtPp601ccrlsZ/Apw6sZvrJw+xv7e3pY+ppNW/G7Mj4UaetTVJxwOfBc4kGxvyPuApPC7EWkBEPArMKbPp/DL7BnB1hfdZA6wpU/4w2b8Ns5qT9Frgy8AHI+LF/DKM7TietAhjOMejNP4T4NpTBrjtuTeyc+VS2uG/tFb/bsyOhNM1rd15nTEzswaTdBRZA+8LEfGVVOzxpGZmDeJGnrUtrzNmZtZ4KSviDuDJiPhkbpPHk5qZNYjTNa2ddew6Y+UUaUxCb28vXcfA9W8uzpiPIn0+Zi3uPOBKYJukR1PZh4Gb8XhSM7OGcCPP2lnHrjNWTpHGJMyfP59Vn/8an9g2sTBjPor0+Zi1soj4FqAKmz2e1FqCpAnAw8BgRFwi6TRgHXAS2U3jKyNiv6SjyTJ5zgF+ArwrIp5O77ECuAp4FfiTiNiUyheQDSeZAHw2Im5u6MFZR6g6XVPSBEnfkbQhPT9N0oOS+iXdnWZ/Q9LR6Xl/2j4z9x4rUvlTki7MlS9IZf2Slg//22ZHyOuMmZmZ2ZG4jmwcf8lK4JaIOB14nqzxRvr9fCq/Je1HWhvycuBNZOP1P5OupScAnyabB2A2cEXa16ymxjImz5XdWorXGTMzM7OxktQN/C7Z7NylcaZvI7tZDIeP5y+N8/8ScH7afyGwLiJ+GRE/JEtHPjf99EfEjojYT9Y7uLD+R2Wdpqp0zVxl/xjwoVxl/4O0y1rgI2QzDS5MjyGr7J8aXtmBH0oqVXZIlT39rVJlf2JcR2aW8TpjZmZmNhZ/DfwZ8Lr0/CTgpxExlJ7nx+AfGLcfEUOSXkj7TwW25N4z/5rh4/znlguiXuP5O20MemkNyK5jssedcuzVjskrRGU3GyuvM2ZmZmbVknQJ8GxEbJXU08xY6jWev9PGoJfWgLz+zUMsXVqMeQAaYdRGXpEqe71nKGzVOxulOxQAp07sbsljyGvV78HMzMxa3nnA70m6GHgNcBzZJCnHS5qYOjjyY/BL4/YHJE0EXk82AUul8fyMUG5WM9X05BWmstd7hsJWvbNRukMBcO0pA7zrXe9qckTj06rfg5mZmbW2iFgBrABInRtLI+Ldkv4ReCfZGLrh4/kXA/+atn8jzdy9HviipE8CvwrMAr5NNvPsrDRb5yDZfBWl4U9mNTPqxCsRsSIiuiNiJllF/EZEvBt4gKwyQ/nKDrnKnsovT7NvnsbByv4QqbKncVOXp33NzMzMzIpgGdm8FP1kw5DuSOV3ACel8g+RlmqKiMeBe8jmmLgPuDoiXk2dI9eQTer2JHBP2tespsazTt4yYJ2km4DvcGhl/4dU2feSNdqIiMcllSr7EKmyA0gqVfYJwBpXdjMzMzNrpojoA/rS4x0cnDAwv88vgN+v8PqPkU1aOLx8I9lkb2Z1M6ZGniu7mZmZmZlZsY1lnTwzMzMzMzMrODfyzMzMzMzM2ogbeWZmZmZmZm3EjTwzMzMzM7M24kaemZmZ1ZSkNZKelfS9XNlHJA1KejT9XJzbtkJSv6SnJF2YK1+QyvolLc+VnybpwVR+d1qCyczMEjfyzMzMrNY+BywoU35LRJyVfjYCSJpNttzSm9JrPiNpgqQJwKeBi4DZwBVpX4CV6b1OB54Hrqrr0XSqCUchCUlM6Z7e7GjMbAzcyDMzM7Oaiohvkq2VW42FwLqI+GVE/BDoJ1ui6VygPyJ2RMR+YB2wUJKAtwFfSq9fC1xa0wOwzKuvMGPZBmYs28CewV3NjsbMxsCNPDMzM2uUayQ9ltI5T0hlU4F8C2IglVUqPwn4aUQMDSs3M7NkTIuhm5mZmR2h24EbgUi/PwG8r55/UNISYAlAV1cXfX19dftb+/btq+v718tjj23jlVf209vby6TJWbv51IndXD95iP25sv29vS15fNC6343ZeLiRZ2ZWYGlc0sPAYERcIuk0srS1k4CtwJURsV/S0cCdwDnAT4B3RcTT6T1WkI1ZehX4k4jYlMoXALcCE4DPRsTNDT046ygR8UzpsaS/Azakp4PAtNyu3amMCuU/AY6XNDH15uX3H/43VwOrAebMmRM9PT3jP5AK+vr6qOf718v8+fOZsWwDO1dewoxl2Vdy7SkD3PbcG9m5cumBsp0rlxIRzQz1iLXqd2M2Hk7XNDMrtuuAJ3PPK004cRXwfCq/Je13pJNaWOGo5SfAkDQl9/QdQGnmzfXA5ZKOTjcxZgHfBh4CZqWZNCeR1eP1kbU0HgDemV6/GLi3EcdgZtYq3MgzMysoSd3A7wKfTc9HmnBiYXpO2n5+2n9Mk1rU/6jsyERLTYAh6S7gX4EzJA1Iugr4K0nbJD0GzAf+K0BEPA7cAzwB3AdcHRGvpl66a4BNZDc67kn7AiwDPiSpn6xX+44GHp6ZWeE5XdPMrLj+Gvgz4HXp+UgTThyYpCIihiS9kPafCmzJvWf+NcMntZhbLohGjmuC4oyfKUIcpXFSpTFS0NyxUdV+JhFxRZniig2xiPgY8LEy5RuBjWXKd5DdqDAzszLcyGs7WTrP5KnT2D3wo2YHY2ZHSNIlwLMRsVVSTzNjaeS4JijO+JkixFEaL1UaIwXNHRtVhM/EzMxG53TNthMtk87TKGn80XckbUjPT5P0oKR+SXensR6k8SB3p/IHJc3MvceKVP6UpAtz5QtSWb+k5Y0+Nmtr5wG/J+lpslTKt5FNknK8pNINuvyEEwcmr0jbX082QUWlSS1GmuzCzMzMWpgbedYJPHGFtZyIWBER3RExk6z+fSMi3k3lCSfWp+ek7d9IE1SMaVKLBhyamZmZ1ZkbedbWPHGFtaFKE07cAZyUyj8ELIcjntTCzMzMWpjH5Fm7K8TEFWbjERF9QF96XHbCiYj4BfD7FV4/pkktzMzMrLW5kWdtq0gTVzR6dsJyijBTYElvby9dx8D1bx5q6kyBeUX6fMzMzMzGw408a2eliSsuBl4DHEdu4orUm1du4oqBKieuYITyQzR6dsJyijQr3vz581n1+a/xiW0TmzpTYF6RPh8zMzOz8fCYPGtbnrjCzMzMzDqRe/KsEy0D1km6CfgOh05c8Q9p4oq9ZI02IuJxSaWJK4ZIE1cASCpNXDEBWOOJK8zMzMys2dzIs47giSvMzMzMrFM4XdPMzMzMzKyNuJFnZmZmZmbWRtzIMzMzMzMzayNu5JmZmZmZmbURN/LMzMyspiStkfSspO/lyk6UtFnS9vT7hFQuSask9Ut6TNLZudcsTvtvl7Q4V36OpG3pNaskqbFHaGYtacJRSGJK9/RmR1J3buSZmZlZrX0OWDCsbDlwf0TMAu5PzwEuIlt/dBawBLgdskYhcAMwl2xG5BtKDcO0z/tzrxv+t8zMDvfqK8xYtoE9g7uaHUnduZFnZmZmNRUR3yRbbzRvIbA2PV4LXJorvzMyW4DjJU0BLgQ2R8TeiHge2AwsSNuOi4gtERHAnbn3snrpoB4Qs3bgdfLMrGGmdE/viLtnZlZWV0TsTo/3AF3p8VQgf2IYSGUjlQ+UKT+MpCVkvYN0dXXR19c3viMYwb59++r6/vXS29vLpMlD7E+/AU6d2M31w8r2r/xLJk0+nf17+lvuOFv1uzEbj1EbeZKmkd0l6wICWB0Rt6Y0iruBmcDTwGUR8XzKi78VuBh4GXhPRDyS3msx8BfprW+KiLWp/Byy1I5jyBaWvi7dnTOzNrJncBczlm1g58pLmh2KmTVRRISkuv8/HxGrgdUAc+bMiZ6enrr9rb6+Pur5/vUyf/78dF5eyoxlGwC49pQBbnvujYeUlR7vXLmUVrtEG8t34+ve9tHpN5arSdccAq6PiNnAPOBqSbNxbr2ZmZlV75mUakn6/WwqHwSm5fbrTmUjlXeXKTerBV/3tonSjeVONWojLyJ2l+5IRMRLwJNkaRHOrTczM7NqrQdKM2QuBu7NlS9Ks2zOA15IaZ2bgAsknZAuji8ANqVtL0qal3pRFuXey2xcfN1r7WJMY/IkzQTeCjxIE3LrzczMrPgk3QX0ACdLGiDr0bgZuEfSVcBO4LK0+0ayVLd+snS39wJExF5JNwIPpf0+GhGlyVw+wMF0t6+nH7OaavZ1b73GlHbKGMXh4027jjm0rN0/g6obeZJeC3wZ+GBEvJhfkqZRufX1HkDdqpW+t8Jg6VY8Fmjd78HMzDIRcUWFTeeX2TeAqyu8zxpgTZnyh4EzxxOj2UiKcN1brzGlrTp+dKyGjze9/s1DLF3aumNLx6qqRp6ko8gq+hci4iup+BlJUyJi9xhy63uGlfcxhtz6eg+gbtVKX6rEcOhg6VatvK36PZiZmVnrK8p1r9l4jDomL+W73wE8GRGfzG1ybr2ZmZlZC5rSPZ1875RlfN1r7aKanrzzgCuBbZIeTWUfxrn1ZmZmZi3JS9pU5OteawujNvIi4ltApVs9zq03MzMzs7bg615rF9Wsk2dmZg0maZqkByQ9IelxSdel8hMlbZa0Pf0+IZVL0ipJ/ZIek3R27r0Wp/23p8V5S+XnSNqWXrNKzt0yMzNrC27kmZkVkxfkNTMzsyPiRp6ZWQF5QV4zMzM7UmNaDN2slUiaRnbh2gUEsDoibk09G3cDM4Gngcsi4vmUqnYr2QDql4H3lC6yU4rbX6S3viki1qbyczg4eHojcF206toVVljtuiBvJUVZK7MIcZTWQS2tgQo0dR3UInwmZmY2OjfyrJ2V0t0ekfQ6YKukzcB7yNLdbpa0nCzdbRmHprvNJUtlm5tLd5tD1ljcKml96hUppbs9SNbIW4BnybIaaucFeSspylqZRYijtA5qaQ1UoKnroBbhMzEzs9E5XdPaltPdrNWNtCBv2l7tgryVyr0gr5mZWRtyT551hE5LdyunCGlWpdSz/b29dB0D1795qKmpZ3lF+HzyqliQ92YOX5D3GknryHqiX4iI3ZI2AR/PTbZyAbAireH0Ylq890GyBXlvq/uBmZmZWd25kWdtrxPT3copQppVKfVs58qlrPr81/jEtolNTT3LK8LnM4wX5DUzM7Mj4kaetbWR0t1SL0e16W49w8r7cLqb1ZEX5DUzM7Mj5TF51raqSHeDw9PdFqVFpeeR0t2ATcAFkk5IKW8XAJvSthclzUt/a1HuvczMrAxJT0vaJulRSQ+nshMlbZa0Pf0+IZVL0ipJ/ZIek3R27n0Wp/23pxmQzcwscSPP2lkp3e1t6WLiUUkXk6W7vV3SduB30nPI0t12kKW7/R1ZKhspta2U7vYQh6e7fTa95gc43c3MrBrzI+KsiJiTni8nm/V4FnB/eg6Hznq8hGxGY3KzHs8FzgVuyI07NTPreE7XbGFTuqezZ3DX6Dt2KKe7mZm1jIUcTItfS5YSv4zcrMfAFkmlWY97SLMeA6TlcRYAdzU27A404ShKY9snT53G7oEfNTkgMyvHjbwWtmdwV5rE4pJmh2JmZlatAP4pTXr1t2liqlrNenyIRs5sXLQZekeTn+04/xvg1IndXD+s7MB+K/+SSZNPz8r29LfEMbfad2NWC27kmZmZWSP9ZkQMSjoV2Czp+/mNtZz1uJEzGxdwht4R5Wc7zv8GuPaUAW577o2HlJXbryizI4+m1b4bs1rwmDwzMzNrmIgYTL+fBb5KNqbumZSGyRhmPS5XbiOY0j0dSQfSLc2sfbmRZ2ZmZg0h6VhJrys9Jput+HvUaNbjBh5KSyoN8yj1xJl1rDS2VBJTuqc3O5q6cLqmmZmZNUoX8NXUkzQR+GJE3CfpIeAeSVcBO4HL0v4bgYvJZjB+GXgvZLMeSyrNegyHznpsZjayV1/JpR2359wWbuS1q3SHwjNfmZlZUUTEDuAtZcp/Qo1mPTYzM6drtq90h8JLLJiZmZmZdRY38szMzMzMzNqIG3lmZmZmZmZtxI08MzMzMzOzNuJGnpmZmZmZWRtxI8/MzMysjZUWQTezzuFGnpmZmVkbKy2CXnNpuaZ2XUzarJW5kddBSnfyfDK2QvFFgplZa/JyTVYwpWtd91y7kddRSnfyfDK2QvFFgpmZmdVA6Vq3Lj3XLcaNPDMzMzMzszbiRp6ZmZmZHbmUdu/Ue7PimNjsAKzO0onXzMzMrC5S2j3AzpWXNDkYMwP35LW/dOJ1brKZmZmZWWcoTCNP0gJJT0nql7S82fG0NadV1JTrrrUy19/i8exw1XHdHZnrUbG5/lq9FaKRJ2kC8GngImA2cIWk2c2Nqo3levc8o+H4uO5aK3P9LaZRZ4fzsiOuu1Vo2iyDrp+jcv0tmDats4Vo5AHnAv0RsSMi9gPrgIVNjqmQfGeucFx3R+E6W2iuvwVS+rcyKi87Aq67xVWqn3v2OGuoMtffGqv6/FlOm9ZZRUSzY0DSO4EFEfFH6fmVwNyIuGbYfkuAJenpGcBTNQ7lZODHNX7PRvMxwIyIOKVWwYykQHW3GkWrG47ncA2ru1Do+luE7wKKEwcUJ5aR4uj0c29RvqNaaKdjgdGPp9POve32/VarXY+7bP1tqdk1I2I1sLpe7y/p4YiYU6/3bwQfQzHVu+5Wo2ifq+NpHY2uv0X5LooSBxQnlqLEUa1G1t1W+2xG0k7HAq17PPWqv636eYxXpx13UdI1B4Fpuefdqcys6Fx3rZW5/lqrct21Vub6a3VXlEbeQ8AsSadJmgRcDqxvckxm1XDdtVbm+mutynXXWpnrr9VdIdI1I2JI0jXAJmACsCbi/2fv3uPlKuu773++JgRRRI4mu9k50BLxBlSEPIQ+9raJnAKlRu8XYmhLolJpH4FiJTXBtg8WxJI2SMEqvUFSEg8Eiqc0BkJEdq19Gg5BTuHQbAMhO3cCVMIhgoTE3/PHuiasTGZ29mEOa2a+79drXnvmWmtmWXdkDQAAIABJREFU/9bMNWvWtdZ1/a5Y04RQmtqdrka8DQ1UoLo7EEV7Xx1PkxW4/hblsyhKHFCcWAoRR0HrbiHemxppp22Bgm1PAepvod6PBuqo7S5E4hUzMzMzMzOrjaJ01zQzMzMzM7MacCPPzMzMzMysjXR8I0/S30t6XNJDkr4naf/csosl9Up6QtIpzYxzTyRNT3H2SprX7HgGQtI4SXdJelTSGkkXpvIDJa2UtDb9PaDZsbaqavVb0kRJr0p6IN3+qYExNbWu9lPvviBpY+49Oa3RsRlIukhSSDo4PZaka1J9eUjSMQ2IoTC/C838vngfPTCSPpren19Lmly2rGWOI0qavY8eLkkLJT0r6ZFcmessxdq3NVqr1+shiYiOvgEnAyPT/fnA/HT/COBBYG/gUODnwIhmx1tlG0ak+H4TGJXiPqLZcQ0g7i7gmHT/bcB/pff974B5qXxe6TPxbUjvcbX6PRF4pAnxNL2u9lPvvgDMafZn1sk3spTiK4D1wMGp7DTgNkDA8cDdDYijEL8Lzf6+eB894Pfpf5BNVN0DTM6Vt8xxRC7mpu+ja7ANHwCOyf/Guc7ufB8KsW9rwna3fL0eyq3jr+RFxB0RsT09XEU2VwnADGBJRLwWEU8CvcBxzYhxAI4DeiNiXURsA5aQxV9oEbEpIu5P918GHgPGksW+KK22CPhwcyJsff3U72Zpel3tp95Z810FfA7IZwSbASyOzCpgf0ld9QyiQL8LTf2+eB89MBHxWEQ8UWFRKx1HlDR9Hz1cEfET4PmyYtdZCrVva7SWr9dD0fGNvDKfJDtjDNkP2Ybcsj6KeyDYSrFWJGki8D7gbmB0RGxKizYDo5sUVrvJ12+AQyX9TNK/SfqfDYqhUHW1rN4BnJ+6sSzs1O48zSJpBrAxIh4sW9TsOtPM34Vmb/tO3kcPSWE+v0FoxZgHwnV2d616zDsU7b59FRVinrx6k/QjYEyFRX8ZET9I6/wlsB34ViNjM5C0L/Ad4DMR8ZKkncsiIiR5no9+DLF+bwLGR8QvJB0LfF/SkRHxUkOCLoAK9e5a4DKyq0iXAVeS/QhajfRXV4HPk3Ulanos/l3YlffRA6sv1hravc5632YlHdHIi4gT+1su6ePA6cAJkTrvAhvJxoeUdKeyImqlWHchaS+yg4dvRcR3U/EzkroiYlPqlvVs8yIsvqHU74h4DXgt3V8t6efAO4H76httMepqpXoXEc/kll8PLGt0XO2uWl2V9G6ycSAPpgZEN3C/pOOoU51pkd+Fpn9fvI/O7Km+VNH0z28IWjHmgeiYOtsi+7ZGa/ftq6jju2tKmk42BuRDEfFKbtFSYKakvSUdCkwC7mlGjANwLzBJ0qGSRgEzyeIvNGVHczcAj0XEl3OLlgKz0/3ZgM+SDlG1+i3pEEkj0v3fJKvf6xoQUtPrarV6VzbO6yPAI+XPtfqIiIcj4h0RMTEiJpJ1pTkmIjaT1Y9ZyhwPvJjrdlUXBfpdaOr3xfvoYWul44iSpu+j68R1lkLt2xqtXet1vzriSt4e/CNZNqGV6Qzyqoj404hYI+kW4FGyS9rnRcSOJsZZVURsl3Q+WVa6EcDCiFjT5LAG4v3A2cDDkh5IZZ8HrgBukXQOWZa9M5sUXzuoWL/Jso9dKul14NfAn0ZE+UD1mitIXa1W786SdDRZd82ngD9pcFxW2XKyDJu9wCvAJxrwPwvxu1CA74v30QMg6SPAV4BDgB9KeiAiTmml44iSAtS5YZN0EzAVOFhSH3AJrrMlhdi3NVo71Ouh0BtXas3MzMzMzKzVdXx3TTMzMzMzs3biRp6ZmZmZmVkbcSOvgCTdLunSCuUzJG2WdJKkuyS9KOmpCutdJulhSdslfaERMZuVDKf+SnqHpJsk/Z+0/D8kTWlY8NbRarDvvUvSc5JekvSgsrn/zBpiuPU3t/7vSgpJX6xrwGZJDfa9T0l6VdLWdLujIYEXnBt5xbQI+KOU2SzvbLI5TV4EFgJ/UeX5vWTZk35YtwjNqhtO/d2XLAvWscCB6bV+qGyeLrN6G+6+90KgKyL2A84FvlmWtdWsnoZbf0tTZlxNNum9WaMMu+4Cvx8R+6Zbw+ZcLTI38orp+8BBwP8sFUg6gGxek8URcU9EfIMqKe8jYlFE3Aa83IhgzcoMuf5GxLqI+HJEbIqIHRFxHTAKOLxBsVtnG+6+96GI2F56COzFrnMzmdXTsOpvchFwB/B4PQM1K1OLumtl3MgroIh4FbgFmJUrPhN4PCIebE5UZgNTy/qbpjQYRXZ12qyualF3JS2T9CuyKyE9wH21jtOskuHWX0kTgE8Cu3WbM6unGh03fCt1l79D0ntrHmQLciOvuBYBZ0h6c3o8K5WZtYJh119J+wHfAP4mIl6scXxm1Qyr7kbE6cDbyOb2uyMifl37EM2qGk79vQb464jYWpfIzPo3nLr7h8BEYAJwF7BC0v41j7DFuJFXUBHxU+C/gQ9L+i3gOODbzY3KbGCGW38l7QP8K9lErX9bnyjNdleLfW9EvJ66zJ8s6UN1CNOsoqHWX0m/D7wtIm6uc4hmFQ1n3xsR/xERr0bEK+mY4QVyXT871chmB2D9Wkx2JuNwYEVEPNPkeMwGY0j1V9LeZP3z+4A/qV94ZlXVat87EvitmkVlNjBDqb8nAJMlbU6P3w7skPTuiHCWWGuUWu17AyhP4tJxfCWv2BYDJwKfInfJWtKb0uXsvbKHerOkUbnle6XlbwJGpuUjGhy72aDrb8rsdivwKjDbXd2sSYZSd98l6VRJ+6R98B8BHwD+rQnxW2cbyrHDXwPvBI5Ot6XA9cAnGhm4dbyh7HvHS3q/pFGp/C+Ag4H/aEL8haKIaHYM1g9JPcB7gTER8Voqm0rW5zjv3yJialp+IzC7bPknIuLGOoZqtpvB1l9Jv0uWrOJVIN/AOzUi/r3uAZslQ6i7/wO4ETgC2AGsBb4UEd9rVMxmJUM5dih7/o1AX0T8VV0DNSszhH3vkcBNZL0mfgU8AMyNiI5PeuVGnpmZmZmZWRtxd00zMzMzM7M24kaemZmZmZlZG3Ejz8zMzGoqJUC4R9KDktZI+ptUfqikuyX1Sro5lzxh7/S4Ny2fmHuti1P5E5JOyZVPT2W9kuY1ehvNzIrMjTwzMzOrtdeAD0bEe8myNU6XdDwwH7gqIg4DtgDnpPXPAbak8qvSekg6ApgJHAlMB74maUTKGP1V4FSyZDdnpXXNzIwWnifv4IMPjokTJwLwy1/+kre+9a3NDWgPih5j0eODgce4evXq/46IQxoQ0pDk6247aIW6MxTN2K6i111oTP0tYp0qWkxFiwd2rb+RZXXbmhbtlW4BfBD4g1S+CPgCcC0wI92HbBqVf5SkVL4kZdl7UlIv2STJAL0RsQ5A0pK07qPV4ivavreIn+Fwteo2ed9bDK1af4ZruNtdrf62bCNv4sSJ3Hdflh21p6eHqVOnNjegPSh6jEWPDwYeo6T19Y9m6PJ1tx20Qt0ZimZsV9HrLjSm/haxThUtpqLFA7vX33S1bTVwGNlVt58DL0TE9rRKHzA23R8LbACIiO2SXgQOSuWrci+bf86GsvIpFWI6FzgXYPTo0SxYsGCom1dzW7duZd999212GDXVqts0bdo073sLoIj7tUYY7nZXO3Zo2UaemZmZFVdE7ACOlrQ/8D3gXU2I4TrgOoDJkydHkQ4g2/GAth22SdJC4HTg2Yg4KpUdCNwMTASeAs6MiC3pavPVwGnAK8DHI+L+9JzZQGmewS9GxKJUfizZnJr7AMuBC8PzmVkdeEyemZmZ1U1EvEA2kfFvA/tLKp1g7gY2pvsbgXEAafnbgV/ky8ueU63cbLhuJBv/mTcPuDMiJgF3pseQjQmdlG7nknU9LjUKLyG7unwccImkA9JzrgU+lXte+f8yqwk38szMzKymJB2SruAhaR/gJOAxssbeGWm12cAP0v2l6TFp+Y/T1Y2lwMyUffNQsoPie4B7gUkpW+cosuQsS+u/ZdbuIuInwPNlxTPIxpCS/n44V744MqvITmJ0AacAKyPi+YjYAqwkSz7UBewXEatS/V6cey2zmnJ3zTrp6h7P5o3ZcIExY8dx0zcXNzkiMxus0vd4zNhxbOp7utnhWBOU6sDVV1/T8t3QGqwLWJTG5b0JuCUilkl6FFgi6YvAz4Ab0vo3AN9IiVWeJ2u0ERFrJN1CllBlO3Be6gaKpPOBFcAIYGFErGnc5lle6XuyYMECzvqjWe24vxwdEZvS/c3A6HR/51jSpDRmtL/yvgrlFZWPKe3p6Rn6FrSAZ599liuvvBKAvfYaxXve8+4mR9QYW7durctn60ZenWzeuIEJc5cBsH7+6U2OxsyGovQ99ne4c5XqwOuvP97sUFpKRDwEvK9C+TreyI6ZL/8V8NEqr3U5cHmF8uVkY5qsyUrfk1Fjtu88wd2uIiIkNWQMXZHHlNbDlVdeyVeey4burp9/Op0yVLFeY1ndXdPMzMzMamPEXkiiq3t8syOppWdSV0vS32dT+WDHjG5M98vLrVx71qOGciOvEUbsxerVq11RzczMrL3teJ0Jc5e12xW9/JjR8rGks5Q5HngxdetcAZws6YCUcOVkYEVa9pKk41Nmzlm517K89qxHDeVGXo11dY8n+97m7HidUWMOc0U1awGl7/Dq1aubHYqZmTWYpJuA/wQOl9Qn6RzgCuAkSWuBE9NjyLoLrwN6geuBTwNExPPAZWQJgu4FLk1lpHW+np7zc+C2RmyXdR6Pyasxj+Exa235sSVmZtZZIuKsKotOqLBuAOdVeZ2FwMIK5fcBRw0nRrOB8JU8M7MWI+nPJa2R9IikmyS9OaWSv1tSr6SbU1p5Uur5m1P53ZIm5l7n4lT+hKRTmrU9ZmbWuSr2grNhcyPPzKyFSBoL/BkwOSKOIksfPxOYD1wVEYcBW4Bz0lPOAbak8qvSekg6Ij3vSLLJeL+W0t2bmZk1TD4jvdWOG3mN5ExBZlYbI4F9JI0E3gJsAj4I3JqWl0/WW5rE91bghDTgfwawJCJei4gnycaH7Jba3szMzFqPx+Q1UsoU5PF6ZjZUEbFR0gLgaeBV4A5gNfBCRJQGEuYn2N05KW9EbJf0InBQKl+Ve+mqk/I2ekLeek0MOxQLFixg1JjtvGNkd2FigmK9R2ZmVjxu5NVAV/d4Z840s4ZI6bhnAIcCLwD/Qtbdsm4aPSFvvSaGHajyffqEucu44JCNzJw5E4AxY8exqe/pZoUHNP89MjMbDh87158beTWQ70vsq3RmVmcnAk9GxHMAkr4LvB/YX9LIdDUvP8FuaVLevtS98+3AL6g+WW/Hq7xPD+/nzcxqxMfO9ecxec2QxuZ5fJ6ZDcHTwPGS3pLG1p0APArcBZyR1imfrLc0ie8ZwI9T2u+lwMyUffNQYBJwT4O2wczMbM98zDxkvpLXDGlsHvjshZkNTkTcLelW4H5gO/Azsq6UPwSWSPpiKrshPeUG4BuSeoHnyTJqEhFrJN1C1kDcDpwXETsaujFmZmb98THzkO3xSp6khZKelfRIruxASSslrU1/D0jlknRNmnfpIUnH5J4zO62/VtLsXPmxkh5Oz7lGnijDzKxfEXFJRLwrIo6KiLNThsx1EXFcRBwWER+NiNfSur9Kjw9Ly9flXufyiPitiDg8Im5r3haZmZlZLQ2ku+aN7D6ofx5wZ0RMAu5MjwFOJevyM4ksE9u1kDUKgUuAKWQpui8pNQzTOp/KPa+uCQTMzMzMzMza2R4beRHxE7IuPnn5eZfK52NaHJlVZIkAuoBTgJUR8XxEbAFWAtPTsv0iYlUaI7I491pmZmZmZmY2SENNvDI6Ijal+5uB0en+zvmYktK8S/2V91UoNzMza6iu7vF4xICZmbWDYSdeiYiQFLUIZk+qTcjb7ElhS5PlAmxL97eVlY3e5431ypcVYULbZr+HA9EKMZpZ6yql9PbgfjOz+vD8eI0z1EbeM5K6ImJT6nL5bCqvNu/SRmBqWXlPKu+usH5F1SbkbfaksNOmTctl/pmTDhLm7FJ2zTe/z5w5lZdlPVWbq9nv4UC0QoxmZmZmVplPpjXOULtr5uddKp+PaVbKsnk88GLq1rkCOFnSASnhysnAirTsJUnHp6yas3KvZWZmVjxp3ibP2WRmZkU1kCkUbgL+EzhcUp+kc4ArgJMkrQVOTI8BlgPrgF7geuDTABHxPHAZcG+6XZrKSOt8PT3n50BnpfH2wYKZWWtJ8za5y5GZmRXVHrtrRsRZVRadUGHdAM6r8joLgYUVyu8DjtpTHG0rHSz4srWZmZmZmdXCULtrmhWepHGS7pL0qKQ1ki5M5QdKWilpbfp7QCqXpGsk9Up6SNIxudeandZfK2l2rvxYSQ+n51wjp+YzMzNrO5IOl/RA7vaSpM9I+oKkjbny03LPuTgdHzwh6ZRc+fRU1itpXuX/aDY8buRZO9sOXBQRRwDHA+dJOgKYB9wZEZOAO9NjgFOBSel2LnAtZI1C4BJgCnAccEmpYZjW+VTuedMbsF1mZmbWQBHxREQcHRFHA8cCrwDfS4uvKi2LiOUA6XhjJnAk2bHB1ySNkDQC+CrZMccRwFlpXbOaciPP2lZEbIqI+9P9l4HHyOZhnAEsSqstAj6c7s8AFkdmFbB/yh57CrAyIp6PiC3ASmB6WrZfRKxKXZUX517LzNpdGlPtcdVmHecE4OcRsb6fdWYASyLitYh4kiz3xHHp1hsR6yJiG7AkrWt74jwWgzLsefLMWoGkicD7gLuB0SmzK8BmYHS6PxbIZ1LoS2X9lfdVKK/0/yvO8dgO2m3+wtJ8luVzW7bTNlqNpDHVgMdVm3WWmcBNucfnS5oF3EfWg2gL2fHAqtw6+WOE8mOKKZX+STseO1SbM3rUmO28Y2Q3F/U37/T8v2XUmMPYtrm3Ld6LknodR7mRZ21P0r7Ad4DPRMRL+WFzERGS6j5RYbU5HttBu81fWJr38qJ3b99lbssizGdpZmbNJWkU8CHg4lR0LVkG+Uh/rwQ+WYv/1Y7HDqXf2PI5oyfMXcYFh/TxlefetUtZpfXa7Te5XsdR7q5pbU3SXmQNvG9FxHdT8TOpqyXp77OpfCMwLvf07lTWX3l3hXIzMzNrT6cC90fEMwAR8UxE7IiIX5NNH3ZcWm+wxxRmNeVGnrWtlOnyBuCxiPhybtFSoJQhczbwg1z5rJRl83jgxdStcwVwsqQDUsKVk4EVadlLko5P/2tW7rXMrAV0dY/fOa7OasOZja3NnUWuq2bppHHyEeCRdH8pMFPS3pIOJUvOdg/ZfNGTJB2argrOTOua1ZQbedbO3g+cDXywLLXxFcBJktYCJ6bHAMuBdWSDo68HPg0QEc+TdcG4N90uTWWkdb6envNz4LZGbJh1Nkn7S7pV0uOSHpP027U8gO4kmzduYMLcZTu7A1lNOLOxtSVJbwVOAr6bK/67dMLhIWAa8OcAEbEGuAV4FLgdOC9d8dsOnE92Avkx4Ja0rllNeUyeta2I+ClQ7ezuCRXWD+C8Kq+1EFhYofw+4KhhhGk2FFcDt0fEGelM8FuAz5MdQF+R5l2aB8xl1wPoKWQHx1NyB9CTycaSrJa0NCUMMBuy1MthU7r/sqR8ZuOpabVFQA9ZHd2Z2RhYlU5idKV1V5ZOqkkqZTbuIWU2TuWlzMY+yWZ1FRG/BA4qKzu7n/UvBy6vUL6c7MSyWd24kWdm1kIkvR34APBxgJSCe5ukmhxAs2vGOBuolNp7zNhxbOp7utnRFIYzG1fXTpmJnZXYGirtbwHvc/vhRt4wdHWPZ/PGDXte0cysdg4FngP+WdJ7gdXAhdTuAHo3jT5QbuTB74IKabxbIbV3KzQQnNm4f+2UmdhZia2hPH3NgLiRNwylsRyuYGZtrlhnDUcCxwAXRMTdkq7mjbFNQO0PoBt9oNzIg9/SwSn0n7K7aKm9i95A6C+zcURsGkRm46ll5T04s7FZS/FFkeZw4hUzsz1JZw0nzF1WhB+qPqAvIu5Oj28la/TVamoQs2FxZmMzy3OCq+ZwI68o0pUCSXR1j292NGZWUBGxGdgg6fBUdAJZ9raaHEA3ajusrTmzsZlZk7m7ZlG4f7GZDdwFwLdSZs11wCfITtrdIukcYD1wZlp3OXAa2cHwK2ldIuJ5SaUDaNj1ANqGyglYnNnYzKwA3MgzM2sxEfEA2dQH5WpyAG3DkE7Y+WSdmZk1k7trmplZx+nqHr8zmY6ZmVm7cSPPzMw6TikRgJmZWTtyI8/MzMzMzKyNuJFnZoa775mZmdVSQ35XU7IrZ6bfnRt5Zma4+56ZmVktNeR3NSW7KsActoXjRp6ZmZmZDZp7QJgVlxt5ZmZmZjZo7gFhVlxu5JmZmdVaGifisSJmZtYMngzdzMw6Qlf3+MaN20jjRABPjG5mZg3nK3lmZtYRSl3L3L3MzMzanRt5ZmZmZmZ7IOkpSQ9LekDSfansQEkrJa1Nfw9I5ZJ0jaReSQ9JOib3OrPT+mslzW7W9lh7cyPPzMzMzGxgpkXE0RExOT2eB9wZEZOAO9NjgFOBSel2LnAtZI1C4BJgCnAccEmpYWhWS27kmZmZmZkNzQxgUbq/CPhwrnxxZFYB+0vqAk4BVkbE8xGxBVgJTG900Nb+nHjFzMzMzGzPArhDUgD/OyKuA0ZHxKa0fDMwOt0fC+QzPfWlsmrlu5F0LtlVQEaPHk1PT0+NNqMxFixYwKgx29mW/gI771cqe8fIbi4awHrVylrt/SnZunVrXWJ3I8/MrAVJGgHcB2yMiNMlHQosAQ4CVgNnR8Q2SXsDi4FjgV8AH4uIp9JrXAycA+wA/iwiVjR+S8ysLaVpRADGjB3Hpr6nmxxQTfxORGyU9A5gpaTH8wsjIlIDsCZSI/I6gMmTJ8fUqVNr9dINMW3aNCbMXcb6+XNy2YbnVC274JA+vvLcu/a4XrWyiJq99Q3V09NDPT5bd9c0M2tNFwKP5R7PB66KiMOALWSNN9LfLan8qrQeko4AZgJHknUV+lpqOJqZDV+aRmTC3GWNm7qkziJiY/r7LPA9sjF1z6RumKS/z6bVNwLjck/vTmXVys1qyo08M7MWI6kb+D3g6+mxgA8Ct6ZVyseFlMaL3AqckNafASyJiNci4kmgl+yApe10dY/feUWhKdIVDU+Kbta6JL1V0ttK94GTgUeApUApQ+Zs4Afp/lJgVsqyeTzwYurWuQI4WdIBKeHKyanMrKbcyDMzaz3/AHwO+HV6fBDwQkRsT4/zYzx2jv9Iy19M6w94XEirK82P1zTpika7XM0w61CjgZ9KehC4B/hhRNwOXAGcJGktcGJ6DLAcWEd2Au164NMAEfE8cBlwb7pdmsraRtNPrBngMXnFlM76tlEfdjOrEUmnA89GxGpJUxv0Pxs6+L/Wg9AHO/i/FgkB6p0YoF4D9c2ssohYB7y3QvkvgBMqlAdwXpXXWggsrHWMRVE6sbZ+/unNDqWjDauRJ+kp4GWyQfvbI2Jymv/jZmAi8BRwZkRsSd2DrgZOA14BPh4R96fXmQ38VXrZL0bEIgqqq3t8/c/GprO+/nKYWQXvBz4k6TTgzcB+ZPvW/SWNTFfr8mM8SuM/+iSNBN5OloBlwONCGj34v9aD0Ac7+L8WCQHqnRigXgP1zcysPdSiu2ZHTQpZOjvR1K4/ZtaxIuLiiOiOiIlkiVN+HBF/CNwFnJFWKx8XUhovckZaP1L5TEl7p8yck8i6IJmZmVmLq8eYPE8KaWbWeHOBz0rqJRtzd0MqvwE4KJV/lnTiLSLWALcAjwK3A+dFxI6GR21mZmY1N9wxeYWYFLKRYxMWDHFMx+h9hjYupFHb1QrjO1ohRrNGiogeoCfdX0eF7JgR8Svgo1Wefzlwef0iNDMzs2YYbiOvEJNCNnJsQmlsBwxuTMc13/w+c+YUd2LHVhjfMZQYJS0ESokqjkplNRs3KulY4EZgH7JMWhdGq87GadZmGjKG2szMrICG1V3Tk0JaC7iR3bv/1nLc6LXAp3LPc1djs4Jo+tQJ5VLmZM+ZZ2ZWY56PdDdDbuR5UkhrBRHxE6B8/pmajBtNy/aLiFXp6t3i3GuZme0qZU72nHlmZjXm+Uh3M5zumqOB76XJDkcC346I2yXdC9wi6RxgPXBmWn85WTe4XrKucJ+AbFJISaVJIaENJ4W0wqnVuNGx6X55+W4aPc9YI7XLWMnyMbP9jaNth+01qyd3lTcza64hN/I8KaS1g1qPG+3n/zR0nrFGaoXxnANRPpfaRe/eXnUcrY8lzfboRuAfyXo4lJS6yl8haV56PJddu8pPIesGPyXXVX4yWaK31ZKWph4Vpa7yd5M18qYDtzVgu8zMWkI9plAwK7pajRvdmO6Xl5uZdTR3lTcza67hZtc0a0WlcaNXsPu40fMlLSE7m/xiRGyStAL4Ui7ZysnAxamr8UtpjOndwCzgK43cEDOzFuKu8jnt0NXd3dzNisuNPGtrkm4CpgIHS+oj6/pzBbUbN/pp3hgXchvuLmRmtkfuKt8eXd3dzd2suNzIs7YWEWdVWVSTcaMRcR9w1HBiNLPa8dx4hfaMpK7UQ2KgXeWnlpX34K7yZmZ75DF5ZmbWNkpz4xVqfjwrqckUS2nZS5KOT5k5Z+Vey8yaoKt7/M55QK0YfCXPzMzMaspd5c06S+kEG8D6+ac3ORoDN/LMzMwab8ReSGLM2HFs6nu62dHUnLvKm5k1l7trmlnHcvcSa5odrzNh7jKPHzQzs7pwI8/MOpbHb5mZDU6nnhyTNE7SXZIelbRG0oWp/AuSNkp6IN1Oyz3nYkm9kp6QdEqufHoq65U0rxnbY+3P3TXNzMzMbEA6eOzVduCiiLhf0tuA1ZJWpmVXRcSC/MqSjgBmAkcCvwH8SNI70+KvAieRzfF4r6SlEfFoQ7bCOoav5JmZDUYaS9XK0wPHAAAgAElEQVTVPb4p/76fs8kHSlopaW36e0Aql6Rr0hnjhyQdk3ut2Wn9tZJmV/ufraB0dcHMrB4iYlNE3J/uvww8Bozt5ykzgCUR8VpEPEmWWOi4dOuNiHURsQ1YktY1qylfyTMzG4w0lqqJZ7CrnU3+OHBnRFyRuv/MA+YCpwKT0m0KcC0wRdKBZBkPJwORXmdpRGxp+BbVQOnqQoddWTCzJpA0EXgfcDfwfuB8SbOA+8j2z1vIGoCrck/r441G4Yay8ilV/s+5wLkAo0ePpqenp2bbUGsLFixg1JjtAGxL98v/VlqWL3vHyG4uGsB6/ZVtW7Cg0O9TJVu3bq1LzG7kmZm1kDRH2KZ0/2VJpbPJM3hj4uhFZJNGz03li1MGw1WS9k8TUU8FVpZS0qeG4nTgpoZtjJlZi5G0L/Ad4DMR8ZKka4HLyE6WXQZcCXyyFv8rIq4DrgOYPHlyTJ06tRYvW1Nd3eN3JpB6oxvvnHTSbde/lZblyy44pI+vPPeuPa7XX9n6+XPIfu5aR09PD/X4bN3IG6B8JTYzK4Kys8mjUwMQYDMwOt0fy+5njcf2U17p/zT0bPJQzmouGOAZ3sGcEc6XDfYM82D+11Dez3qd+TWz6iTtRdbA+1ZEfBcgIp7JLb8eKGXy2giMyz29O5XRT3nLKVwvijSkAmjbKWoGyo28AWpKJXZFNbMqKpxN3rksIkJSzU5lNvps8lDOak6bNm1AZ3gHc0Y4XzbYM8yD+V9DOetcrzO/ZlaZsp3sDcBjEfHlXHlX7gTbR4BH0v2lwLclfZks8cok4B5AwCRJh5I17mYCf9CYregAaUgFdFxioN24kVdkrqhmVkGls8nAM6WDjdQd89lUXu1s8kbe6N5ZKu+pZ9xmZi3s/cDZwMOSHkhlnwfOknQ0WXfNp4A/AYiINZJuAR4lG0t9XkTsAJB0PrACGAEsjIg1jdwQ6wxu5JmZtZBqZ5PJzhrPBq5If3+QKz9f0hKywf0vpobgCuBLpSycwMnAxY3YBstxjw3rBKmet3Idj4ifkl2FK7e8n+dcDlxeoXx5f88zqwVPoWBm1lpKZ5M/WDb57hXASZLWAiemx5AdSKwjS999PfBpgJRw5TLg3nS7tJSEpVW0xaTMqcfGhLnLPO7b2leq567jZo3jK3lmZi2kn7PJACdUWD+A86q81kJgYe2ia6wOnpTZzMysX76SZ2ZmZmZmA9IWvSg6gBt5ZmZmZmY2IKVeFKWeFFZMbuSZmZmZmZm1ETfyzMyspZS6CrWdlIGwq3t8syMxM7MW50aemXWctm0kdIh8wpW24gyEZmZWI27kmVnHadtGgplZnfjkmLkOtBY38szMzMysXz45Zi1XBzq8C7wbeWZmZmZWfx1+0G0N1uFd4N3I64fnATFrL+5q0ro6an+cDoR9MGxtp8MPultRR+1724wbef0o1DwgPvtlNiT5H6iW62piOxVqf1xv6UDYB8Nm1mwdte9tM27ktQqf/TIbEv9AmZmZDU5b9Xzp0N4RbuSZmVlhtdWBxlC4F4eZNUFb9Xzp0N4RbuSZWVvq+MZBm2irA42hKPXi2Ly5I89EW3N5PFZn8efdXtzIazUdesnZbLDq3jjwd9EaKX8mevNmVq9e7Xpndefu7p2lIz7vDuod4UZeBYW+AtChl5zNCsffxbp56KGHfTa5PzteZ9SYw3Ze3euEgxVrMz5JZs3SQTkuCtPIkzRd0hOSeiXNa2YsLdM9qIPORhRZkepup3NXk8ErUv0tfX6vv76t/c8m10KFrpwj996nY34XilR3203dT3b7JFmh6m+hL27USwecaChEI0/SCOCrwKnAEcBZko5oblQtoIPORhSV626xdERXkxoqQv31FBc1kDtg3rHtVx0xhq8Idbfd+LvYOEWovx3/eXfAiYZCNPKA44DeiFgXEduAJcCMRvzjUiUvnf1syTMZubMRnXQWtyCaVnftDU0/C5m+g/n9SIt8B+tafyvtX0v3S3/dMK+TsjF8LVo/++N97zCVfz+b9l1s3f3ncDS8/hbm8y6iCnWwHY6nFRHNjgFJZwDTI+KP0+OzgSkRcX7ZeucC56aHhwNPpPsHA//doHCHqugxFj0+GHiMEyLikHoHAzWpu+2gFerOUDRjuxpWd6HQ9beIdapoMRUtHvC+d7CK+BkOV6tuk/e9xdCq9We4hrvdFevvyGG8YMNFxHXAdeXlku6LiMlNCGnAih5j0eOD1oixmmp1tx208ufSn3bdrqFodP0t4ntftJiKFk9RFXnf246fYTtuUzMVuf7WQ6fWn3ptd1G6a24ExuUed6cys6Jz3bVW5vprrcp111qZ66/VXVEaefcCkyQdKmkUMBNY2uSYzAbCdddameuvtSrXXWtlrr9Wd4XorhkR2yWdD6wARgALI2LNIF6iFS5lFz3GoscHBYyxBnW3HRTuc6mRdt2unQpcf4v43hctpqLF01AFrruD0Y6fYTtuU821Sf2th06tP3XZ7kIkXjEzMzMzM7PaKEp3TTMzMzMzM6sBN/LMzMzMzMzaSNs08iR9QdJGSQ+k22nNjglA0nRJT0jqlTSv2fFUIukpSQ+n9+2+ZscDIGmhpGclPZIrO1DSSklr098Dmhljp5P0UUlrJP1a0uSyZRenOv+EpFOaFeNQtMJ3tt1I+ntJj0t6SNL3JO1fZb2676v29PlL2lvSzWn53ZIm1iOO9L/GSbpL0qPpu3ZhhXWmSnox99v3/9YrHqsN7zvNdtWpdafev2lt08hLroqIo9NtebODkTQC+CpwKnAEcJakI5obVVXT0vtWlPlJbgSml5XNA+6MiEnAnemxNc8jwP8CfpIvTHV8JnAk2Wf4tfRdKLwW+862k5XAURHxHuC/gIv7Wbdu+6oBfv7nAFsi4jDgKmB+rePI2Q5cFBFHAMcD51Wpj/+e++27tI7xWG1432mWuO7U7zet3Rp5RXMc0BsR6yJiG7AEmNHkmFpCRPwEeL6seAawKN1fBHy4oUHZLiLisYh4osKiGcCSiHgtIp4Eesm+C63A39kmiIg7ImJ7eriKbM6oZhjI55/fD90KnCBJ9QgmIjZFxP3p/svAY8DYevwvaxzvO8124bpTJ+3WyDs/dfdZWJCufGOBDbnHfRTzBzqAOyStlnRus4Ppx+iI2JTubwZGNzMYq6pV6n0lrRx7u/gkcFuVZfXeVw3k89+5TmqYvggcVIdYdpG6hb4PuLvC4t+W9KCk2yQdWe9YrG5aef/TyrFbc3Vy3anrb1oh5skbKEk/AsZUWPSXwLXAZWRv2GXAlWQHC7ZnvxMRGyW9A1gp6fF0Ja2wIiIkef6POuvvOxcRP2h0PNa6BlKXJP0lWRfFb1V5mZbbV9WCpH2B7wCfiYiXyhbfD0yIiK1pLPr3gUmNjtF25X2nmQ1AXX/TWqqRFxEnDmQ9SdcDy+oczkBsBMblHnenskKJiI3p77OSvkd26byIB07PSOqKiE2SuoBnmx1Quxvod65MS9T7Klo59kLbU12S9HHgdOCEqDKBawP2VQP5/Evr9EkaCbwd+EUNY9iFpL3IGnjfiojvli/PN/oiYrmkr0k6OCL+u14x2Z5539lSsVtzdWzdqfdvWtt010wH/SUfIRvY3Gz3ApMkHSppFNmA6qVNjmkXkt4q6W2l+8DJFOO9q2QpMDvdnw34bGgxLQVmpiyEh5JdVbinyTENVOG/s+1I0nTgc8CHIuKVKus0Yl81kM8/vx86A/hxtUbpcKWxfjcAj0XEl6usM6Y0JlDScWS/63VrdFpded9pnagj604jftNa6kreHvydpKPJums+BfxJc8PJxmtIOh9YAYwAFkbEmiaHVW408L10jDAS+HZE3N7ckEDSTcBU4GBJfcAlwBXALZLOAdYDZzYvQpP0EeArwCHADyU9EBGnRMQaSbcAj5J1vTsvInY0M9aBapHvbDv6R2Bvsu4qAKsi4k8l/Qbw9Yg4jQbsq6p9/pIuBe6LiKVkja5vSOolSw41s5YxlHk/cDbwsKQHUtnngfEp3n8ia2j+P5K2A68CM+vV6LTa8L7T7A0dXHfq/psm/xaYmZmZmZm1j7bprmlmZmZmZmZu5BWSpNtT96Dy8hmSNks6SdJdkl6U9FSV17hQ0pOSfinpMUnvrHvgZgyv/koaL2lr2S0kXdSwDbCONdx9r6SjJf17Wt4n6a8bErgZNam//7ekeyS9rGw6qt9pSODW8QZQd/9C0iOpbj4p6S/K1puY6vYrkh6XNJTER23HjbxiWgT8UWkwfc7ZZKnFXwQWAn9R/kQASX8MnAP8HrAvWcY6Z1qzRhly/Y2IpyNi39INeDfwa7Lsgmb1Nqx9L/BtssxoBwK/C3xa0ofqFKtZuSHXX0kHAv8K/D2wP/B3wL+qGHMOW/vbU90VMAs4AJhONi92fjz0TcDPyOYs/UvgVkmH1D3qgvOYvAKStA/ZZN+/X5ovI+1oNwFTIuLBVHYiWVKCibnnvoksKcnHI+LORsduNpz6W+G1LgGmRsS0ugduHW+4dVfSK8DkiHg0Pf4X4P6I+NvGbYV1qmEeO5wOzI+II3Nl/5XKbmjcVlgnGmjdza1/DVkb5oLUU+1h4OCIeDkt/3eyaWf+qZHbUTS+kldAEfEqcAvZWYuSM4HHyyt6Bd3pdpSkDemy9t+kxp9Z3Q2z/u6UzujNIjvDZ1Z3Nai7/wDMkrSXpMOB3wZ+VPtIzXZXg/pbfhVFwFE1Cs+sqsHU3XRs8D+BUgbOI4F1pQZe8mAq72g+8C+uRcAZkt6cHg/0YLc7/T2ZrKvbNOAssu6bZo0y1Pqb9ztkKYZvrWVgZnswnLq7jGxKg1eBx4EbIuLe2odoVtVQ6+9/Ar8h6ax0kmI28FvAW+oUp1m5gdbdL5C1X/45Pd6XrCty3ovA2+oQY0txI6+gIuKnZOPoPizpt4DjyMZ77Mmr6e/fRcQLEfEU8L+B0+oSqFkFw6i/ebOB70TE1lrHZ1bNUOtuGtN0O3Ap8GZgHHCKpE/XMVyzXQy1/kbEL4AZwGeBZ8jGPf0I6KtftGZvGEjdVTaf3izg9yLitVS8Fdiv7OX2A16mw7XTZOjtaDFZZT4cWBERzwzgOU8A28gmhS/xwEtrhqHUX2Bn//yPAh+pU2xm/RlK3f1NYEdELE6P+yQtITvB9rX6hGlW0ZD2vRHxb8D/BSBpJLAOuLJeQZpVULXuSvokMA/4QETkTz6sAX5T0ttyXTbfy+BPLLcdX8krtsXAicCnyF2ylvSmdDl7r+yh3ixpFEBEvALcDHxO0tskdQPnknUjMmukQdffnI8AW4C7GhWsWc5Q6u5/pbI/SOuNAT4GPNTg2M2GtO+V9L7UVXM/YAGwISJWNDh262zV6u4fAl8CToqIdfknRMR/AQ8Al6Q6/RHgPTgrtxt5RZa6Wv5/wFuBpblFHyDrlrkcGJ/u35Fbfj7Z5ev/Q9bP/ttkaZPNGmYY9ReyrprfCKf/tSYYSt2NiJeA/wX8OdkJigeAR4AvNipuMxjWvvdzZN3lNgBduCeFNVg/dfeLZNMj3Ks35tDNZ86cCUwm2/deAZwREc81Juri8hQKZmZmZmZmbcRX8szMzMzMzNqIG3lmZmZmZmZtxI08MzMzMzOzNuJGnpmZmZmZWRtp2XnyDj744Jg4cWLdXv+Xv/wlb33rW+v2+kXQrtu4evXq/46IQ5odRzWDqbut8BkVPcZWiq/odRfqv+8drqJ/3nvSyvEXvf5Wqrut/H7Xgrc/2/6i113wcW8ljjlTrf62bCNv4sSJ3HfffXV7/Z6eHqZOnVq31y+Cdt1GSeubHUN/BlN3W+EzKnqMrRRf0esu1H/fO1xF/7z3pJXjL3r9rVR3W/n9rgVvf7b9Ra+74OPeShxzplr9dXdNMzMzMzOzNuJGnpmZmZmZWRtxI8/MzMzMzKyNuJFnbU/SCEk/k7QsPT5U0t2SeiXdLGlUKt87Pe5NyyfmXuPiVP6EpFNy5dNTWa+keY3eNjMzMzOzcm7kdbiu7vFIoqt7fLNDqacLgcdyj+cDV0XEYcAW4JxUfg6wJZVfldZD0hHATOBIYDrwtdRwHAF8FTgVOAI4K61rBVCq2w899HCzQzEbkK7u8axevbrd98dWEK5v1qpKv++uv/1zI6/Dbd64gQlzl7F544Zmh1IXkrqB3wO+nh4L+CBwa1plEfDhdH9GekxafkJafwawJCJei4gngV7guHTrjYh1EbENWJLWtQIo1e3XX9/W7FDMBmTzxg2MGnNY2+6PrVhc36xVlX7fXX/717JTKJgN0D8AnwPelh4fBLwQEdvT4z5gbLo/FtgAEBHbJb2Y1h8LrMq9Zv45G8rKp1QKQtK5wLkAo0ePpqenZ0DBb926dcDrNktRY1ywYAGjxmznHSO7CxlfSVHfPzMzM2tdbuRZ25J0OvBsRKyWNLWZsUTEdcB1AJMnT46BzpHSCnPAFDXGadOmMWHuMi44pI+PfexjzQ6nqqK+f2bWIUbshSTGjB3Hpr6nmx2NmdWIu2taO3s/8CFJT5F1pfwgcDWwv6TSCY5uYGO6vxEYB5CWvx34Rb687DnVys2GTdJCSc9KeiRX9gVJGyU9kG6n5ZYNKjlQtQREZta+SmOZspEIyY7X23rYhlmnciPP2lZEXBwR3RExkSxxyo8j4g+Bu4Az0mqzgR+k+0vTY9LyH0dEpPKZKfvmocAk4B7gXmBSOlgelf7H0gZsmnWGG8kS/ZS7KiKOTrflMOTkQNUSEJlZmyqNZZowd1mzQzGzOnMjzzrRXOCzknrJxtzdkMpvAA5K5Z8F5gFExBrgFuBR4HbgvIjYkcb1nQ+sIMveeUta12zYIuInwPMDXH1QyYH2kIDIGqjilRUzM7Nh8pg86wgR0QP0pPvryA5+y9f5FfDRKs+/HLi8QvlyYHkNQzXbk/MlzQLuAy6KiC0MPjlQfwmIdjPUxEHN0GqJbOb8+YWMGnMYANs29zJ6nyxpUCttg1m7kfTnwB8DATwMfALoIjtRdhCwGjg7IrZJ2htYDBxLNsTjYxHxVHqdi8l6SewA/iwiVqTy6WTDR0YAX4+IKxq3ddYp3MgzM2sd1wKXkR14XAZcCXyy3v90qImDmqHVEtmUEgQBrJ8/h2u++X3mzJlD1lPczBpN0ljgz4AjIuJVSbeQdYc/jayL+xJJ/0TWeLuW3By7kmaSdYX/WFk3+t8AfiTpnenffBU4iewE272SlkbEow3cTOsA7q5pZtYiIuKZ1FX418D1vHFFerDJgX5B9QRE1mwp26EkT/Rr1hwjgX3SPvItwCY8x661GF/J60Bd3ePZvHEDCxYsaHYoZjYIkroiYlN6+BGglHlzKfBtSV8mO2NcSg4kUnIgskbcTOAPIiIklRIQLWHXBETWbCnbIcD6+ac3ORizzhIRGyUtAJ4GXgXuIOue2TJz7A5FK3V1L82D24rd2xv5PruR14FK2bVGjdm+55XNrCkk3QRMBQ6W1AdcAkyVdDRZd82ngD+BLDlQ6lL0KLCdlBwovU4pOdAIYGEuOdBcYImkLwI/440ERGZmHUvSAWRX1g4FXgD+hcqZjuuukV3lW6mre6mb+0Xv3t5y3dsb+T67kWdmVkARcVaF4qoNscEmB6qWgMjMrMOdCDwZEc8BSPou2by7+0sama7mVZpjt2+Ac+zST7lZzXhMnpmZmZlZ5mngeElvSWPrTiDrJeE5dq2l+EqemZmZmRkQEXdLuhW4n6z7+8/Iukz+kMpd3G8AvpHm2H2erNE21G70ZjXjRp6ZmZmZWRIRl5CNg87zHLvWUtxd08zMzMzMrI24kWdmZmZmZtZG3MgzMzMzMzNrI27kmZmZmZmZtRE38szMzMzMzNqIG3lmZmZmZmZtxI08MzMzqylJCyU9K+mRXNmBklZKWpv+HpDKJekaSb2SHpJ0TO45s9P6ayXNzpUfK+nh9Jxr0qTVZmaW7LGRJ2mcpLskPSppjaQLU7l31mZmZlbJjcD0srJ5wJ0RMQm4Mz0GOBWYlG7nAtdCdpxBNlfZFLL5yS4pHWukdT6Ve175/7LBGrEXkpBEV/f4ZkdjZsM0kCt524GLIuII4HjgPElH4J21tQBJb5Z0j6QH00mKv0nlh0q6O51YuFnSqFS+d3rcm5ZPzL3Wxan8CUmn5Mqnp7JeSfPKYzAzK9fVPZ52Pp8ZET8Bni8rngEsSvcXAR/OlS+OzCpgf0ldwCnAyoh4PiK2ACuB6WnZfhGxKiICWJx7LRuqHa8zYe4yJsxdxuaNG5odjZkN08g9rRARm4BN6f7Lkh4DxpLtlKem1RYBPcBccjtrYJWk0s56KmlnDSCptLPuIe2sU3lpZ31bbTbROtxrwAcjYqukvYCfSroN+CxwVUQskfRPwDlkJxvOAbZExGGSZgLzgY+lExszgSOB3wB+JOmd6X98FTgJ6APulbQ0Ih5t5EaaWWvZvHEDE+YuY/3805sdSiONTscUAJuB0en+WCDfquhLZf2V91Uo342kc8lOODN69Gh6enp2Wb5169bdytrZggULGDVmOwDbFixg9D5vlG0rW9YJ70unff7WWfbYyMtLVzXeB9xNAXfWtdTOX/zSDr18596O25tONmxND/dKtwA+CPxBKl8EfIGskTcj3Qe4FfjH1H14BrAkIl4DnpTUS3ZFGqA3ItYBSFqS1nUjz8ysiogISdGA/3MdcB3A5MmTY+rUqbss7+npobysHXV1j995dW7C3GUArJ8/h2u++X3mzJmTTjjM2WVZ9vPZ3jrl87fONOBGnqR9ge8An4mIl/LdTIqys66ldv7iT5s2jQlzl3HRu7fvsnNv1x26pBHAauAwsqtuPwdeiIjtaZX8iYWdJyMiYrukF4GDUvmq3Mvmn1N+8mJKhRiGdIKiFU42FDXG0gmMd4zs5sorrwRgr71G8Z73vLvJke2qqO+fWR08I6krIjalHj7PpvKNwLjcet2pbCNv9Bgqlfek8u4K61sVHXrl2KyjDaiRl7q5fQf4VkR8NxV7Z20tISJ2AEdL2h/4HvCuJsQwpBMUrXCyoagxlk5mXHBIH195LvvI188/vXAnM4r6/pnVwVJgNnBF+vuDXPn5qSfEFODFdGyxAvhSbvz+ycDFEfG8pJckHU/Ws2gW8JVGboiZWdENJLumgBuAxyLiy7lFpZ017L6znpWybB5P2lkDK4CTJR2QdtgnAyvSspckHZ/+16zca5nVTES8ANwF/DbZwP7SSY78iYWdJynS8rcDv6D/kxeVys3MOpakm4D/BA6X1CfpHLLG3UmS1gInpscAy4F1QC9wPfBpgDSG/zLg3nS7tDSuP63z9fScn+Nx/GZmuxjIlbz3A2cDD0t6IJV9nmznfEvaca8HzkzLlgOnke14XwE+AdnOWlJpZw2776xvBPYh21F7Z201IekQ4PWIeEHSPmQJUuaTNfbOAJaw+0mK2WQHJ2cAP07dkZcC35b0ZbLEK5OAewABkyQdSta4m8kbY/3MzDpSRJxVZdEJFdYN4Lwqr7MQWFih/D7gqOHEaFZN6vnzdbI6FsAngSeAm4GJwFPAmRGxJV2guJrs2PcV4OMRcX96ndnAX6WX/WJELErlx/LGce9y4MIoWjcTa3kDya75U7ID2Uq8s7ai6wIWpXF5bwJuiYhlkh4Flkj6IvAzsqvVpL/fSIlVnidrtBERayTdQpZQZTtwXuoGiqTzya5UjwAWRsSaxm2embW1NHfZmLHj2NT3dLOjMesUVwO3R8QZaYqlt5Bd4LgzIq5I0yXNI8sqn586bApZErcpuanDJpM1FFen7NtbeGPqsLvJGnnT8QUOq7FBZdc0azUR8RBZRtjy8nW8kR0zX/4r4KNVXuty4PIK5cvJdtJmZrWV5i5zwgyzxpD0duADwMcBImIbsE2Spw6zluJGnpmZmZlZ5lDgOeCfJb2XLDv3hXjqsMIonwqsVeKGxr7PbuSZmZmZmWVGAscAF0TE3ZKuJuuauZOnDmuu8qnAWmk4YyPf5z1m1zQzMzMz6xB9QF9E3J0e30rW6HsmdcNkEFOHVSv31GFWd27kmZkVkKSFkp6V9Eiu7EBJKyWtTX8PSOWSdI2kXkkPSTom95zZaf21KdNbqfxYSQ+n51yTMsSZmXW0iNgMbJB0eCo6gSzpmqcOs5biRp6ZWTHdSJZxLW8eWXa3ScCdvNGFKJ/d7VyyzG3ksrtNIUs0dEluYulSdrfS88r/l5lZp7oA+Jakh4CjgS/heR6txXhMnplZAUXETyRNLCt2djczszqLiAfIpj4o56nDrGW4kWdm1joant0NGpvhbbhaJUNcKTvctvQXYNuCBTuzxeXLSuu1wnZZm/D8jGYtz428DtLVPZ7NGzdUXph26IB36mYtoFHZ3dL/aliGt+FqlQxxpexw6+fPYcLcZQCsnz+Ha775febM2bWstF4rZZCzFuf5Gc1ansfkdZDNGzfsPHDYTdqhT5i7rHpD0MyazdndzMzMbI/cyDMzax3O7mZmZmZ75O6aZmYFJOkmssQpB0vqI8uSeQVwi6RzgPXAmWn15cBpZJnaXgE+AVl2N0ml7G6we3a3G4F9yBKuOOmKmZlZm3Ajz8ysgCLirCqLnN3NzMzM+uXummbWNrq6xyNpZxIhsyJx/bRGcn0z62xu5JlZ2yglF6qaYMisiVw/rZFc38w6mxt5ZmZmZmZmbcSNPDMzMzMzszbiRp61LUnjJN0l6VFJayRdmMoPlLRS0tr094BULknXSOqV9JCkY3KvNTutv1bS7Fz5sZIeTs+5Rh78YGZmZmZN5kaetbPtwEURcQRwPHCepCOAecCdETEJuDM9BjgVmJRu5wLXQtYoJEtfPwU4Drik1DBM63wq97zpDdguMzMzM7Oq3MizthURmyLi/nT/ZeAxYCwwA1iUVlsEfDjdnwEsjswqYH9JXcApwMqIeD4itgArgelp2X4RsSqlsOXI5rYAACAASURBVF+cey0zs9oZsdfOTIld3eObHY2ZmRWc58mzjiBpIvA+4G5gdERsSos2A6PT/bHAhtzT+lJZf+V9FcrNzGprx+s7sySun396k4MxM7OicyPP2p6kfYHvAJ+JiJfyw+YiIiRFA2I4l6wLKKNH///s3X24XFV99//3xwSQ8lB5DGlOINwS9QYUlPwI/eFtE1EISI3eF/LQlgSh0l4CRUvagG0vrKg32EQKVGlRIqEqkaLUFIEYKflR76tBEkSeApIGAjlNiALyIEpI/P7+2GuSncmcx5nZe8/M53Vd5zoza++Z+e6ZdfbZa9Za3zWOZcuWDetxr7zyyrD3LUuVYpw3bx47H7AZgE3p9v5j+7goV1aVWGuq9P6ZmVlG0hhgBdAfESdLOhhYBOwDrATOjIhNknYhG8lzFPAccFpEPJWe4xLgHGAL8GcRsSSVzwCuAsYAX42Iyws9OOsJbuRZV5O0E1kD7xsR8Z1U/Kyk8RGxPg253JjK+4GJuYf3pbJ+YFpd+bJU3tdg/x1ExHXAdQBTpkyJadOmNdptB8uWLWO4+5alSjFOnz4919sxh4Pm3sYF+63jmp+9bWtZNrK2Oqr0/pmZ2VYXkk3z2DPdvwK4MiIWSfpHssbbten3CxFxiKTT036npRwApwOHAb8D/EDSW9JzfQl4P9kIoPskLY6IR4s6MOsNnpNnXStlurweWBURX8xtWgzUMmTOBr6bK5+VsmweA7yYhnUuAY6XtFdKuHI8sCRte0nSMem1ZuWey8zMzDqQpD7gA8BX030B7wVuSbvUz+evzfO/BTgu7T8TWBQRr0XEk8BqsuRtRwOrI2JNRGwi6x2c2f6jsl7jnjzrZscCZwIPSXoglX0KuBy4WdI5wFrg1LTtduAkshPxq8BHASLieUmXAfel/T4TEc+n2x8HbgB2Be5IP2ZmZp0vJfwBOGDCRNave7rkgArz98BfAnuk+/sAv4iIzel+fg7+1nn7EbFZ0otp/wnA8txz5h9TP89/aqMgRjvVYzQ6aepAbWrGuF2z250SNxT7PruRZ10rIn4IDLRu3XEN9g/gvAGeawGwoEH5CuDwJsI0MzOrph5M+CPpZGBjRKyUNK3MWEY71WM0OmnqQG1qxkVv38ycOdWbhjGYIt9nN/LMzMzMzDLHAh+UdBLwRrI5eVeRLas0NvXm5efg1+bzr5M0FvhtsgQsA83zZ5Bys5bxnDwzMzMzMyAiLomIvoiYRJY45d8j4g+Bu4FT0m718/lr8/xPSftHKj9d0i4pM+dk4EdkUz8mSzpY0s7pNRYXcGjWY9yTZ2ZmZmY2uLnAIkmfBX5MltiN9PufJa0GnidrtBERj0i6GXgU2AycFxFbACSdT5bUbQywICIeKfRIrCcM2ZMnaYGkjZIezpXtLWmppCfS771SuSRdLWm1pAclvSv3mNlp/yckzc6VHyXpofSYq5VfxMzMzMy6iqSn0v/9ByStSGUtu64wa5WIWBYRJ6fbayLi6Ig4JCI+EhGvpfJfp/uHpO1rco//XES8OSLeGhF35Mpvj4i3pG2fK/7IrBcMZ7jmDcCMurKLgbsiYjJwV7oPcCJZd/RksmxA10J28gYuJcsedDRwae0Envb5WO5x9a9lZmZm3WV6RBwZEVPS/VZeV/S08X0H4u/LzWzIRl5E3EPW/ZyXXxOkfq2QGyOznGyS6njgBGBpRDwfES8AS4EZadueEbE8jV++MfdcZmZm1htacl1RdNBVtKH/ma0ZMc2sd412Tt64tBA0wAZgXLq9da2QpLYmyGDl6xqUm5mZWXcK4PuSAvinlCa+VdcV2xlqnbFOWhtsuGpriG1Kv4Gtt+vLauuMDXf/bnuvuvHzN6tpOvFKREQ6UbedF4VsTv2JfLCTe7cdu5mZVca7I6Jf0v7AUkmP5Te28rpiqHXGOmltsOGqrSG29oo5uTXu5jQsu/rr/8qcOY23NSrrpPXIhqMbP3+zmtE28p6VND4i1qdhExtT+UBrgvQD0+rKl6Xyvgb7N+RFIZtTf+KvLSLZCydyMzOrhojoT783SrqVbE5dq64rzMyM0a+Tl18TpH6tkFkpG9YxwItp+MUS4HhJe6WJ0ccDS9K2lyQdk7Jqzso9l5mZmXURSbtJ2qN2m+x64GFadF1R4KGYmVXakD15km4i+7ZsX0nryLJZXQ7cLOkcYC1watr9duAkYDXwKvBRgIh4XtJlZAtAAnwmImrJXD5OlsFzV+CO9GNmZgOQ9BTwMrAF2BwRU1K2wW8Bk4CngFMj4oX0BdpVZOfmV4GzIuL+9Dyzgb9OT/vZiFiIWXuNA25N2R/HAt+MiDsl3UfrrivMzHrekI28iDhjgE3HNdg3gPMGeJ4FwIIG5SuAw4eKw8zMtjM9In6eu19LQX+5pIvT/blsn4J+KlkK+qm5FPRTyBJhrJS0OGUqtBYa33cgG/qfGXrHHpDWEDuiQflztOi6wtpkzE5I4oAJE1m/7umyozGzIYx2uKZ1iNp6OV4zx6zrOQV9RTmlvXWFLa9z0Nzb/IWFWYdoOrumVVv+4mLtFSeXHI2ZtUhhKeih2MzGzapiZuTRprQfeL/5zJ8/n5122pl3vOPt5RyUmZlVmht5Zmadp7AU9On5Csts3KwqZkYebUr7gfe7KJWd7EzIBnhIsJntyMM1zcw6TD4FPbBdCnqAEaSgb1RuZh2mNmrHw4LNrMaNPDOzDuIU9GZmZjYUN/JsRymD1vi+A8uOpGmSFkjaKOnhXNnekpZKeiL93iuVS9LVklZLelDSu3KPmZ32fyKlna+VHyXpofSYq+UMN9XWHXV7HPBDST8BfgR8LyLuJFva5v2SngDel+5DloJ+DVkK+q+QLVtDSjdfS0F/H05Bb2Zm1jXcyLMddVcGrRvYMWNgLdX8ZOCudB+2TzV/LlmqeXKp5qeSDYu7tNYwTPt8LPc4Zyessi6o2xGxJiKOSD+HRcTnUvlzEXFcREyOiPfVGmwpq+Z5EfHmiHh7Wram9lwLIuKQ9PO1so7JzMzMWsuNPOtqEXEPUN870ZJU82nbnhGxPK3ldGPuuczMzKzDSJoo6W5Jj0p6RNKFqdyjgKyjOLum9aJWpZqfkG7Xl+9gtCnoq5gOvl6VYmyUen7/sX1c1CAdfVVirtL7Z2ZmbAYuioj70/znlZKWAmeRjQK6XNLFZKOA5rL9KKCpZCN8puZGAU0hW/ZmpaTF6cvi2iige8mG1M8A7ijwGK0HuJFnPa3VqeYHeZ1RpaCvYjr4elWKsZaqHralnr9gv3Vc87O3bVe29oo5lUk9X6X3z8xsSGluM8ABEyayft3TJQfUWulL4PXp9suSVpF9gTsTmJZ2WwgsI2vkbR0FBCyXVBsFNI00CgggNRRnSFpGGgWUymujgNzIs5ZyI8960bOSxkfE+hGkmp9WV74slfc12N/MzKw7pbnNAGuvOLnkYNpL0iTgnWQ9bh0zCmg0OmlUSW3Uzrhds9udEjcU+z67kWe9qJZq/nJ2TDV/vqRFZEMuXkwNwSXA53PJVo4HLomI5yW9lNLS3wvMAq4p8kDMzKx3eRH09pG0O/Bt4BMR8VJ+2lzVRwGNRieNKqmN2rno7ZuZM/cS2PJ6x/QqF/k+O/GKdTVJNwH/CbxV0jpJ59DaVPMfB76aHvNfeLiFmRWlO5YEsSbUFkG31pK0E1kD7xsR8Z1U/Gwa/cMIRgENVO5RQK3SBVmz28U9edbVIuKMATYd12DfAM4b4HkWAAsalK8ADm8mRjPrPoX0sKSLm24fMmdWpJTp8npgVUR8MbfJo4Cso7iRZ2Zm1mL5HhY3wsw6yrHAmcBDkh5IZZ8ia9zdnEYErQVOTdtuB04iG9HzKvBRyEYBSaqNAoIdRwHdAOxKNgLIo4Cs5dzIMzMzM+sQnofXXhHxQ2Cgdes8Csg6hufkdanxfQcir61pZmbWVWq9xJ6LZ2aDcSOvS3kytpmZmbWVk/+YVZYbeWZmZmY2cs5saFZZbuTZwNI3dP6WzsyswnyuNjOzOk68YgNL39CBs8OZmVWWz9VmZlbHPXlmZmYt4qRX1i6uW2Y2Em7kmZmZtYiTXlm7VLpueciwWeV4uKaZmZmZjZ6HDJtVjnvyuoyHc5iZFa8y516ntDczM9zI6zqVHs5hVhW+ELYWq8y51yntu0rty4NKfIFgZh3FjTwz63gj7kXxhbCZdYDalweV+ALBzDqKG3lm1vEq04tiPaXSvSxOhGFl8UgJs0pwI68LVPpCw8ysS1W6lyX1VrvHujNVZo7naHikhFkluJHXBQq50PA3c2ZmZm1Va9x1xegE9yablaoyjTxJMyQ9Lmm1pIvLjsfq+Ju5AbnudjBfhLj+jkJH97J0kW6su13RuKvJ9yZv2NDT59lGurH+WrVUopEnaQzwJeBE4FDgDEmHlhtV9flCo3yuux2ux4e0uf6OTsddiHfhSIxuqrs9MeWi9kVxauyN3WXXnv6CrZvqb5EGve7twvNcsyrRyAOOBlZHxJqI2AQsAmaWHFMl5f8ZlHKh4Z6Peq673aI3/0G4/g5TR1+I111gd0k978i6m69HtYZOped2tlqqi1s2/XqHXr58w692uwvq6UA6sv6WbdDr3u48zzVFEVF2DEg6BZgREX+c7p8JTI2I8+v2Oxc4N919K/B4G8PaF/h5G5+/Crr1GA+KiP2KeKEC6m4nfEZVj7GT4ius7kJlz73NqvrnPZROjr8Tz72d/H63go8/O36fezuzLjjmTMP6O7bFL9JWEXEdcF0RryVpRURMKeK1ytILx1gVo627nfAZVT1Gx9e8Is+9zeqE93MwnR5/1QxVd3v9/fbxV/v4fd07OMc8uKoM1+wHJubu96Uys6pz3bVO5vprncp11zqZ66+1XVUaefcBkyUdLGln4HRgcckxmQ2H6651Mtdf61Suu9bJXH+t7SoxXDMiNks6H1gCjAEWRMQjJYfVEUOTmtQLx9hWBdTdTviMqh6j4xtARc+9zar65z2UTo+/EC2su73+fvv4S1DRc28n1gXHPIhKJF4xMzMzMzOz1qjKcE0zMzMzMzNrATfyzMzMzMzMuogbeXUkfUTSI5J+I2lK3bZLJK2W9LikE8qKsRUkzUjHsVrSxWXHY9uT9HeSHpP0oKRbJb0pt630etgJfydVrOOSFkjaKOnhXNnekpZKeiL93qvMGLuBpE9L6pf0QPo5qeyYhlLF+torOrG+NKvX65ukpyQ9lD7vFWXHU5aqX2sMphPqsKSJku6W9Gi6ZrowlRfyf9+NvB09DPxv4J58oaRDybIfHQbMAL4saUzx4TUvxf0l4ETgUOCMdHxWHUuBwyPiHcBPgUugUvWw0n8nFa7jN5C9L3kXA3dFxGTgrnTfmndlRByZfm4vO5jBVLi+9pKOqS/Ncn3banr6vDtqnbUWq/q1RkMdVIc3AxdFxKHAMcB5Kc5C/u+7kVcnIlZFxOMNNs0EFkXEaxHxJLAaOLrY6FrmaGB1RKyJiE3AIrLjs4qIiO9HxOZ0dznZGjpQkXrYAX8nlazjEXEP8Hxd8UxgYbq9EPhQoUFZFVSyvlrXcn0zoPrXGoPoiDocEesj4v50+2VgFTCBgv7vu5E3fBOAZ3L316WyTtRNx9ILzgbuSLer/tlVJb6qxDEc4yJifbq9ARhXZjBd5Pw0BGlBBwyB7aT62q06qb40y/UNAvi+pJWSzi07mIrwtUYbSZoEvBO4l4L+71dinbyiSfoBcECDTX8VEd8tOh7rTcOph5L+iqy7/xtFxpZe238nBYuIkOR1bYZhsPoJXAtcRnYhdxkwn+wCxnqU64vVeXdE9EvaH1gq6bE00qLrVP1aoxdI2h34NvCJiHhJ0tZt7fy/35ONvIh43yge1g9MzN3vS2WdqJuOpWMNVQ8lnQWcDBwX2xa0LOyz6/C/k6rEMRzPShofEesljQc2lh1QJxhu/ZT0FeC2NofTrE6qrx2py+pLs3q+vkVEf/q9UdKtZMP/urKRV/VrjVGqenxbSdqJrIH3jYj4Tiou5P++h2sO32LgdEm7SDoYmAz8qOSYRus+YLKkgyXtTDa5dnHJMVmOpBnAXwIfjIhXc5uqXg+rEl8n1fHFwOx0ezbgXtImpX+aNR8mSxRUZZ1UX7tOB9aXZvV0fZO0m6Q9areB4+n+z7yhDr7W6Ig6rKzL7npgVUR8MbepkP/7PdmTNxhJHwauAfYDvifpgYg4ISIekXQz8ChZl/Z5EbGlzFhHKyI2SzofWAKMARZExCMlh2Xb+wdgF7JhJADLI+JPq1IPq/53UtU6LukmYBqwr6R1wKXA5cDNks4B1gKnlhdh1/iCpCPJht89BfxJueEMrqr1tYd0VH1plusb44Bb0//WscA3I+LOckMqTaWvNQbSQXX4WOBM4CFJD6SyT1HQ/31t65k1MzMzMzOzTufhmmZmZmZmZl3EjTwzMzMzM7Mu4kZeBUm6U9JnGpTPlLRB0l9IeljSy5KelPQXdftdJukhSZslfbqwwM1orv5K2l/STZL+W9KLkv6vpKnFHoH1qhace++W9DNJL0n6iaTKLc5r3avZ+pvb//ckhaTPtj9qs5ace5+S9CtJr6Sf7xcXfXW5kVdNC4E/Sll58s4kW8NEwCxgL2AG2SKup+f2W02WLel7BcRqVq+Z+rs7Wdaso4C903N9T9kaM2bt1uy590JgfETsCZwLfL0uc6NZOzVbf2vp3q8iW7DZrChN113g9yNi9/RzfNsj7gBOvFJBknYFNpBV2HtS2V7AemBqRPykbv+ryT7LC+rKvw6sjohPFxK4Ga2rv7ntLwHTI2JleyO3XtfKuiuptu7WeyKiSqnHrUu1ov5KupjsC7b9gXUR8ddFxW+9q9m6K+kp4I8j4geFBl5x7smroIj4FXAz2bcWNacCjzWo6AL+F1DF1LHWg1pZf1Na853JeqfN2qoVdVfSbZJ+TdYTsgxY0c6YzWqarb+SDgLOBnYYNmfWTi26bvhGGi7/fUlHtDXgDuFGXnUtBE6R9MZ0f1Yqq/dpss/xawXFZTYcTddfSXsC/wz8bUS82KY4zeo1VXcj4mRgD+Ak4PsR8Zv2hWq2g2bq79XA30TEK22N0KyxZuruHwKTgIOAu4Elkt7Utkg7hBt5FRURPwR+DnxI0puBo4Fv5vdRthDkLOADEfFa8VGaNdZs/U1DN/6NbGHW/1NM1GatOfdGxOsRcQdwvKQPFhC2GTD6+ivp94E9IuJbBYdsBjR37o2I/xsRv4qIV9M1wy/Ievt62tiyA7BB3UhWmd8KLImIZ2sbJJ0NXEw232NdSfGZDWZU9VfSLsC/AuuAPykuXLOtWnXuHQu8uW1RmjU2mvp7HDBF0oZ0/7eBLZLeHhHOEmtFadW5N8iStfQ0J16pMEmTgJ8CG4FPRsS/pPI/BOaTJaNY1eBxOwFjgAXAGuCzwOsRsaWYyM1GV39T3f0OsAU4JSI2FxmzGYy67r4NOJhsHt5m4DSyc/AxEXF/UbGbjbL+7gHsliu6Cvhv4LKIeL6AsM1GW3cPBCaSZeZ+A3ABWYb5t0XEc4UFX0Fu5FWcpGXAEcABuWEVTwJ9QH6Y0Ncj4k/T9huA2XVP9dGIuKHd8ZrljbT+Svo9sovkXwH5uUwnRsR/FBK0GaOqu/8TuAE4lOxLiieAz0fErUXGbQaju3aoe/wNOLumlWAU597DgJvIRk38GngAmBsRPZ/0yo08MzMzMzOzLuLEK2ZmZmZmZl3EjTzrWpImSrpb0qOSHpF0YSrfW9JSSU+k33ulckm6WtJqSQ9KelfuuWan/Z+QNDtXfpSkh9Jjrk7rt5iZmZmZlcaNPOtmm4GLIuJQ4BjgPEmHkmVnuisiJgN3pfsAJwKT08+5wLWQNQqBS4GpZCl9L601DNM+H8s9bkYBx2VmZmZmNiA38qxrRcT6Wla7iHgZWAVMAGaybYHNhcCH0u2ZwI2RWQ68SdJ44ARgaUQ8HxEvAEuBGWnbnhGxPLLJrTfmnsvMzMzMrBQdu07evvvuG5MmTRr2/r/85S/Zbbfdht6x4rrhONp9DCtXrvx5ROyXL0tped8J3AuMi4j1adMGYFy6PQF4JvewdalssPJ1Dcp3IOlcst5Bdt1116MmTpw40sNqud/85je84Q3V+J6nSrFAefH89Kc/3aHuVs1Iz701VTx3VS2mqsUDI4up0bm3SkZbdwdTtc/M8QxuoHiqXndh6Ppblfe6KnFAdWIp47oXOriRN2nSJFasGH521GXLljFt2rT2BVSQbjiOdh+DpLV193cHvg18IiJeyk+bi4iQ1PYUsxFxHXAdwJQpU2IkdbddqlSXqhQLlBdPfd2topGee2uq9hlD9WKqWjwwspiqXn9HW3cHU7XPzPEMbqB4ql53Yej6W5X3uipxQHViKfq6t6Y6X52btUFaXPvbwDci4jup+Nk01JL0e2Mq7ydbULOmL5UNVt7XoNzMzMzMrDRu5FnXSpkurwdWRcQXc5sWs22x+NnAd3Pls1KWzWOAF9OwziXA8ZL2SglXjgeWpG0vSTomvdas3HOZmZmZmZXCjTzrZscCZwLvlfRA+jkJuBx4v6QngPel+wC3A2uA1cBXgI8DRMTzwGXAfennM6mMtM9X02P+C7ijiAOz7nH22Wez//77c/jhh28tK2KZj4Few8ysl0l6o6QfSfpJWn7pb1P5wZLuTefSb0naOZXvku6vTtsn5Z7rklT+uKQTcuUzUtlqSRfXx2DWCm7kdYjxfQciiQcffKjsUDpGRPwwIhQR74iII9PP7RHxXEQcFxGTI+J9tQZbyqp5XkS8OSLeHhErcs+1ICIOST9fy5WviIjD02POT1k2bRRqdXx834Flh1Kos846izvvvLO+uIhlPgZ6DSvJgw8+1JN/A9a82vlz5cqVrj/New14b0QcARxJlk37GOAK4MqIOAR4ATgn7X8O8EIqvzLtR1qy6XTgMLLz7pcljZE0BvgS2fn8UOCMtK+1WK+fU93I6xAb+p/hoLm38frrm8oOxaylxvcdyMqVK7fW8Q39zwz9oC7ynve8h7333ru+uIhlPgZ6DSvJ669v6sm/AWte7fy58wGHuP40KZ1fX0l3d0o/AbwXuCWV15+Xa+fSW4Dj0oiJmcCiiHgtIp4kG/FzdPpZHRFrImITsCjtay1Q+8JDUs+fUzs2u6aZdYcN/c+w8wGHlB1G1RSxzMdAr7GD/BIg48aNY9myZSM8HHjllVdG9bh2qlpMfX19XHTAZjbNm1eZuKr2HpkVIfW2rQQOIet1+y/gFxGxOe2SP5duPf9GxGZJLwL7pPLluafNP6b+fD11gDiGfe6tyt9q2XHM+eSFW68p9h/760qcU8t6T9zIMzOrsCKW+RjqNeqXABlNKuiqpLLOq1pM8+fP55qfvY21V8yhKiO/q/YemRUhIrYAR0p6E3Ar8LaS4hj2ubcqf6tlxzF9+nQOmnsbABfst64S59Sy3hMP1zQzq54ilvkY6DXMzAyIiF8AdwO/SzY8vtY5kj+Xbj3/pu2/DTzHyM/XZi01rEaepE+mDEMPS7opZR5yliEzs/YoYpmPgV7DrGmSFkjaKOnhXNmnJfXXZTuubRvR9cFA1yBmzZK0X+rBQ9KuwPuBVWSNvVPSbvXn5dq59BTg39Mc6MXA6em6+GCyxFc/IsvSPTnV4Z3JkrMsbv+RWa8ZspEnaQLwZ8CUiDgcGENWIZ1lyMysSWeccQa/+7u/y+OPPw7wDknnUMwyHwO9hlkr3MC2TK55V+azHcOorw8GugYxa9Z44G5JD5KdT5dGxG3AXODPJa0mm3N3fdr/emCfVP7npEzFEfEIcDPwKHAncF5EbEnz+s4n+3JuFXBz2tespYY7J28ssKuk14HfAtaTZRn6g7R9IfBpslTdM9NtyLIM/UN9liHgyfTHcHTab3VErAGQVMsy9OjoD8vMqmx834E9m+2q3k033bT1tqQHI6J24XBc/b7p2+HzGj1PRCwAFjQoXwEc3qD8uUavYdYKEXFPfiTPEEZ0fSBpFQNfg5g1JSIeBN7ZoHwN2+plvvzXwEcGeK7PAZ9rUH472Zd2Zm0zZCMvIvolzQOeBn4FfJ8s41ClswzVKzvbT7PmzZvHzgdsZv+xfR19HND5n4U1r5buG2DtFSeXHI2ZFeh8SbOAFcBFacmPkV4f7MPA1yDbaUVm2MFU5f9Z7Rph3K7Z7SrEBNV5f2qqFo9ZOw3ZyEvzO2YCBwO/AP6FxkMw2q6ZDG9lZ/tpVi1b0AX7reO0004rO5ymdPpnYWZmo3It2bDiSL/nA2e38wVbkRl2MFX5f1a7Rrjo7ZuZM8fZWQdStXjM2mk4wzXfBzwZET8DkPQd4FhSlqH0TVqjLEPrhplliEHKzczMrAtExLO125K+AtyW7o70+uA5Br4GMTMzhpdd82ngGEm/lebWHUc2X85ZhszMzGxYakt2JB8Gapk3R3R9kK4pBroGMTMzhjcn715JtwD3A5uBH5MNffgesEjSZ1NZPsvQP6eJ08+TnZSJiEck1bIMbSZlGQKQVMsyNAZY4CxDZmZmnUvSTcA0YF9J64BLgWmSjiQbrvkU8Ccw6uuDuTS+BjEzM4aZXTMiLiU7Qec5y5CZmZntICLOaFA8YENspNcHA2U6NDOzzLAWQzczMzMzM7PO4EaemVXLmJ2QhCTG9x1YdjRmZmZmHWe4i6GbmTVtWIugb3nda+iZmZmZNcE9eWZWmPwi6GZmZmbWHm7kmZmZmZmZdRE38qyrSVogaaOkh3Nln5bUL+mB9HNSbtslklZLelzSCbnyGalstaSLc+UHS7o3lX8rreVkZmZmZlYaN/Ks290AzGhQfmVEHJl+bgeQdCjZuo6Hpcd8WdIYSWOALwEnAocCZ6R9Aa5Iz3UI8AJwTluPxszMzMxsCG7kWVeLiHuA54e5+0xgUUS8FhFPAqvJ1mE6GlgdI1zx2AAAIABJREFUEWsiYhOwCJgpScB7gVvS4xcCH2rpAZiZmZmZjZCza1qvOl/SLGAFcFFEvABMAJbn9lmXygCeqSufCuwD/CIiNjfYfzuSzgXOBRg3bhzLli1r0WGM3iuvvFJ4HPPmzWPnAzazKf0G2DRvHuN2HXhbGe9VGe+NmZmZWau4kWe96FrgMiDS7/nA2e18wYi4DrgOYMqUKTFt2rR2vtywLFu2jKLjmD59OgfNvY21V8zJLZMwh6u//q/MmTOn4baIKDRGKOe9MTMzM2sVD9e0nhMRz0bEloj4DfAVsuGYAP3AxNyufalsoPLngDdJGltXbtY0SW/NJQd6QNJLkj7hxEFmZmY2FDfyrOdIGp+7+2GglnlzMXC6pF0kHQxMBn4E3AdMThfEO5MlZ1kcWRfT3cAp6fGzge8WcQzW/SLi8VpyIOAo4FXg1rTZiYPMzMxsQG7kWVeTdBPwn8BbJa2TdA7wBUkPSXoQmA58EiAiHgFuBh4F7gTOSz1+m4HzgSXAKuDmtC/AXODPJa0mm6N3fYGHZ73jOOC/ImLtIPs4cZCZmZkBnpNnXS4izmhQPGBDLCI+B3yuQfntwO0NytewbbinWbucDtyUu19o4iAzMzPrLG7kmZlVWBoi/EHgklRUeOKgVmSHrWLG0qrF1NfXx0Upw2xV4qrae2RmZsMzrEaepDcBXwUOJ7uwOBt4HPgWMAl4Cjg1Il5IQ4CuAk4im0NyVkTcn55nNvDX6Wk/GxELU/lRZItW70rWW3JhlJFSz8ysek4E7o+IZyFLHFTbIOkrwG3p7kAJghigfGvioNSbN2DioFZkh61ixtKqxTR//nyu+dnbSssq20jV3iMzMxue4c7Juwq4MyLeBhxBNi/pYuCuiJgM3JXuQ3ZBMjn9nEv2rTOS9gYuJRsmdDRwqaS90mOuBT6We9yM5g7LzKxrnEFuqKYTB5mZmdlQhmzkSfpt4D2keUwRsSkifkE2yX9h2i0/YX8mcGNklpN9UzweOAFYGhHPp/kjS4EZadueEbE8XXTciCf/m5khaTfg/cB3csVOHGRmZmaDGs5wzYOBnwFfk3QEsBK4EBgXEevTPhuAcen2BHac5D9hiPJ1Dcp30My8kE6fVzBv3jx2PmAz+4/t6+jjgM7/LMyKEhG/JGt85cvOHGR/Jw4yM2uCpIlkHQ7jyKYoXRcRV6URaZ6mZB1jOI28scC7gAsi4l5JV7FtaCYAERGS2l45m5kX0unzCqZPn85Bc2/jgv3Wcdppp5UdTlM6/bMwMzOrrDE7IYkDJkxk/bqny46mE20my1p8v6Q9gJWSlgJnkU1TulzSxWTXwnPZfprSVLIpSFNz05SmkDUWV0panEaz1aYp3UvWyJsB3FHgMVoPGM6cvHXAuoi4N92/hazR92xtbkj6vTFtH2jy/2DlfQ3KzczMzGwktrzOQXNvY0P/tsFT4/sORBLj+w4sMbDOEBHraz1xEfEy2TD3CXiaknWYIRt5EbEBeEbSW1PRcWRzPhaTTdSH7SfsLwZmKXMM8GIa1rkEOF7SXinhyvHAkrTtJUnHpC7vWXjyv5mZmVlLbOh/ZoeGnw1N0iTgnWQ9boVPUzJrxnDXybsA+EbKzLYG+ChZA/FmSecAa4FT0763k41LXk02NvmjABHxvKTLyDK9AXwmIp5Ptz/OtrHJd+AuazMzs44laQFwMrAxIg5PZZ7TZB1D0u7At4FPRMRLWTXNFDVNaSS5KKqS76DsOGo5LAD2H1uNtUfLek+G1ciLiAfIxhTXO67BvgGcN8DzLAAWNChfQbYGn5mZmXW+G4B/IBuKVlNbeslzmqzSJO1E1sD7RkTUshs/K2l8RKwfwTSlaXXlyxjBNKWR5KKoSr6DsuOo5bAAuGC/dZVYe7Ss92S46+SZmZmZDUtE3AM8X1fsOU1Wealn+XpgVUR8MbfJ05Sso7iRZ2ZmZkXwnCbrBMcCZwLvlfRA+jkJuBx4v6QngPel+5D1JK8hm6b0FbIpSKQpSbVpSvex4zSlr6bH/BfuhW5aLblQQynjbK8lHxrunDwzMzOzlqjinKbRKHv+UU1tHtK4Xbfdzs9DalRWhKq8PzXDiScifggM0FrwNKWqqiUXWnvFyTtuTBlngcbbu5QbeWZmZlaESs9pGo2y5x/V1OYhXfT2zcyZMydd7G6bh1TbXvTcpKq8PzVVi8esnTxc08zMzIrgOU1mZgVxT56ZVVcaR3/AhImsX/d02dGY2TBJuomsF25fSevIsmRejpdeMjMrhBt5ZlZdaRx9L42hN+sGEXHGAJs8p8nMrAAermlmZmZmZtZF3MizriZpgaSNkh7Ole0taamkJ9LvvVK5JF0tabWkByW9K/eY2Wn/JyTNzpUfJemh9JirNWD+XjMzs+qopZzvpZTyZr3EjTzrdjcAM+rKLgbuiojJwF3pPsCJwOT0cy5wLWSNQrL5JFOBo4FLaw3DtM/Hco+rfy0zM7PKqaWc39D/zNA7m1nHcSPPulpE3AM8X1c8E1iYbi8EPpQrvzEyy4E3pTTfJwBLI+L5iHgBWArMSNv2jIjlaU7JjbnnMjMzMzMrhROvWC8al1JwA2wAxqXbE4D8V5rrUtlg5esalO+g3QvyjkYZi9TmF+Td+YDNAGyaN2+HBXzz27yAr5mZmdnIuJFnPS0iQlLbV4Zt94K8o1HGorD5BXkPmnsbAGuvmMPVX//X7RbwzW/r5QV8JT0FvAxsATZHxJQ0fPhbwCTgKeDUiHghzQe9iiwV/avAWRFxf3qe2cBfp6f9bEQsTOVHsS0N/e3AhVHkG21mwza+70A29D/T3JIyaVkaM+t+Hq5pvejZNNSS9HtjKu8HJub260tlg5X3NSg3a6XpEXFkRExJ9z2n1KwHtWQOXVqWpvZFmpl1LzfyrBctBmoZMmcD382Vz0pZNo8BXkzDOpcAx0vaK10cHw8sSdteknRM6kWZlXsus3bxnFIzMzMb1LCHa0oaA6wA+iPiZEkHA4uAfYCVwJkRsUnSLmQXC0cBzwGnRcRT6TkuAc4hG3r0ZxGxJJXPIBtmNAb4akRc3qLjsx4n6SZgGrCvpHVkPRqXAzdLOgdYC5yadr+dbKjbarLhbh8FiIjnJV0G3Jf2+0xE1JK5fJxtw93uSD9mrRLA99OQ4n9Kw347ck5pFec5Vi2mvr4+LiphDupgqvYemZnZ8IxkTt6FwCpgz3T/CuDKiFgk6R/JGm/Xpt8vRMQhkk5P+50m6VDgdOAw4HeAH0h6S3quLwHvJ7vIuE/S4oh4tMljMyMizhhg03EN9g3gvAGeZwGwoEH5CuDwZmI0G8S7I6Jf0v7AUkmP5Td20pzSqsxzzKtaTPPnz+ean72t8Dmog6nae2RmZsMzrOGakvqADwBfTfcFvBe4Je1SP2SoNpToFuC4tP9MYFFEvBYRT5L1lhydflZHxJqI2ETWOziz2QMzM+t0EdGffm8EbiU7X3pOqVkvS8lTvJC5mQ1muD15fw/8JbBHur8P8IuI2Jzu54f5bB0aFBGbJb2Y9p8ALM89Z/4x9UOJpjYKopkhQ50+5KSWXn7/sX0dfRzQ+Z+FWREk7Qa8ISJeTrePBz7Dtjmll7PjnNLzJS0iO4e+GBHrJS0BPp9LtnI8cEkahvxSmn96L9mc0muKOj4zG6WUPAVg7RUnlxyMmVXVkI08SScDGyNipaRp7Q9pYM0MGer0ISe11PMX7LeO0047rexwmtLpn4VZQcYBt6Z052OBb0bEnZLuw3NKzczMbBDD6ck7FvigpJOAN5LNybuKLHPb2NSblx/mUxsatE7SWOC3yRKwDDRkiEHKzcx6UkSsAY5oUP4cnlNqZmZmgxhyTl5EXBIRfRExiSxxyr9HxB8CdwOnpN3qhwzV0tOfkvaPVH66pF1SZs7JwI/Ivl2eLOlgSTun11jckqMzMzMzMzPrMSPJrllvLrBI0meBHwPXp/LrgX+WtBp4nqzRRkQ8Iulm4FFgM3BeRGwBkHQ+2VpkY4AFEfFIE3GZmZmZmZn1rBE18iJiGbAs3V5Dlumtfp9fAx8Z4PGfAz7XoPx2svkkZmZmZtaslIXTzHrTsJZQMDMzM7MOkrJw1jJxmllvcSPPzMzMrBO5t87MBuBGnpmZmVknyq2ZZ2aW50aemZmZmZlZF3Ejz8zMzMwMkLRA0kZJD+fK9pa0VNIT6fdeqVySrpa0WtKDkt6Ve8zstP8Tkmbnyo+S9FB6zNXyeFtrEzfyzMzMzMwyNwAz6souBu6KiMnAXek+wIlk6z5PBs4FroWsUQhcCkwly0R/aa1hmPb5WO5x9a9l1hJu5JmZmVlhJD2VejIekLQilbWsp8SsGRFxD9k6z3kzgYXp9kLgQ7nyGyOzHHiTpPHACcDSiHg+Il4AlgIz0rY9I2J5RARwY+65zFrKjTwzMzMr2vSIODIipqT7rewpMWu1cRGxPt3eAIxLtycAz+T2W5fKBitf16DcrOVGtBi6mZmZWRvMBKal2wuBZcBccj0lwHJJtZ6SaaSeEgBJS8mGvd1UbNjWayIiJEURryXpXLIvNxg3bhzLli0bcN9XXnll0O1FKSuOefPmsfMBm9mUfgPsP7aPi+rKNs2bV3h8Zb0nbuSZmZlZkQL4frpQ/qeIuI7W9ZSYtcOzksZHxPr0JcPGVN4PTMzt15fK+tn2pUWtfFkq72uwf0Ppb+M6gClTpsS0adMG2pVly5Yx2PailBXH9OnTOWjubay9Ys7WZUUu2G8d1/zsbduVrb1iDtl3RsUp6z1xI8/MzMyK9O6I6Je0P7BU0mP5ja3sKRlJT8hoFPkNfaOeik11ZeN2Hd5+9WXtOoaq9C7VNBHPYmA2cHn6/d1c+fmSFpENHX4xNQSXAJ/PDSE+HrgkIp6X9JKkY4B7gVnANaM+ILNBuJFnPUvSU8DLwBZgc0RMSfM8vgVMAp4CTo2IF1KK46uAk4BXgbMi4v70PLOBv05P+9mIWIiZmTUUEf3p90ZJt5LNqWtVT0n9aw27J2Q0ivyGvlFPRe127fdFb9/MnDlzhtyvvqxdPRtV6V2qGU48km4iq1v7SlpHNvfzcuBmSecAa4FT0+63k10XrCa7NvgoQGrMXQbcl/b7TG1oMfBxsgyeuwJ3pB+zlnMjz3rd9Ij4ee5+bfL/5ZIuTvfnsv3k/6lkk/+n5ib/TyEbgrRS0uKUTcvMzHIk7Qa8ISJeTrePBz5Di3pKCjwU61IRccYAm45rsG8A5w3wPAuABQ3KVwCHNxOj2XA4u6bZ9lqSJrnooKtsfN+BSMLrvY6MpImS7pb0qKRHJF2Yyj8tqT+ln39A0km5x1ySUs0/LumEXPmMVLY6fXlRKz9Y0r2p/FuSdi72KK0HjQN+KOknwI+A70XEnWSNu/dLegJ4X7oPWU/JGrKekq+Q9YKQekVqPSX3sX1PiZlZzxuyJ0/SRLJ1PMaR9VRcFxFXtXJYm6Sj2NZ1fTtwYRQ9K9J6UWGT/9s9L2Q0iporMeeTF7LzAYcAsGnD6obzQoYzj6TI96wi80g2AxdFxP2S9iDrJV6atl0ZEfPyO0s6FDgdOAz4HeAHkt6SNn8JeD9Z/bwv9TY/ClyRnmuRpH8EziGlqDdrh4hYAxzRoPw5WtRTYmZmwxuuOdCFxlm0bljbtcDHyCah3k7WE+Ixyg1t6xE5YMJE1q97uuR4Olphk//bPS9kNIqaK1GbRwIDzwu5+uv/OuQ8kiK/96nCPJL0ZcP6dPtlSasYPHvgTGBRRLwGPClpNdlcJ4DV6eKaNOxtZnq+9wJ/kPZZCHwaN/LMzMw63pCNvEEuNFqypo2kZcCeaQgckm4kGyLnRl5DkbsAPrnkWDpbkZP/zZohaRLwTrIvwo4lm6M0C1hB9iXcC2Tn5eW5h+V7let7m6cC+wC/iIjNDfavf/2me6Ir0ju6narF1Ne3bU2nqsRVtffIzMyGZ0SJV+ouNFo1rG1Cul1fbtY2nvxvnULS7sC3gU9ExEuSriWbixTp93zg7HbG0Iqe6Cr0jtarWkzz58/fuqZTVWYsVO09sjYYsxOSPDrIrMsMu5HX4EJj67ZWDmsbIoZRf5vc6d9G1uYr7T82+6YXqNS3vSNRkc9iHHBrqsdjgW9GxJ2S7qN1aZKtVdJFCPTWMGVJO5Gdd78REd8BiIhnc9u/AtyW7g7U28wA5c+RJRAam3rzBl2U18yKN77vQDb0PzP0js3Y8noaFu/RQWbdZFiNvEYXGrRuWFt/ul2//w6a+Ta507+NrM1rumC/dVzzs7cBVOrb3pGowmfhyf8dJl2EQO8MU05JrK4HVkXEF3Pl43OjKD4MPJxuLwa+KemLZIlXJpNlLxQwWdLBZOfW04E/SF/O3Q2cAixi+55rM6uADf3P9Ny5z8xaY8glFAa60GDbsDbYcVjbLGWOIQ1rA5YAx0vaKw1tOx5Ykra9JOmY9Fqz8IWGmdmxwJnAe+uWS/iCpIckPQhMBz4JEBGPADcDjwJ3AudFxJbUS3c+2Tl4FXBz2heyedR/npK07EN2rjezAuWXmRnfd2DZ4ZhZlxhOT17tQuMhSQ+ksk+RzVlq1bC2j7NtCYU7cNIVM+txEfFDsl64ercP8pjPAZ9rUH57o8elHu2j68vNrDjurTMbvUKGNHeo4WTXHOhCA1o0rC0iVgCHDxWLmZmZmbVBj859ts5W+5LEX5DsaETZNc3MzMysC/Xg3GezbjbknDwzMzMzK0DqTRu7y65be9XMzEbDjTwzMzOzKki9aVs2/Xprr5qZ2Wi4kWdmZmZmZt0v9Zb3QiZbN/LMzMzMzKz7pd7yXsjI6UaemZmZWcFq6+NVUg/1dph1KzfyzMzMzAqWXx+vcnqot8OsW7mRZ2ZmZtZGtV4794yZWVHcyDOztqj0UCQzswLVeu3cM2ZmRXEjz8zaotJDkczMzMy6mBt5ZmZmJan1eLvX28zMWmls2QGYmZn1qu17vB8rNRZrjfF9B7Kh/xkOmDCR9eue3n5jylppZtZu7skzMzMza5FB59+lrJUeym42eh4BMTxu5JlZZ/H6TWZmZj2r9kWKvywZnBt5ZtZZvH6TmZmZ2aAq08iTNEPS45JWS7q47HjMhst1dxsPoeg8rr/WqSpfd9OoA58PrZHK199ul/v77NaRQZVo5EkaA3wJOBE4FDhD0qHlRlU+XzBXn+vu9jyEorO4/lqnqlrdbfj/uhvm3/XAhXAZqlZ/O0VL19/N/X1268igSjTygKOB1RGxJiI2AYuAmSXHVLohL5g9N6kKXHcpaeFzX3y0gutvFbluD0el6m7XfsHVAxfCJalU/e0UXn93ZBQRZceApFOAGRHxx+n+mcDUiDi/br9zgXPT3bcCj4/gZfYFft6CcMvWDcfR7mM4KCL2a+Pzb1VQ3W2XKtWlKsUC5cVTWN2Fwutv1T5jqF5MVYsHRhZTL557q/aZOZ7BDRRPN5x7q/JeVyUOqE4spVz3dtQ6eRFxHXDdaB4raUVETGlxSIXrhuPohmMYqWbqbrtU6XOoUixQvXjK1or6W8X3tGoxVS0eqGZMI9Huc2/V3h/HM7iqxTOUkdTfqhxbVeKA6sRSVhxVGa7ZD0zM3e9LZWZV57prncz11zqV6651Mtdfa7uqNPLuAyZLOljSzsDpwOKSYzIbDtdd62Suv9apXHetk7n+WttVYrhmRGyWdD6wBBgDLIiIR1r8MpUaKteEbjiObjgGoLC62y5V+hyqFAtUL562KLj+VvE9rVpMVYsHqhlTlc69VXt/HM/gKhFPm+pvJY6N6sQB1YmllDgqkXjFzMzMzMzMWqMqwzXNzMzMzMysBdzIMzMzMzMz6yJd38iT9HeSHpP0oKRbJb0pt+0SSaslPS7phDLjHIqkGSnO1ZIuLjue4ZA0UdLdkh6V9IikC1P53pKWSnoi/d6r7Fh7laRPS+qX9ED6OamEGCpVtyU9Jemh9H6sKDuebiHpIkkhad90X5KuTp/7g5LeVWAslfu/UPbfgc/XozdYfSowhsqcRweqS2WTNEbSjyV1zWrakj6S3uPfSJpSt620a9yyry0q9vdQ2jVF1zfygKXA4RHxDuCnwCUAkg4ly2Z0GDAD+LKkMaVFOYgU15eAE4FDgTNS/FW3GbgoIg4FjgHOS3FfDNwVEZOBu9J9K8+VEXFk+rm9yBeucN2ent6P0tfX6QaSJgLHA0/nik8EJqefc4FrCwypUv8XKvJ34PP16DWsT0WpSP3JG6gule1CYFXZQbTYw8D/Bu7JF1bkGreUa4sK/j1ASdcUXd/Ii4jvR8TmdHc52VokADOBRRHxWkQ8CawGji4jxmE4GlgdEWsiYhOwiCz+SouI9RFxf7r9MtnJdQJZ7AvTbguBD5UToVVAR9ZtG7Ergb8E8pm+ZgI3RmY58CZJ44sIpoL/F0r/O/D5evQGqU9FKb3+5A1Sl0ojqQ/4APDVMuNotYhYFRGPN9jUSde4rVapv4cydX0jr87ZwB3p9gTgmdy2dZR8EhpEJ8XakKRJwDuBe4FxEbE+bdoAjCspLMucn4YZLShhKFYV63YA35e0UtK5JcfS8STNBPoj4id1m6ry2Vfh/0JV3gvA5+sm5etTUSpVf/Lq6lKZ/p7si6bflBxHUapQJ8q6tqjCseeVdk1RiXXymiXpB8ABDTb9VUR8N+3zV2RDCL5RZGwGknYHvg18IiJekrR1W0SEJK/j0UaD/X2QDZG7jOwkdBkwn+wipZe9OyL6Je0PLJX0WETcM+SjetgQdexTZEM1C+X/C6Pj83Vjrk8jV1+XSozjZGBjRKyUNK2sOEZrOHWvDL62GLbSrim6opEXEe8bbLuks4CTgeNi28KA/cDE3G59qayKOinW7Ujaiewk/42I+E4qflbS+IhYn4ZnbSwvwu431N9HjaSvAEVPSK9c3Y6I/vR7o6RbyYZ+uJE3iIHqmKS3AwcDP0mNhT7gfklH0+bPvsP+L1Ti78Dn64GNsj4VpRL1J2+AulSWY4EPpuQfbwT2lPT1iPijkuMaluH+D6/T9jpR4WuLSv09lHlN0fXDNSXNIOui/2BEvJrbtBg4XdIukg4mm/z/ozJiHIb7gMmSDpa0M9lk2sUlxzQkZVd11wOrIuKLuU2Lgdnp9mygtG+iel3dHKgPk03iLlKl6rak3STtUbtN1gNV9HvSNSLioYjYPyImRcQksmEz74qIDWSf8yxljgFezA0LbKsK/l8o/e/A5+vRG6Q+FaX0+pM3SF0qRURcEhF96Rx0OvDvndLAa0Kp17glX1tU5u+h7GuKrujJG8I/ALuQdZECLI+IP42IRyTdDDxKNrzivIjYUmKcA4qIzZLOB5YAY4AFEfFIyWENx7HAmcBDkh5IZZ8CLgdulnQOsBY4taT4DL4g6UiyIRVPAX9S5ItXsG6PA25N54qxwDcj4s4S4+lmtwMnkSUEeBX4aIGvXan/CxX5O/D5evQa1qeiXrwi9SevYV0qOntzL5D0YeAaYD/ge5IeiIgTKnCNW9q1RcX+Hkq9plDxowrMzMzMzMysXbp+uKaZmZmZmVkvcSPPzMzMzMysi7iRV0GS7pT0mQblMyVtkPQXkh6W9LKkJyX9RYN9L0zbfilplaS3FBO99bpm6q+kAyW9UvcTki4q9iisFzV77pV0pKT/kPSipHWS/qa46K3XtaD+/r+SfpS2Pyjp3cVFb71smHV3jaSXJP23pCsljc3tN0nS3ZJelfSYpNFkJO06buRV00Lgj1KGqrwzydbfETAL2AuYQbbg5Om1nST9MXAO8AFgd7K0zj8vIG4zaKL+RsTTEbF77Qd4O9nitd8uLHrrZU2de4FvkqXG3hv4PeDjkj7Y9qjNMqOuv5L2Bv4N+DvgTcAXgH9TsYtYW+8aqu5+hywz857A4cARwJ/l9rsJ+DGwD9k6fbdI2q/tUVecE69UkKRdgQ3A79cWTEwn2vXA1Ij4Sd3+V5N9lhdIegNZBrSzIuKugkM3a6r+NniuS4FpETG9/ZFbr2u27kp6FZgSEY+m+/8C3B8R/6fAw7Ae1eS1w8nAFRFxWG77T1PZ9YUdhPWkkdRdSfsA3wJ+GhEfTyPVHgL2jYiX0z7/QbZG4z8WfCiV4p68CoqIXwE3k33jVnMq8FiDk7SA/wXU0sP2pZ/DJT2ThmT8bWr8mbVdk/W3ftsssm/4zNquBXX378nW/ttJ0luB3wV+0N6ozTItqL/1vSgi6zUxa6vh1F1JfyDpJbKRaUcA/5T2OwxYU2vgJT9J5T3NF/7VtRA4RdIb0/2BLnY/TfY5fi3d70u/jycb6jYdOINs+KZZUUZbf/PeTbbGzC3tCNBsAM3U3duAU4BfAY8B10fEfe0L1WwHo62//wn8jqQz0pcUs4E3A7/V5njNagatuxHxzTRc8y3APwLPpk27Ay/WPdeLwB7tDbf63MirqIj4Idm3FR+S9GbgaLL5HlspW+xxFvCBiHgtFf8q/f5CRPwiIp4i+7bjpEICN6Op+ps3G/h2RLzS7njNakZbd9OcpjuBzwBvBCYCJ0j6eIHhW48bbf2NiOeAmcCfk108zyDrhV5XXPTWy4ZTd9N+T5D1QH85Fb0C7Fm3257Ay/S4sUPvYiW6kexE/FZgSUTUvrVA0tnAxcB7IiJ/En4c2ATkJ1t64qWVYTT1t7Z9V+AjwIcLitUsbzR1938AWyLixnR/naRFZF+wfRmz4ozq3BsR/x/w/6T9xgJrgPlFBW3GIHW3zliynmbIGnz/Q9IeuSGbR9Cggdhr3JNXbTcC7wM+Rq7LWtIfAp8H3h8Ra/IPiIhXySak/qWkPST1AeeSDSMyK9KI62/Oh4EXgLvbHaRZA6Opuz/NdtEfSHqDpAOA04AHC4rZrGZU515J70xDNfcE5gHPRMSSgmI2g4Hr7h9L2j/dPhS4BLgLICJ+CjwAXCrpjZI+DLwDZ+V2ds2qk7SM7BuJA3LDgp4km3vGZ/35AAAgAElEQVSXH+L29Yj407R9T+A6siUUfgF8Bbgs/GFbwUZTf9M+S4AfRYTXGbNSjPLc+17gCrI5I78iS0l/Yfryzawwo6y/N7FtasedwAURsbGwoM0YsO5+jaxu7g78DPgX4G8i4tdp+yTgBmAq8DRwXkT0fNIrN/LMzMzMzMy6iIdrmpmZmZmZdRE38szMzMzMzLqIG3lmZmZmZmZdxI08MzMzMzOzLtKx6+Ttu+++MWnSJH75y1+y2267lR1OU7rhGKA6x7Fy5cqfR8R+ZccxkFrdrVeV96/dfJwDq3rdhYHrb5G6vQ516vFVvf5W/dzrOMqLo+p1FxrX36p8VkXqtWMezvEOVH87tpE3adIkVqxYwbJly5g2bVrZ4TSlG44BqnMcktaWHcNganW3XlXev3bzcQ6s6nUXBq6/Rer2OtSpx1f1+lv1c6/jKC+OqtddaFx/q/JZFanXjnk4xztQ/fVwTTMzMzMzsy7iRp6ZmZmZmVkXcSPPzMzMzGwYJI2R9GNJt6X7B0u6V9JqSd+StHMq3yXdX522T8o9xyWp/HFJJ5RzJNbt3MiriPF9ByIJSYzvO7DscMzaYnzfgaxcudJ13Jry4IMP+VxpLVOrT65TNkwXAqty968AroyIQ4AXgHNS+TnAC6n8yrQfkg4FTgcOA2YAX5Y0pqDYO0btuth/k6PnRl5FbOh/hoPm3sZBc29jQ/8zZYdj1hYb+p9h5wMOcR23prz++iafK61lavXJdcqGIqkP+ADw1XRfwHuBW9IuC4EPpdsz033S9uPS/jOBRRHxWkQ8Cfz/7d1/sF1lfe/x96cJwZSqgFdO0pwQ0kuqEymiZkgcO06OSAhcxugdf4R2TFBq7gyJpS2ZhrSdi1ekF+490Qql8YKkJGqJ1Pojk0ZjpDnjdMZgQGkgYC5HBJMzCaGEH0aL+XG/94/17JN1TvY+v/bvtT+vmTN77Wetvfbz7P2cZz/PWt/1rH7g0saUoH2U+sX+n5y4tp1d08zMzJrr4x//OFu3buW8884bTJN0LvBV4ALgGeDDEfFi6uB+HrgK+BVwbUT8KL1mOfBXaRefiYiNKf0dwH3AVGAbcENERKX3qGdZzYC/Af4ceG16/gbgpYg4kZ4fAGak5RnAfoCIOCHp5bT9DGBXbp/51wwhaQWwAqCrq4u+vr4h648ePXpaWlH09vYyZdoJjvX2DiljkctcTjXl9SDPCk3SnwJ/BATwGPAxYDqwmayxfQT4aEQck3QmsAl4B/AC8JGIeCbtZy1Z6MVJ4I8jYntKX0zWaZkEfDEibmtc6czMmuvaa69l1apVLFu2LJ98E/BgRNwm6ab0fA1wJTAn/c0H1gPz04DtZmAeWVv9iKQtadC2HvgE8BDZIG8x8O0R3sOsLiRdDRyOiEckLWzEe0bE3cDdAPPmzYvhU+kX+XYCPT09zFqzlWdvX01EDKYXuczlVFNeh2taYUmaAfwxMC8iLiIbiC2lRvHzKYb+LrKOy1zgmrStWU1I+lNJeyU9Lul+Sa+p5UX+khantP7UUTYbl3e/+92ce+65w5PzYWrDw9c2RWYXcLak6cAVwI6IOJIGdjuAxWnd6yJiV2S9vE2UD4XLv4dZvbwLeJ+kZ8gOFL+H7CDv2ZJKJ026gYG0PADMBEjrX092AHkwvcxrzGrGZ/Ks6CYDUyUdB34TOEjWMP9BWr8R+BTZ0eIlaRmy+Pm/HR4/D/xMUj5+vj8ingaQtDlt+0Sdy2QdIHeQYm5E/IekB8gONlxFdpBis6QvkB2cWE/uIIWk0sGMjww7SPHbwPck/W56m7uAy8nChXansyeuv1atrog4mJYPAV1peTB8LSmFqY2UfqBM+kjvMcRo4W4A3d3d3Dgti7YbHhrWSK0ShuZ8lBcRa4G1AOlM3uqI+ENJ/wh8kGzgtxz4VnrJlvT8B2n9v6RQ4y3AP0j6LFmbPAf4YSPLYp3BgzwrrIgYkNQL/Bz4D+C7ZOGZtYyfH94xmV+Holjn8kEKa2upUxujb1mf9xgt3A1g3bp13Pn8mwFOCw1rpFYJQ3M+xm0NsFnSZ4AfA/em9HuBL6U29wjZwTYiYm86aPcEcAJYGREnG59tKzoP8qywJJ1D1mmdDbwE/CNZuGUz8jLq0eRWO2pZD729vXRNzR6LXtZqv89WOkgxlvrbSKUzL80861JP7dYWHDp0iF/+8pf5pOckTY+Igynk8nBKrxSmNgAsHJbel9K7y2w/0nuY1V1E9JHVUdKBstNmx4yIV4EPVXj9rcCt9cuhmQd5VmzvBX4WEc8DSPo6WUz92ZImp45yufj5A+OInx9TXP1Yjia30VHLCevp6eGOL3+T1aubd7S8Uar9PlvpIMVY6m8jlc68NPOsSz21W1vwzDPPcNZZZ+WTSmFqt3F6+NqqdNZ4PvByGqRtB/461XmARcDaiDgi6RVJC8gmXlkG3DnKe5iZGZ54xYrt58ACSb+ZwtYuIwuP2EkWHw/l4+chFz+f0pemiS1mcyp+fjcwJ02EMYUsFGNLA8plnWHwIEVEHAeGHKRI21Rzkb8v/reqXXPNNbzzne9k3759ABdLuo5s4HW5pKfI6nFp1uFtwNNk9wW7B7geICKOALeQtam7gU+nNNI2X0yv+SnZzJqM8B5mZobP5FmBRcRDkr4G/Igs7v3HZGcj/pkaxc9LWgVsJ5u5c0NE7G1U+azwBg9SkIVrXgY8zKmDFNVe5C/SQQqywd1STl3rZzYm999//+CypD0RUWpPLxu+bTpotrLcfiJiA7ChTPrDwEVl0l8o9x5mZpbxIM8KLSJuJrv/Ul7N4ucjYhvZ0WmzmvJBCjMzM5soD/LMzFqUD1KYmZnZRPiavAaY3n0+kpjefX6zs2JmZmZmZgXnQV4DHBrYz6w1Wzk0sH/0jc3MzMzMzKow6iBP0gZJhyU9nks7V9IOSU+lx3NSuiTdIalf0h5Jb8+9Znna/ilJy3Pp75D0WHrNHWkWxGKadIbP6JmZmZmZWV2N5UzefZx+b6abgAcjYg7wYHoOcCXZzG1zyG6cux6yQSHZdSXzya4luTl3P5z1wCdyr2vKfaAa4uTx7IzeoUNI8oDPzMzMzIxTlzcV+XxPI406yIuI75PN1Ja3BNiYljcC78+lb4rMLrL7OU0HrgB2RMSRiHgR2AEsTuteFxG70tTKm3L7Kq402HMIp5mZmZnZqcubZq3Z2uysFMJEZ9fsioiDafkQ0JWWZwD5UcuBlDZS+oEy6WVJWkF2hpCuri76+vo4evQofX19EyxGY/T29jJl2gmODXsEONbby9GjRwe3KaW1epnKaYfvwszMzMyslqZ3n8+hgf1MmzGTgwd+3uzsADW4hUK62W7UIjNjeK+7ye4Txbx582LhwoX09fWxcOHCRrz9hPX09DBrzVaevX31kEeAZ29fzc6dO1m9OpfWuxZOHm+pijIW7fBdmJmZmdn41XIgU9oX0Hb93XJKZyGf7f3AYLjppCmv4eSxV5tWvonOrvlcCrUkPR5O6QPAzNx23SltpPTuMumdrXTtnkM5zczMzKwF1HK2+HxoZqH6u7lLsk4ee7Wp5ZvoIG8LUJohcznwrVz6sjTL5gLg5RTWuR1YJOmcNOHKImB7WveKpAVpVs1luX2ZmZmZmVkrSbPFewLB8WvkvbPHcguF+4EfAG+SdEDSdcBtwOWSngLem54DbAOeBvqBe4DrASLiCHALsDv9fTqlkbb5YnrNT4Fv16ZoZmZmZmZWUx0wgWBpMDb5zKmDA9rScjUDtEbeO3vUa/Ii4poKqy4rs20AKyvsZwOwoUz6w8BFo+XDzMzMzMw6SDprCNm1e/d/eVND3nbwGrvbr87No3H1YFo7mGi4ppmZmZlZR5D0Gkk/lPRvkvZK+h8pfbakhyT1S/qqpCkp/cz0vD+tvyC3r7UpfZ+kK5pTosbJ3/+u7Fmw3EDuNB1w1rBePMgzMzMzMxvZr4H3RMRbgUvI7ve8ALgd+FxEXAi8CFyXtr8OeDGlfy5th6S5wFLgLcBi4O8kTWpoSeogP5AbHtY46iQraSBXWLlrGBvJgzwzMzMzsxFE5mh6ekb6C+A9wNdS+kbg/Wl5SXpOWn9ZmmRwCbA5In4dET8jm5Pi0gYUoa7yA7kRZ5Uc6axdUeXORjZS1ffJMzMzMyuR9Cbgq7mk3wH+O3A28Ang+ZT+FxGxLb1mLdmZj5PAH0fE9pS+GPg8MAn4YkTcltJnA5uBNwCPAB+NiGN1Lpp1uHTG7RHgQuAusgkDX4qIE2mTA8CMtDwD2A8QESckvUxWX2cAu3K7zb9m+PutAFYAdHV10dfXN2T90aNHT0urtT17HuP48exfq7e3lynTTnAsPQIc6+2lr69vcF0prbTdaetu/59MmXYhxw71l91+tLRGlHmksubLNZ7tR9rH4GcsQWS3Hj/jjClcfPHvVVVeD/Ks0CSdTTZ760VkR9w+Duwj64BcADwDfDgiXkxH2D4PXAX8Crg2In6U9rMc+Ku0289ExMaU/g7gPmAq2eyyN6QJiMyq5vpr7Sgi9pGFs5U6xQPAN4CPkYW19ea3Hxa+9tvA9yT9blp9F3A5WUd4t6QtEfEEp0LkNkv6AtkAcX3dC2cdLSJOApektvkbwJvr/H53A3cDzJs3LxYuXDhkfV9fH8PTaq2np6fMxCOrc2mriYhh260e3G6kdRNJ27lzZ93LnC93pTxFxJAbugPjLlfp5/bUew2d5CUiqvqOHa5pRfd54DsR8WbgrcCTwE3AgxExB3gwPQe4EpiT/laQOgySzgVuBuaThVTcnO73SNrmE7nXLW5AmaxzuP5au7sM+GlEPDvCNpXC1y4F+iPi6XSWbjOwJB3QqBQiZ1Z3EfESsBN4J3C2pNJJk26ygxqkx5kAaf3rgRfy6WVeY20kH6LainwmzwpL0uuBdwPXAqROwjFJS4CFabONQB+whqyjsSmdydgl6WxJ09O2O0r3dpS0g+yC6z7gdRGxK6VvIuto+F6PVjXXXyuIpcD9ueerJC0DHgZujIgXGTl8bf+w9PlkIW+VQuSGGC3cDaC7u5sbh4WeNUOjwtCcj4mR9EbgeES8JGkq2Rnm28kGex8kOwixHPhWesmW9PwHaf2/RERI2gL8g6TPkp25ngP8sKGFaZSiXn/XJuXyIM+KbDbZtR9/L+mtZHH0NwBdEXEwbXMI6ErLg/HzSanjMFL6gTLppxlLR6PVftDqobe3l66p2WPRy1qD77Ot6m8jlTrlzeyQ11NR2oI0lfz7gLUpaT1wC1no8S3AOrIQ5LoZLdwNYN26ddz5fBZ1lw+harRGhN45H1WZDmxMIci/ATwQEVslPQFslvQZ4MfAvWn7e4EvSeoHjpAd8CAi9kp6AHgCOAGsTGGgxZMmHGmX+8qNWZuUy4M8K7LJwNuBT0bEQ5I+z6nQNiCbLUtS3X/Rx9LRaMEftJrr6enhji9/k9Wrm9eRapQafJ9tVX8bqdQpb2aHvJ4K1BZcCfwoIp4DKD0CSLoHKMU4jRS+Vi79BVKIXDqb53A3q7uI2AO8rUz605SZHTMiXgU+VGFftwK31jqPtTT8erOK2uSsVstpwOfma/KsyA4AByLiofT8a2Sd5udSGBvp8XBaX6mjMVJ6d5l0s1pw/bV2dw25UM1SvU0+ADyelrcAS9PNo2dzKnxtNzAn3Wx6CtmZkC0pJLkUIgdDQ+TMrAZK15uNquj3uKuXBtxWwYM8K6yIOATsT9N5QzYBwBOcipOH0+PnlymzAHg5hcVtBxZJOidNWLEI2J7WvSJpQZoIYBnuaFiNuP5aO5N0Ftk1S1/PJf8vSY9J2gP0AH8KWfgaUApf+w4pfC2dpVtFVoefJAuP25v2tQb4sxQK9wZOhciZmRkO17Ti+yTwlXQU+GmyKbx/A3hA0nXAs8CH07bbyKaf7yebgv5jABFxRNItZEeVAT5dmsQCuJ5TU9B/G09aYbXl+mttKSJ+STb4yqd9dITty4avpfvobSuTXjZEzszMMh7k1cmYY5mtriLiUWBemVWXldk2gJUV9rMB2FAm/WGye5iZ1Zzrr5mZmU2EwzXrpNXvnWFmZmZmZpnp3ecjqTATyXiQZ2ZmZmZmHa1oJ2g8yDMzMzMzMysQD/LMzMzMzMwKxIM8MzMzMzOzAvEgz8zMzMzMrEA8yDMzMzMzMysQD/KabdIZPPLII83OhZmZmZmZtYJJZyCJPXsem/AuPMhrtpPHmTLtwmbnwszMzMysdaUTI9O7z292Turv5HFmrdnK8ePHJrwLD/LMzMzMzKy1pRMjhwb2NzsnbcGDPDMzMzMzswLxIM/MzMzMzKxAPMgzMzMzMzMrEA/yamx69/lIanY2zMzMzKxGJM2UtFPSE5L2SrohpZ8raYekp9LjOSldku6Q1C9pj6S35/a1PG3/lKTlzSqTFZsHeTV2aGA/s9ZsbXY2zMzMmkrSM5Iek/SopIdTWs06xJLekfbfn17rI6xWTyeAGyNiLrAAWClpLnAT8GBEzAEeTM8BrgTmpL8VwHrI/geAm4H5wKXAzaX/A7Na8iDPCk/SJEk/lrQ1PZ8t6aHUMfiqpCkp/cz0vD+tvyC3j7UpfZ+kK3Lpi1Nav6Sbhr+3WbVcf63N9UTEJRExLz2vZYd4PfCJ3OsW17841qki4mBE/Cgt/wJ4EpgBLAE2ps02Au9Py0uATZHZBZwtaTpwBbAjIo5ExIvADlx3rQ4mNzsDZg1wA1lj/Lr0/HbgcxGxWdIXgOvIOgvXAS9GxIWSlqbtPpKO1C0F3gL8NvA9Sb+b9nUXcDlwANgtaUtEPNGogllHcP21IlkCLEzLG4E+YA25DjGwS1KpQ7yQ1CEGkLQDWCypD3hd6jwjaRNZ5/rbDSuJdax0EO1twENAV0QcTKsOAV1peQaQn+v/QEqrlF7ufVaQHfSgq6uLvr6+IeuPHj16Wlqt9Pb2MmXaCY6lR2BwebxptdhHKa1rapa3epS7t4nlqpR23uTuCZfVgzwrNEndwH8BbgX+LIXzvAf4g7TJRuBTZJ3kJWkZ4GvA36btlwCbI+LXwM8k9ZMdUQboj4in03ttTtu6k2w14fprbS6A70oK4P9ExN3UrkM8Iy0PTzerK0m/BfwT8CcR8Uo+SjgiItX3mkj/M3cDzJs3LxYuXDhkfV9fH8PTaqWnp4dZa7by7O2rBy9DKi2PN60W+yil3fHlb7J69Wqy40H1KXMzylUp7ZNvPMBHPvKRCZXHgzwrur8B/hx4bXr+BuCliDiRnuc7BoOdiYg4IenltP0MYFdun/nXDO98zC+XidGOxkF9j8i1it46H4VrJTX6Plui/ppN0O9HxICk84Adkn6SX1nrDnE5Y2l7u7u7uTF39LxZbVOr/AY4H5VJOoNsgPeViPh6Sn5O0vSIOJjOPh9O6QPAzNzLu1PaAKfOZpfS++qZb+tMHuRZYUm6GjgcEY9IWtjMvIx2NA7qe0SuVfT09NT1KFwrqfb7bKX6O5aOciOVOuXN7JDXUyt2biciIgbS42FJ3yA7g1yrDvFAWh6+/fA8jNr2rlu3jjuffzOQHT1vVtvUKr8Bzkd5KTLiXuDJiPhsbtUWYDlwW3r8Vi59VYqSmA+8nOr9duCvc9eWLgLWNqIM1o6EJKbNmMnBAz8f1ys9yLMiexfwPklXAa8hu6bp82QXP09OZ0PyHYNSJ+OApMnA64EXqNz5YIR0s2q1TP0dS0e5kUqd8mZ2yOup1Tq3EyHpLOA3IuIXaXkR8Glq1CGOiCOSXpG0gOy6qGXAnY0qn3WkdwEfBR6T9GhK+wuyuvyApOuAZ4EPp3XbgKuAfuBXwMcAUt29Bdidtvt06ZpTs9NFCuW8etyvrGp2TdV5emSzakTE2ojojogLyCae+JeI+ENgJ/DBtNnwTkap/n0wbR8pfWmavXA22SxuPyRroOek2Q6npPfY0oCiWQdw/bU21wX8q6R/I6tv/xwR3yHrEF8u6Sngvek5ZB3ip8k6xPcA10PWIQZKHeLdDO0QXw98Mb3mp3jSFaujiPjXiFBEXJxmjL0kIrZFxAsRcVlEzImI95bqZ5pVc2VE/OeI+L2IeDi3rw0RcWH6+/vmlcqKrBZn8noi4t9zz0vTI9+WpuS+iWzmrPz0yPPJJgqYn5seeR7ZRdqPpBneXqxB3szKWQNslvQZ4Mdk4Rekxy+liSmOkHV6iYi9kh4gm5DiBLAyIk4CSFoFbAcmARsiYm9DS2KdyPXXWl6a0OetZdJfAC4rkx7Aygr72gBsKJP+MHBR1Zk1MwCmd5/PoYH9o29obaEe4Zo1mR4ZuL8OebMOFRF9pAubU+fj0jLbvAp8qMLrbyWb4XB4+jayI9BmdeP6a2Zm9XZoYH9udsfxhwdaa6l2kFfP6ZFPU+7i/1a7QH0s9xWpdM+P07dfx7p16wA444wpXHzx7zWnUGPUat+FmZmZmVknqnaQ19Dpkctd/N9qF6iP5b4ile75cfr2Nw45otLqEwy02ndhZmZmZjaSooapVjXxSn56ZGDI9MgA45ge2TMUmpmZmZlZQ+XDVItkwoM8SWdJem1pmWxa48cZOsPb8JnflqVZNheQpkcmu+h/kaRz0kyci1KamZmZmZlVML37fCQxvfv8ZmfFWkw14ZpdwDeye0MyGfiHiPiOpN34fiFmZmZmZnVVOgvliVJsuAkP8hoxPbKZmZmZmdmgSWeQTjIxbcZMDh74eZMz1JqquibPGihVaJ+ONzMzM7OOdfI4s9ZsZdaarYWcMKVWPMhrF6lCuzKbmZmZmdlIPMgzMzMzM7OOUZqwphT2WUQe5JmZmZmZWccoTVhTxFsnlHiQZ2ZmZmZmViAe5NVAJ5zyNTNrJrezZmZmY1fNffIsKZ3yBXyfEjOzOsi3s/CTpubFzMys1flMnpmZmZmZWYF4kGdmZmZmNgpJGyQdlvR4Lu1cSTskPZUez0npknSHpH5JeyS9Pfea5Wn7pyQtb0ZZrPg8yGs36abokph85tQhj75Z+lCSZkraKekJSXsl3ZDSa9YgS3qHpMfSa+6QLxiyGnH9tXY1Qt39lKQBSY+mv6tyr1mb6uE+SVfk0hentH5JN+XSZ0t6KKV/VdKUxpbSOtR9wOJhaTcBD0bEHODB9BzgSmBO+lsBrIesDQduBuYDlwI3l9pxs1ryIK/dpJuiz1qzlZPHXh3y6Juln+YEcGNEzAUWACslzaW2DfJ64BO51w1v/M0myvXX2lWlugvwuYi4JP1tA0jrlgJvIauDfydpkqRJwF1kdXsucE1uP7enfV0IvAhc16jCWeeKiO8DR4YlLwE2puWNwPtz6Zsisws4W9J04ApgR0QciYgXgR247bU68MQrVlgRcRA4mJZ/IelJYAZZw7swbbYR6APWkGuQgV2SSg3yQlKDDCBpB7BYUh/wutR4I2kTWeP+7UaUz4rN9dfa1Qh1t5IlwOaI+DXwM0n9ZAckAPoj4mkASZuBJWl/7wH+IG2zEfgU6cCGWYN1pToPcAjoSsszgPyR9wMprVL6aSStIDtoR1dXF319fUPWHz16lN7eXqZMO8Gx3t7T1o/Vnj2Pcfz4scF9ARzL7bfatFrso5TWNZXT8jmRcteirLUsV6W08yZ3c+MEv18P8qwjSLoAeBvwELVrkGek5eHp5d5/xIYassZ6og10u+jNNdBFL2stv89m11+ziRpWd98FrJK0DHiY7Gzfi2T1blfuZfm6OLzuzgfeALwUESfKbD/8/Udte7u7s04UTLzDWAut8hvgfExcRISkqOH+7gbuBpg3b14sXLhwyPq+vj5Wr17NrDVbefb21WTH+Mavp6cn7ePq3Gzxp/ZbbVot9lFKu+PL3xwscyltIuUulblVylUp7ZNvPMCdz795QuX0IK8K07vPd3hkG5D0W8A/AX8SEa/kLzuqdYNcyWgNNWSNdbn0Iunp6RlsoCf6Y9QuavV9tkL9HUtHud7yR12rObLZDtqxc1tOmbq7HrgFiPS4Dvh4PfMwlrZ33bp13Pn8m4GJdxhroVV+A5yPcXtO0vSIOJiiJw6n9AFgZm677pQ2wKlojFJ6XwPyaR3Gg7wqlO7b5HvjtS5JZ5B1Mr4SEV9PybVqkAfS8vDtzWqiVervWDrK9ZY/6lrNkc120Ead24rK1d2IeC63/h6gdOPDSnWXCukvkF3fNDmdzXPba820BVgO3JYev5VLX5XCjOcDL6d2ezvw17lroxcBa6vORZqYD2DajJkcPPDzqndp7c0Tr1hhpZkC7wWejIjP5laVGmQ4vUFelmYpXEBqkIHtwCJJ56RGeRGwPa17RdKC9F7Lcvsyq4rrr7WrSnU3HZQo+QBQmoZ+C7BU0pmSZpNNAvRDYDcwJ82kOYVscpYt6brTncAH0+vz/wdmdSPpfuAHwJskHZB0Hdng7nJJTwHvTc8BtgFPA/3APcD1AOn66FvI6vdu4NOla6arkpuYz1FmBj6TVzzpSI6P4gDZ9R8fBR6T9GhK+wuyBviB1Dg/C3w4rdsGXEXWIP8K+BhkDbKkUoMMQxvk68mmVJ5KNmGFJ62wWnH9tXZVqe5eI+kSsnDNZ4D/BhAReyU9ADxBNjPnyog4CSBpFdmBiknAhojYm/a3Btgs6TPAj8kGlWZ1FRHXVFh1WZltA1hZYT8bgA01zJrZaTzIK5p0JMchpBAR/wpUuu9XTRrkiHgYuKiKbJqV5fpr7WqEurtthNfcCtxaJn1budelGTcvHZ5uZh3GJzcqcrhmwU3vPt83SrfWkxpl10szM7P66Ig+YDq5MdYQ1dJn0gk8yCu40uQwjtG2luJrB8zMzOqqE/uAow1sS59JJ3C4ZlHlZlkyMzMzMyu6/CCu0y9d8pm8okpnSoZIA7/CnrI3MzMzs4pKZ7omnzl18IxXIfjkxmk8yOsk44xbNjMzM7M2M8KAp3Sm6+SxVwdDOQuh3MmNDudB3jjlY33NzACAGu0AAAeYSURBVMzMzFqKBzyGB3njlr+I1czMzMzMrNV4kGdmZmZmVmQdfM1ap0bheZBnZmZmZlZknRjCmQa2nRqF50GemdVVpx5BMzMzsybqxIFtjgd5ZlZXnXoEzczMzKxZPMgbo9LZCDMzMzMzs1bmQd4Ylc5GmJmZmZmZtTIP8kbga4nMzMzMzKzdeJBXRmlwV9hridJsQ5KY3n1+s3NjZmbWGtLvo38bzazdeZBXRuFDM9NsQ7PWbOXQwP5m58bMzKw1pN9H/zaaWbtrmUGepMWS9knql3RTs/NjNlauu9bOXH+tXbnuWjtz/bV6a4lBnqRJwF3AlcBc4BpJcxuZh469/s6hKVVphbrbqsY8I63rYNO4/lq7qnvd9WUNVkdue60RWmKQB1wK9EfE0xFxDNgMLGnEGxf++rvRlEJTDh0a/EGbfObUIY/V/MjlB88F/aFsWt1tJaXvOV9nxhz2PKwOFrSetKqWr78jHixwR7yT1bfu5i9rSG1TqX1zXbMaaPm219qfIqLZeUDSB4HFEfFH6flHgfkRsWrYdiuAFenpm4B9wH8C/r2B2a2HIpQBWqccsyLijY14oyrr7nCt8vnVm8tZWcPqLtS8/jZS0etQu5bPbW91nI+hGpmPdm17W+W7aqROK/NYylu2/k6uT37qIyLuBu7Op0l6OCLmNSlLNVGEMkBxylEP5erucJ3y+bmc7Wcs9beRivTZllP08jVSO7W9zkdr5qOZRqu/nfgZdVqZqylvq4RrDgAzc8+7U5pZq3PdtXbm+mvtynXX2pnrr9VdqwzydgNzJM2WNAVYCmxpcp7MxsJ119qZ66+1K9dda2euv1Z3LRGuGREnJK0CtgOTgA0RsXeML2+ZEKIqFKEMUJxyjFmVdXe4Tvn8XM4WUeP620gt/9lWqejlq1pB217nY6hWyUfN1bD+FvYzGkGnlXnC5W2JiVfMzMzMzMysNlolXNPMzMzMzMxqwIM8MzMzMzOzAinEIE/S/5b0E0l7JH1D0tnNztNYSVosaZ+kfkk3NTs/4yVppqSdkp6QtFfSDc3OUzuR9KH0uf0/SfOGrVub6sU+SVc0K4+10u51fSSSNkg6LOnxXNq5knZIeio9ntPMPBZBEetQpTbU9acxmlmnJD0j6TFJj0p6OKXV/XsfT3ulzB3p89kj6e11zsenJA2kz+RRSVfl1hXqN7FWRupHFEkR2/+RlPv/GK9CDPKAHcBFEXEx8H+BtU3Oz5hImgTcBVwJzAWukTS3ubkatxPAjRExF1gArGzDMjTT48B/Bb6fT0yf4VLgLcBi4O9SfWlLBanrI7mP7HvKuwl4MCLmAA+m5zZBBa5DldpQ1586a5E61RMRl+Tug9WI7/0+xt5eXQnMSX8rgPV1zgfA59JncklEbIPi/SbWWNl+RJG0yP9qo91H+f+PMSvEIC8ivhsRJ9LTXWT3G2kHlwL9EfF0RBwDNgNLmpyncYmIgxHxo7T8C+BJYEZzc9U+IuLJiNhXZtUSYHNE/Doifgb0k9WXdtX2dX0kEfF94Miw5CXAxrS8EXh/QzNVPIWsQyO0oa4/9deKdaru3/s426slwKbI7ALOljS9jvmopGi/iTUzQj+iSFrxf7Wuxvn/UVYhBnnDfBz4drMzMUYzgP255wdo4wGSpAuAtwEPNTcnhVCoukHxyjMWXRFxMC0fArqamZkCKHwdGtaGuv7UX7PrVADflfSIpBUprVnfe6X3bcZntCqFhm7Ihas2+7uy5vL3PwEtcZ+8sZD0PWBamVV/GRHfStv8JVnoy1camTcDSb8F/BPwJxHxSrPz00rGUnet2CIiJPl+NVbR8DZU0uA615/C+v2IGJB0HrBD0k/yK5v1vTe5vq0HbiEbAN8CrCM7eN/R3I+wiWibQV5EvHek9ZKuBa4GLov2ufnfADAz97w7pbUVSWeQdU6+EhFfb3Z+Ws1odbeCQtSNnKKVZyyekzQ9Ig6m8KbDzc5QmytsHarQhrr+1F9T61REDKTHw5K+QRaS1qzvvdL7NvQziojnSsuS7gG2NiMfrWaC/Ygi6ejvf6IKEa4paTHw58D7IuJXzc7POOwG5kiaLWkK2UXFW5qcp3FRdrj5XuDJiPhss/NTIFuApZLOlDSb7KL3HzY5T9Vo+7o+AVuA5Wl5OeCjrdUpZB0aoQ11/am/ptUpSWdJem1pGVhENoFGs773Su+7BViWZtlcALycC+usuWHX+32A7DMp5aNIv4k2PoVs/+suItr+j+wC3P3Ao+nvC83O0zjyfhXZjKA/JTvt3vQ8jTP/v08WVrEn9/lf1ex8tcsf2Y/YAeDXwHPA9ty6v0z1Yh9wZbPzWoOytnVdH6Vs9wMHgePp+7wOeAPZLHVPAd8Dzm12Ptv9r4h1qFIb6vrTsM+/KXUK+B3g39Lf3tJ7N+J7H097BYhsVsOfAo8B8+qcjy+l99lD1omfntu+UL+JNfwcK/YjivRXxPZ/lPKe9v8x3n0o7cjMzMzMzMwKoBDhmmZmZmZmZpbxIM/MzMzMzKxAPMgzMzMzMzMrEA/yzMzMzMzMCsSDPDMzMzMzswLxIM/MzMzMzKxAPMgzMzMzMzMrkP8PQXrdOKAdBuQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x1080 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check data distributions\n",
    "train_df.hist(bins=50, figsize=(15, 15), edgecolor=\"black\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 135
    },
    "id": "nRKl0iPON_3d",
    "outputId": "20128715-be4f-4b65-cfe1-38d0b59b9ecd"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check boxplots\n",
    "plt.figure(figsize=(30,5))\n",
    "sns.boxplot(data=train_df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "9Kg2L100OuuZ",
    "outputId": "21ea9a5b-e8bf-4085-ed65-4df14cf15c30"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_78b9c_row0_col0, #T_78b9c_row1_col1, #T_78b9c_row2_col2, #T_78b9c_row3_col3, #T_78b9c_row4_col4, #T_78b9c_row5_col5, #T_78b9c_row6_col6, #T_78b9c_row7_col7, #T_78b9c_row8_col8, #T_78b9c_row9_col9, #T_78b9c_row10_col10, #T_78b9c_row11_col11, #T_78b9c_row12_col12, #T_78b9c_row13_col13, #T_78b9c_row14_col14, #T_78b9c_row15_col15, #T_78b9c_row16_col16, #T_78b9c_row17_col17, #T_78b9c_row18_col18, #T_78b9c_row19_col19, #T_78b9c_row20_col20, #T_78b9c_row21_col21, #T_78b9c_row22_col22, #T_78b9c_row23_col23, #T_78b9c_row24_col24, #T_78b9c_row25_col25, #T_78b9c_row26_col26, #T_78b9c_row27_col27, #T_78b9c_row28_col28, #T_78b9c_row29_col29 {\n",
       "  color: blue;\n",
       "}\n",
       "#T_78b9c_row0_col1, #T_78b9c_row0_col3, #T_78b9c_row0_col7, #T_78b9c_row0_col10, #T_78b9c_row0_col12, #T_78b9c_row0_col19, #T_78b9c_row0_col20, #T_78b9c_row0_col22, #T_78b9c_row0_col24, #T_78b9c_row0_col26, #T_78b9c_row0_col29, #T_78b9c_row1_col0, #T_78b9c_row1_col2, #T_78b9c_row1_col4, #T_78b9c_row1_col5, #T_78b9c_row1_col6, #T_78b9c_row1_col7, #T_78b9c_row1_col8, #T_78b9c_row1_col9, #T_78b9c_row1_col11, #T_78b9c_row1_col12, #T_78b9c_row1_col13, #T_78b9c_row1_col15, #T_78b9c_row1_col16, #T_78b9c_row1_col17, #T_78b9c_row1_col19, #T_78b9c_row1_col20, #T_78b9c_row1_col21, #T_78b9c_row1_col22, #T_78b9c_row1_col25, #T_78b9c_row1_col26, #T_78b9c_row2_col1, #T_78b9c_row2_col3, #T_78b9c_row2_col7, #T_78b9c_row2_col10, #T_78b9c_row2_col19, #T_78b9c_row2_col21, #T_78b9c_row2_col22, #T_78b9c_row2_col23, #T_78b9c_row2_col24, #T_78b9c_row2_col25, #T_78b9c_row2_col26, #T_78b9c_row3_col0, #T_78b9c_row3_col2, #T_78b9c_row3_col4, #T_78b9c_row3_col5, #T_78b9c_row3_col6, #T_78b9c_row3_col7, #T_78b9c_row3_col8, #T_78b9c_row3_col9, #T_78b9c_row3_col11, #T_78b9c_row3_col12, #T_78b9c_row3_col13, #T_78b9c_row3_col14, #T_78b9c_row3_col15, #T_78b9c_row3_col16, #T_78b9c_row3_col17, #T_78b9c_row3_col18, #T_78b9c_row3_col23, #T_78b9c_row3_col24, #T_78b9c_row3_col25, #T_78b9c_row3_col27, #T_78b9c_row3_col28, #T_78b9c_row4_col1, #T_78b9c_row4_col3, #T_78b9c_row4_col10, #T_78b9c_row4_col12, #T_78b9c_row4_col14, #T_78b9c_row4_col18, #T_78b9c_row4_col21, #T_78b9c_row4_col23, #T_78b9c_row4_col26, #T_78b9c_row4_col29, #T_78b9c_row5_col1, #T_78b9c_row5_col3, #T_78b9c_row5_col7, #T_78b9c_row5_col10, #T_78b9c_row5_col12, #T_78b9c_row5_col14, #T_78b9c_row5_col18, #T_78b9c_row5_col19, #T_78b9c_row5_col20, #T_78b9c_row5_col22, #T_78b9c_row5_col27, #T_78b9c_row5_col28, #T_78b9c_row6_col1, #T_78b9c_row6_col3, #T_78b9c_row6_col7, #T_78b9c_row6_col10, #T_78b9c_row6_col18, #T_78b9c_row6_col19, #T_78b9c_row6_col20, #T_78b9c_row6_col22, #T_78b9c_row6_col24, #T_78b9c_row6_col25, #T_78b9c_row6_col27, #T_78b9c_row6_col28, #T_78b9c_row6_col29, #T_78b9c_row7_col0, #T_78b9c_row7_col1, #T_78b9c_row7_col2, #T_78b9c_row7_col3, #T_78b9c_row7_col5, #T_78b9c_row7_col6, #T_78b9c_row7_col11, #T_78b9c_row7_col12, #T_78b9c_row7_col13, #T_78b9c_row7_col15, #T_78b9c_row7_col16, #T_78b9c_row7_col17, #T_78b9c_row7_col20, #T_78b9c_row7_col21, #T_78b9c_row7_col22, #T_78b9c_row7_col23, #T_78b9c_row7_col25, #T_78b9c_row8_col1, #T_78b9c_row8_col3, #T_78b9c_row8_col10, #T_78b9c_row8_col18, #T_78b9c_row8_col23, #T_78b9c_row8_col25, #T_78b9c_row8_col27, #T_78b9c_row9_col1, #T_78b9c_row9_col3, #T_78b9c_row9_col10, #T_78b9c_row9_col14, #T_78b9c_row9_col18, #T_78b9c_row9_col19, #T_78b9c_row9_col21, #T_78b9c_row9_col23, #T_78b9c_row9_col27, #T_78b9c_row9_col29, #T_78b9c_row10_col0, #T_78b9c_row10_col2, #T_78b9c_row10_col4, #T_78b9c_row10_col5, #T_78b9c_row10_col6, #T_78b9c_row10_col8, #T_78b9c_row10_col9, #T_78b9c_row10_col11, #T_78b9c_row10_col13, #T_78b9c_row10_col14, #T_78b9c_row10_col15, #T_78b9c_row10_col16, #T_78b9c_row10_col17, #T_78b9c_row10_col21, #T_78b9c_row10_col25, #T_78b9c_row11_col1, #T_78b9c_row11_col3, #T_78b9c_row11_col7, #T_78b9c_row11_col10, #T_78b9c_row11_col18, #T_78b9c_row11_col20, #T_78b9c_row11_col21, #T_78b9c_row11_col23, #T_78b9c_row11_col25, #T_78b9c_row11_col27, #T_78b9c_row11_col29, #T_78b9c_row12_col0, #T_78b9c_row12_col1, #T_78b9c_row12_col3, #T_78b9c_row12_col4, #T_78b9c_row12_col5, #T_78b9c_row12_col7, #T_78b9c_row12_col14, #T_78b9c_row12_col15, #T_78b9c_row12_col16, #T_78b9c_row12_col17, #T_78b9c_row12_col21, #T_78b9c_row12_col27, #T_78b9c_row12_col28, #T_78b9c_row13_col1, #T_78b9c_row13_col3, #T_78b9c_row13_col7, #T_78b9c_row13_col10, #T_78b9c_row13_col14, #T_78b9c_row13_col18, #T_78b9c_row13_col20, #T_78b9c_row13_col22, #T_78b9c_row13_col24, #T_78b9c_row13_col26, #T_78b9c_row13_col27, #T_78b9c_row13_col28, #T_78b9c_row14_col3, #T_78b9c_row14_col4, #T_78b9c_row14_col5, #T_78b9c_row14_col9, #T_78b9c_row14_col10, #T_78b9c_row14_col12, #T_78b9c_row14_col13, #T_78b9c_row14_col17, #T_78b9c_row14_col18, #T_78b9c_row14_col20, #T_78b9c_row14_col23, #T_78b9c_row14_col24, #T_78b9c_row14_col25, #T_78b9c_row14_col27, #T_78b9c_row15_col1, #T_78b9c_row15_col3, #T_78b9c_row15_col7, #T_78b9c_row15_col10, #T_78b9c_row15_col12, #T_78b9c_row15_col18, #T_78b9c_row15_col19, #T_78b9c_row15_col20, #T_78b9c_row15_col23, #T_78b9c_row15_col25, #T_78b9c_row15_col27, #T_78b9c_row15_col29, #T_78b9c_row16_col1, #T_78b9c_row16_col3, #T_78b9c_row16_col7, #T_78b9c_row16_col10, #T_78b9c_row16_col12, #T_78b9c_row16_col18, #T_78b9c_row16_col20, #T_78b9c_row16_col21, #T_78b9c_row16_col22, #T_78b9c_row16_col25, #T_78b9c_row16_col28, #T_78b9c_row17_col1, #T_78b9c_row17_col3, #T_78b9c_row17_col7, #T_78b9c_row17_col10, #T_78b9c_row17_col12, #T_78b9c_row17_col14, #T_78b9c_row17_col18, #T_78b9c_row17_col19, #T_78b9c_row17_col20, #T_78b9c_row17_col22, #T_78b9c_row17_col24, #T_78b9c_row17_col27, #T_78b9c_row17_col28, #T_78b9c_row17_col29, #T_78b9c_row18_col3, #T_78b9c_row18_col4, #T_78b9c_row18_col5, #T_78b9c_row18_col6, #T_78b9c_row18_col8, #T_78b9c_row18_col9, #T_78b9c_row18_col11, #T_78b9c_row18_col13, #T_78b9c_row18_col14, #T_78b9c_row18_col15, #T_78b9c_row18_col16, #T_78b9c_row18_col17, #T_78b9c_row18_col19, #T_78b9c_row18_col20, #T_78b9c_row18_col23, #T_78b9c_row18_col24, #T_78b9c_row18_col25, #T_78b9c_row18_col27, #T_78b9c_row18_col29, #T_78b9c_row19_col0, #T_78b9c_row19_col1, #T_78b9c_row19_col2, #T_78b9c_row19_col5, #T_78b9c_row19_col6, #T_78b9c_row19_col9, #T_78b9c_row19_col15, #T_78b9c_row19_col17, #T_78b9c_row19_col18, #T_78b9c_row19_col21, #T_78b9c_row19_col22, #T_78b9c_row19_col24, #T_78b9c_row19_col26, #T_78b9c_row19_col28, #T_78b9c_row20_col0, #T_78b9c_row20_col1, #T_78b9c_row20_col5, #T_78b9c_row20_col6, #T_78b9c_row20_col7, #T_78b9c_row20_col11, #T_78b9c_row20_col13, #T_78b9c_row20_col14, #T_78b9c_row20_col15, #T_78b9c_row20_col16, #T_78b9c_row20_col17, #T_78b9c_row20_col18, #T_78b9c_row20_col21, #T_78b9c_row20_col22, #T_78b9c_row20_col23, #T_78b9c_row20_col24, #T_78b9c_row20_col25, #T_78b9c_row20_col26, #T_78b9c_row20_col28, #T_78b9c_row20_col29, #T_78b9c_row21_col1, #T_78b9c_row21_col2, #T_78b9c_row21_col4, #T_78b9c_row21_col7, #T_78b9c_row21_col9, #T_78b9c_row21_col10, #T_78b9c_row21_col11, #T_78b9c_row21_col12, #T_78b9c_row21_col16, #T_78b9c_row21_col19, #T_78b9c_row21_col20, #T_78b9c_row21_col23, #T_78b9c_row21_col24, #T_78b9c_row21_col27, #T_78b9c_row21_col29, #T_78b9c_row22_col0, #T_78b9c_row22_col1, #T_78b9c_row22_col2, #T_78b9c_row22_col5, #T_78b9c_row22_col6, #T_78b9c_row22_col7, #T_78b9c_row22_col13, #T_78b9c_row22_col16, #T_78b9c_row22_col17, #T_78b9c_row22_col19, #T_78b9c_row22_col20, #T_78b9c_row22_col24, #T_78b9c_row22_col25, #T_78b9c_row22_col27, #T_78b9c_row22_col29, #T_78b9c_row23_col2, #T_78b9c_row23_col3, #T_78b9c_row23_col4, #T_78b9c_row23_col7, #T_78b9c_row23_col8, #T_78b9c_row23_col9, #T_78b9c_row23_col11, #T_78b9c_row23_col14, #T_78b9c_row23_col15, #T_78b9c_row23_col18, #T_78b9c_row23_col20, #T_78b9c_row23_col21, #T_78b9c_row23_col25, #T_78b9c_row23_col27, #T_78b9c_row23_col28, #T_78b9c_row24_col0, #T_78b9c_row24_col2, #T_78b9c_row24_col3, #T_78b9c_row24_col6, #T_78b9c_row24_col13, #T_78b9c_row24_col14, #T_78b9c_row24_col17, #T_78b9c_row24_col18, #T_78b9c_row24_col19, #T_78b9c_row24_col20, #T_78b9c_row24_col21, #T_78b9c_row24_col22, #T_78b9c_row24_col25, #T_78b9c_row24_col27, #T_78b9c_row25_col1, #T_78b9c_row25_col2, #T_78b9c_row25_col3, #T_78b9c_row25_col6, #T_78b9c_row25_col7, #T_78b9c_row25_col8, #T_78b9c_row25_col10, #T_78b9c_row25_col11, #T_78b9c_row25_col14, #T_78b9c_row25_col15, #T_78b9c_row25_col16, #T_78b9c_row25_col18, #T_78b9c_row25_col20, #T_78b9c_row25_col22, #T_78b9c_row25_col23, #T_78b9c_row25_col24, #T_78b9c_row26_col0, #T_78b9c_row26_col1, #T_78b9c_row26_col2, #T_78b9c_row26_col4, #T_78b9c_row26_col13, #T_78b9c_row26_col19, #T_78b9c_row26_col20, #T_78b9c_row26_col28, #T_78b9c_row27_col3, #T_78b9c_row27_col5, #T_78b9c_row27_col6, #T_78b9c_row27_col8, #T_78b9c_row27_col9, #T_78b9c_row27_col11, #T_78b9c_row27_col12, #T_78b9c_row27_col13, #T_78b9c_row27_col14, #T_78b9c_row27_col15, #T_78b9c_row27_col17, #T_78b9c_row27_col18, #T_78b9c_row27_col21, #T_78b9c_row27_col22, #T_78b9c_row27_col23, #T_78b9c_row27_col24, #T_78b9c_row27_col28, #T_78b9c_row28_col3, #T_78b9c_row28_col5, #T_78b9c_row28_col6, #T_78b9c_row28_col12, #T_78b9c_row28_col13, #T_78b9c_row28_col16, #T_78b9c_row28_col17, #T_78b9c_row28_col19, #T_78b9c_row28_col20, #T_78b9c_row28_col23, #T_78b9c_row28_col26, #T_78b9c_row28_col27, #T_78b9c_row29_col0, #T_78b9c_row29_col4, #T_78b9c_row29_col6, #T_78b9c_row29_col9, #T_78b9c_row29_col11, #T_78b9c_row29_col15, #T_78b9c_row29_col17, #T_78b9c_row29_col18, #T_78b9c_row29_col20, #T_78b9c_row29_col21, #T_78b9c_row29_col22 {\n",
       "  color: green;\n",
       "}\n",
       "#T_78b9c_row0_col2, #T_78b9c_row0_col4, #T_78b9c_row0_col5, #T_78b9c_row0_col6, #T_78b9c_row0_col8, #T_78b9c_row0_col9, #T_78b9c_row0_col11, #T_78b9c_row0_col13, #T_78b9c_row0_col14, #T_78b9c_row0_col15, #T_78b9c_row0_col16, #T_78b9c_row0_col17, #T_78b9c_row0_col18, #T_78b9c_row0_col21, #T_78b9c_row0_col23, #T_78b9c_row0_col25, #T_78b9c_row0_col27, #T_78b9c_row0_col28, #T_78b9c_row1_col3, #T_78b9c_row1_col10, #T_78b9c_row1_col14, #T_78b9c_row1_col18, #T_78b9c_row1_col23, #T_78b9c_row1_col24, #T_78b9c_row1_col27, #T_78b9c_row1_col28, #T_78b9c_row1_col29, #T_78b9c_row2_col0, #T_78b9c_row2_col4, #T_78b9c_row2_col5, #T_78b9c_row2_col6, #T_78b9c_row2_col8, #T_78b9c_row2_col9, #T_78b9c_row2_col11, #T_78b9c_row2_col12, #T_78b9c_row2_col13, #T_78b9c_row2_col14, #T_78b9c_row2_col15, #T_78b9c_row2_col16, #T_78b9c_row2_col17, #T_78b9c_row2_col18, #T_78b9c_row2_col20, #T_78b9c_row2_col27, #T_78b9c_row2_col28, #T_78b9c_row2_col29, #T_78b9c_row3_col1, #T_78b9c_row3_col10, #T_78b9c_row3_col19, #T_78b9c_row3_col20, #T_78b9c_row3_col21, #T_78b9c_row3_col22, #T_78b9c_row3_col26, #T_78b9c_row3_col29, #T_78b9c_row4_col0, #T_78b9c_row4_col2, #T_78b9c_row4_col5, #T_78b9c_row4_col6, #T_78b9c_row4_col7, #T_78b9c_row4_col8, #T_78b9c_row4_col9, #T_78b9c_row4_col11, #T_78b9c_row4_col13, #T_78b9c_row4_col15, #T_78b9c_row4_col16, #T_78b9c_row4_col17, #T_78b9c_row4_col19, #T_78b9c_row4_col20, #T_78b9c_row4_col22, #T_78b9c_row4_col24, #T_78b9c_row4_col25, #T_78b9c_row4_col27, #T_78b9c_row4_col28, #T_78b9c_row5_col0, #T_78b9c_row5_col2, #T_78b9c_row5_col4, #T_78b9c_row5_col6, #T_78b9c_row5_col8, #T_78b9c_row5_col9, #T_78b9c_row5_col11, #T_78b9c_row5_col13, #T_78b9c_row5_col15, #T_78b9c_row5_col16, #T_78b9c_row5_col17, #T_78b9c_row5_col21, #T_78b9c_row5_col23, #T_78b9c_row5_col24, #T_78b9c_row5_col25, #T_78b9c_row5_col26, #T_78b9c_row5_col29, #T_78b9c_row6_col0, #T_78b9c_row6_col2, #T_78b9c_row6_col4, #T_78b9c_row6_col5, #T_78b9c_row6_col8, #T_78b9c_row6_col9, #T_78b9c_row6_col11, #T_78b9c_row6_col12, #T_78b9c_row6_col13, #T_78b9c_row6_col14, #T_78b9c_row6_col15, #T_78b9c_row6_col16, #T_78b9c_row6_col17, #T_78b9c_row6_col21, #T_78b9c_row6_col23, #T_78b9c_row6_col26, #T_78b9c_row7_col4, #T_78b9c_row7_col8, #T_78b9c_row7_col9, #T_78b9c_row7_col10, #T_78b9c_row7_col14, #T_78b9c_row7_col18, #T_78b9c_row7_col19, #T_78b9c_row7_col24, #T_78b9c_row7_col26, #T_78b9c_row7_col27, #T_78b9c_row7_col28, #T_78b9c_row7_col29, #T_78b9c_row8_col0, #T_78b9c_row8_col2, #T_78b9c_row8_col4, #T_78b9c_row8_col5, #T_78b9c_row8_col6, #T_78b9c_row8_col7, #T_78b9c_row8_col9, #T_78b9c_row8_col11, #T_78b9c_row8_col12, #T_78b9c_row8_col13, #T_78b9c_row8_col14, #T_78b9c_row8_col15, #T_78b9c_row8_col16, #T_78b9c_row8_col17, #T_78b9c_row8_col19, #T_78b9c_row8_col20, #T_78b9c_row8_col21, #T_78b9c_row8_col22, #T_78b9c_row8_col24, #T_78b9c_row8_col26, #T_78b9c_row8_col28, #T_78b9c_row8_col29, #T_78b9c_row9_col0, #T_78b9c_row9_col2, #T_78b9c_row9_col4, #T_78b9c_row9_col5, #T_78b9c_row9_col6, #T_78b9c_row9_col7, #T_78b9c_row9_col8, #T_78b9c_row9_col11, #T_78b9c_row9_col12, #T_78b9c_row9_col13, #T_78b9c_row9_col15, #T_78b9c_row9_col16, #T_78b9c_row9_col17, #T_78b9c_row9_col20, #T_78b9c_row9_col22, #T_78b9c_row9_col24, #T_78b9c_row9_col25, #T_78b9c_row9_col26, #T_78b9c_row9_col28, #T_78b9c_row10_col1, #T_78b9c_row10_col3, #T_78b9c_row10_col7, #T_78b9c_row10_col12, #T_78b9c_row10_col18, #T_78b9c_row10_col19, #T_78b9c_row10_col20, #T_78b9c_row10_col22, #T_78b9c_row10_col23, #T_78b9c_row10_col24, #T_78b9c_row10_col26, #T_78b9c_row10_col27, #T_78b9c_row10_col28, #T_78b9c_row10_col29, #T_78b9c_row11_col0, #T_78b9c_row11_col2, #T_78b9c_row11_col4, #T_78b9c_row11_col5, #T_78b9c_row11_col6, #T_78b9c_row11_col8, #T_78b9c_row11_col9, #T_78b9c_row11_col12, #T_78b9c_row11_col13, #T_78b9c_row11_col14, #T_78b9c_row11_col15, #T_78b9c_row11_col16, #T_78b9c_row11_col17, #T_78b9c_row11_col19, #T_78b9c_row11_col22, #T_78b9c_row11_col24, #T_78b9c_row11_col26, #T_78b9c_row11_col28, #T_78b9c_row12_col2, #T_78b9c_row12_col6, #T_78b9c_row12_col8, #T_78b9c_row12_col9, #T_78b9c_row12_col10, #T_78b9c_row12_col11, #T_78b9c_row12_col13, #T_78b9c_row12_col18, #T_78b9c_row12_col19, #T_78b9c_row12_col20, #T_78b9c_row12_col22, #T_78b9c_row12_col23, #T_78b9c_row12_col24, #T_78b9c_row12_col25, #T_78b9c_row12_col26, #T_78b9c_row12_col29, #T_78b9c_row13_col0, #T_78b9c_row13_col2, #T_78b9c_row13_col4, #T_78b9c_row13_col5, #T_78b9c_row13_col6, #T_78b9c_row13_col8, #T_78b9c_row13_col9, #T_78b9c_row13_col11, #T_78b9c_row13_col12, #T_78b9c_row13_col15, #T_78b9c_row13_col16, #T_78b9c_row13_col17, #T_78b9c_row13_col19, #T_78b9c_row13_col21, #T_78b9c_row13_col23, #T_78b9c_row13_col25, #T_78b9c_row13_col29, #T_78b9c_row14_col0, #T_78b9c_row14_col1, #T_78b9c_row14_col2, #T_78b9c_row14_col6, #T_78b9c_row14_col7, #T_78b9c_row14_col8, #T_78b9c_row14_col11, #T_78b9c_row14_col15, #T_78b9c_row14_col16, #T_78b9c_row14_col19, #T_78b9c_row14_col21, #T_78b9c_row14_col22, #T_78b9c_row14_col26, #T_78b9c_row14_col28, #T_78b9c_row14_col29, #T_78b9c_row15_col0, #T_78b9c_row15_col2, #T_78b9c_row15_col4, #T_78b9c_row15_col5, #T_78b9c_row15_col6, #T_78b9c_row15_col8, #T_78b9c_row15_col9, #T_78b9c_row15_col11, #T_78b9c_row15_col13, #T_78b9c_row15_col14, #T_78b9c_row15_col16, #T_78b9c_row15_col17, #T_78b9c_row15_col21, #T_78b9c_row15_col22, #T_78b9c_row15_col24, #T_78b9c_row15_col26, #T_78b9c_row15_col28, #T_78b9c_row16_col0, #T_78b9c_row16_col2, #T_78b9c_row16_col4, #T_78b9c_row16_col5, #T_78b9c_row16_col6, #T_78b9c_row16_col8, #T_78b9c_row16_col9, #T_78b9c_row16_col11, #T_78b9c_row16_col13, #T_78b9c_row16_col14, #T_78b9c_row16_col15, #T_78b9c_row16_col17, #T_78b9c_row16_col19, #T_78b9c_row16_col23, #T_78b9c_row16_col24, #T_78b9c_row16_col26, #T_78b9c_row16_col27, #T_78b9c_row16_col29, #T_78b9c_row17_col0, #T_78b9c_row17_col2, #T_78b9c_row17_col4, #T_78b9c_row17_col5, #T_78b9c_row17_col6, #T_78b9c_row17_col8, #T_78b9c_row17_col9, #T_78b9c_row17_col11, #T_78b9c_row17_col13, #T_78b9c_row17_col15, #T_78b9c_row17_col16, #T_78b9c_row17_col21, #T_78b9c_row17_col23, #T_78b9c_row17_col25, #T_78b9c_row17_col26, #T_78b9c_row18_col0, #T_78b9c_row18_col1, #T_78b9c_row18_col2, #T_78b9c_row18_col7, #T_78b9c_row18_col10, #T_78b9c_row18_col12, #T_78b9c_row18_col21, #T_78b9c_row18_col22, #T_78b9c_row18_col26, #T_78b9c_row18_col28, #T_78b9c_row19_col3, #T_78b9c_row19_col4, #T_78b9c_row19_col7, #T_78b9c_row19_col8, #T_78b9c_row19_col10, #T_78b9c_row19_col11, #T_78b9c_row19_col12, #T_78b9c_row19_col13, #T_78b9c_row19_col14, #T_78b9c_row19_col16, #T_78b9c_row19_col20, #T_78b9c_row19_col23, #T_78b9c_row19_col25, #T_78b9c_row19_col27, #T_78b9c_row19_col29, #T_78b9c_row20_col2, #T_78b9c_row20_col3, #T_78b9c_row20_col4, #T_78b9c_row20_col8, #T_78b9c_row20_col9, #T_78b9c_row20_col10, #T_78b9c_row20_col12, #T_78b9c_row20_col19, #T_78b9c_row20_col27, #T_78b9c_row21_col0, #T_78b9c_row21_col3, #T_78b9c_row21_col5, #T_78b9c_row21_col6, #T_78b9c_row21_col8, #T_78b9c_row21_col13, #T_78b9c_row21_col14, #T_78b9c_row21_col15, #T_78b9c_row21_col17, #T_78b9c_row21_col18, #T_78b9c_row21_col22, #T_78b9c_row21_col25, #T_78b9c_row21_col26, #T_78b9c_row21_col28, #T_78b9c_row22_col3, #T_78b9c_row22_col4, #T_78b9c_row22_col8, #T_78b9c_row22_col9, #T_78b9c_row22_col10, #T_78b9c_row22_col11, #T_78b9c_row22_col12, #T_78b9c_row22_col14, #T_78b9c_row22_col15, #T_78b9c_row22_col18, #T_78b9c_row22_col21, #T_78b9c_row22_col23, #T_78b9c_row22_col26, #T_78b9c_row22_col28, #T_78b9c_row23_col0, #T_78b9c_row23_col1, #T_78b9c_row23_col5, #T_78b9c_row23_col6, #T_78b9c_row23_col10, #T_78b9c_row23_col12, #T_78b9c_row23_col13, #T_78b9c_row23_col16, #T_78b9c_row23_col17, #T_78b9c_row23_col19, #T_78b9c_row23_col22, #T_78b9c_row23_col24, #T_78b9c_row23_col26, #T_78b9c_row23_col29, #T_78b9c_row24_col1, #T_78b9c_row24_col4, #T_78b9c_row24_col5, #T_78b9c_row24_col7, #T_78b9c_row24_col8, #T_78b9c_row24_col9, #T_78b9c_row24_col10, #T_78b9c_row24_col11, #T_78b9c_row24_col12, #T_78b9c_row24_col15, #T_78b9c_row24_col16, #T_78b9c_row24_col23, #T_78b9c_row24_col26, #T_78b9c_row24_col28, #T_78b9c_row24_col29, #T_78b9c_row25_col0, #T_78b9c_row25_col4, #T_78b9c_row25_col5, #T_78b9c_row25_col9, #T_78b9c_row25_col12, #T_78b9c_row25_col13, #T_78b9c_row25_col17, #T_78b9c_row25_col19, #T_78b9c_row25_col21, #T_78b9c_row25_col26, #T_78b9c_row25_col27, #T_78b9c_row25_col28, #T_78b9c_row25_col29, #T_78b9c_row26_col3, #T_78b9c_row26_col5, #T_78b9c_row26_col6, #T_78b9c_row26_col7, #T_78b9c_row26_col8, #T_78b9c_row26_col9, #T_78b9c_row26_col10, #T_78b9c_row26_col11, #T_78b9c_row26_col12, #T_78b9c_row26_col14, #T_78b9c_row26_col15, #T_78b9c_row26_col16, #T_78b9c_row26_col17, #T_78b9c_row26_col18, #T_78b9c_row26_col21, #T_78b9c_row26_col22, #T_78b9c_row26_col23, #T_78b9c_row26_col24, #T_78b9c_row26_col25, #T_78b9c_row26_col27, #T_78b9c_row26_col29, #T_78b9c_row27_col0, #T_78b9c_row27_col1, #T_78b9c_row27_col2, #T_78b9c_row27_col4, #T_78b9c_row27_col7, #T_78b9c_row27_col10, #T_78b9c_row27_col16, #T_78b9c_row27_col19, #T_78b9c_row27_col20, #T_78b9c_row27_col25, #T_78b9c_row27_col26, #T_78b9c_row27_col29, #T_78b9c_row28_col0, #T_78b9c_row28_col1, #T_78b9c_row28_col2, #T_78b9c_row28_col4, #T_78b9c_row28_col7, #T_78b9c_row28_col8, #T_78b9c_row28_col9, #T_78b9c_row28_col10, #T_78b9c_row28_col11, #T_78b9c_row28_col14, #T_78b9c_row28_col15, #T_78b9c_row28_col18, #T_78b9c_row28_col21, #T_78b9c_row28_col22, #T_78b9c_row28_col24, #T_78b9c_row28_col25, #T_78b9c_row28_col29, #T_78b9c_row29_col1, #T_78b9c_row29_col2, #T_78b9c_row29_col3, #T_78b9c_row29_col5, #T_78b9c_row29_col7, #T_78b9c_row29_col8, #T_78b9c_row29_col10, #T_78b9c_row29_col12, #T_78b9c_row29_col13, #T_78b9c_row29_col14, #T_78b9c_row29_col16, #T_78b9c_row29_col19, #T_78b9c_row29_col23, #T_78b9c_row29_col24, #T_78b9c_row29_col25, #T_78b9c_row29_col26, #T_78b9c_row29_col27, #T_78b9c_row29_col28 {\n",
       "  color: red;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_78b9c_\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th class=\"col_heading level0 col0\" >V1</th>\n",
       "      <th class=\"col_heading level0 col1\" >V2</th>\n",
       "      <th class=\"col_heading level0 col2\" >V3</th>\n",
       "      <th class=\"col_heading level0 col3\" >V4</th>\n",
       "      <th class=\"col_heading level0 col4\" >V5</th>\n",
       "      <th class=\"col_heading level0 col5\" >V6</th>\n",
       "      <th class=\"col_heading level0 col6\" >V7</th>\n",
       "      <th class=\"col_heading level0 col7\" >V8</th>\n",
       "      <th class=\"col_heading level0 col8\" >V9</th>\n",
       "      <th class=\"col_heading level0 col9\" >V10</th>\n",
       "      <th class=\"col_heading level0 col10\" >V11</th>\n",
       "      <th class=\"col_heading level0 col11\" >V12</th>\n",
       "      <th class=\"col_heading level0 col12\" >V13</th>\n",
       "      <th class=\"col_heading level0 col13\" >V14</th>\n",
       "      <th class=\"col_heading level0 col14\" >V15</th>\n",
       "      <th class=\"col_heading level0 col15\" >V16</th>\n",
       "      <th class=\"col_heading level0 col16\" >V17</th>\n",
       "      <th class=\"col_heading level0 col17\" >V18</th>\n",
       "      <th class=\"col_heading level0 col18\" >V19</th>\n",
       "      <th class=\"col_heading level0 col19\" >V20</th>\n",
       "      <th class=\"col_heading level0 col20\" >V21</th>\n",
       "      <th class=\"col_heading level0 col21\" >V22</th>\n",
       "      <th class=\"col_heading level0 col22\" >V23</th>\n",
       "      <th class=\"col_heading level0 col23\" >V24</th>\n",
       "      <th class=\"col_heading level0 col24\" >V25</th>\n",
       "      <th class=\"col_heading level0 col25\" >V26</th>\n",
       "      <th class=\"col_heading level0 col26\" >V27</th>\n",
       "      <th class=\"col_heading level0 col27\" >V28</th>\n",
       "      <th class=\"col_heading level0 col28\" >V29</th>\n",
       "      <th class=\"col_heading level0 col29\" >V30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row0\" class=\"row_heading level0 row0\" >V1</th>\n",
       "      <td id=\"T_78b9c_row0_col0\" class=\"data row0 col0\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row0_col1\" class=\"data row0 col1\" >0.015507</td>\n",
       "      <td id=\"T_78b9c_row0_col2\" class=\"data row0 col2\" >-0.015559</td>\n",
       "      <td id=\"T_78b9c_row0_col3\" class=\"data row0 col3\" >0.005334</td>\n",
       "      <td id=\"T_78b9c_row0_col4\" class=\"data row0 col4\" >-0.009902</td>\n",
       "      <td id=\"T_78b9c_row0_col5\" class=\"data row0 col5\" >-0.000177</td>\n",
       "      <td id=\"T_78b9c_row0_col6\" class=\"data row0 col6\" >-0.021839</td>\n",
       "      <td id=\"T_78b9c_row0_col7\" class=\"data row0 col7\" >0.003626</td>\n",
       "      <td id=\"T_78b9c_row0_col8\" class=\"data row0 col8\" >-0.009599</td>\n",
       "      <td id=\"T_78b9c_row0_col9\" class=\"data row0 col9\" >-0.019278</td>\n",
       "      <td id=\"T_78b9c_row0_col10\" class=\"data row0 col10\" >0.003578</td>\n",
       "      <td id=\"T_78b9c_row0_col11\" class=\"data row0 col11\" >-0.011240</td>\n",
       "      <td id=\"T_78b9c_row0_col12\" class=\"data row0 col12\" >0.003362</td>\n",
       "      <td id=\"T_78b9c_row0_col13\" class=\"data row0 col13\" >-0.010310</td>\n",
       "      <td id=\"T_78b9c_row0_col14\" class=\"data row0 col14\" >-0.002944</td>\n",
       "      <td id=\"T_78b9c_row0_col15\" class=\"data row0 col15\" >-0.014766</td>\n",
       "      <td id=\"T_78b9c_row0_col16\" class=\"data row0 col16\" >-0.023218</td>\n",
       "      <td id=\"T_78b9c_row0_col17\" class=\"data row0 col17\" >-0.007048</td>\n",
       "      <td id=\"T_78b9c_row0_col18\" class=\"data row0 col18\" >-0.000303</td>\n",
       "      <td id=\"T_78b9c_row0_col19\" class=\"data row0 col19\" >0.016057</td>\n",
       "      <td id=\"T_78b9c_row0_col20\" class=\"data row0 col20\" >0.000829</td>\n",
       "      <td id=\"T_78b9c_row0_col21\" class=\"data row0 col21\" >-0.001097</td>\n",
       "      <td id=\"T_78b9c_row0_col22\" class=\"data row0 col22\" >0.017551</td>\n",
       "      <td id=\"T_78b9c_row0_col23\" class=\"data row0 col23\" >-0.001289</td>\n",
       "      <td id=\"T_78b9c_row0_col24\" class=\"data row0 col24\" >0.002329</td>\n",
       "      <td id=\"T_78b9c_row0_col25\" class=\"data row0 col25\" >-0.002013</td>\n",
       "      <td id=\"T_78b9c_row0_col26\" class=\"data row0 col26\" >0.004085</td>\n",
       "      <td id=\"T_78b9c_row0_col27\" class=\"data row0 col27\" >-0.044459</td>\n",
       "      <td id=\"T_78b9c_row0_col28\" class=\"data row0 col28\" >-0.231744</td>\n",
       "      <td id=\"T_78b9c_row0_col29\" class=\"data row0 col29\" >0.118519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row1\" class=\"row_heading level0 row1\" >V2</th>\n",
       "      <td id=\"T_78b9c_row1_col0\" class=\"data row1 col0\" >0.015507</td>\n",
       "      <td id=\"T_78b9c_row1_col1\" class=\"data row1 col1\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row1_col2\" class=\"data row1 col2\" >0.015350</td>\n",
       "      <td id=\"T_78b9c_row1_col3\" class=\"data row1 col3\" >-0.004632</td>\n",
       "      <td id=\"T_78b9c_row1_col4\" class=\"data row1 col4\" >0.005220</td>\n",
       "      <td id=\"T_78b9c_row1_col5\" class=\"data row1 col5\" >0.004119</td>\n",
       "      <td id=\"T_78b9c_row1_col6\" class=\"data row1 col6\" >0.020025</td>\n",
       "      <td id=\"T_78b9c_row1_col7\" class=\"data row1 col7\" >0.002628</td>\n",
       "      <td id=\"T_78b9c_row1_col8\" class=\"data row1 col8\" >0.001865</td>\n",
       "      <td id=\"T_78b9c_row1_col9\" class=\"data row1 col9\" >0.011013</td>\n",
       "      <td id=\"T_78b9c_row1_col10\" class=\"data row1 col10\" >-0.006297</td>\n",
       "      <td id=\"T_78b9c_row1_col11\" class=\"data row1 col11\" >0.013002</td>\n",
       "      <td id=\"T_78b9c_row1_col12\" class=\"data row1 col12\" >0.001395</td>\n",
       "      <td id=\"T_78b9c_row1_col13\" class=\"data row1 col13\" >0.018952</td>\n",
       "      <td id=\"T_78b9c_row1_col14\" class=\"data row1 col14\" >-0.003989</td>\n",
       "      <td id=\"T_78b9c_row1_col15\" class=\"data row1 col15\" >0.005743</td>\n",
       "      <td id=\"T_78b9c_row1_col16\" class=\"data row1 col16\" >0.021561</td>\n",
       "      <td id=\"T_78b9c_row1_col17\" class=\"data row1 col17\" >0.011934</td>\n",
       "      <td id=\"T_78b9c_row1_col18\" class=\"data row1 col18\" >-0.007469</td>\n",
       "      <td id=\"T_78b9c_row1_col19\" class=\"data row1 col19\" >0.022579</td>\n",
       "      <td id=\"T_78b9c_row1_col20\" class=\"data row1 col20\" >0.010194</td>\n",
       "      <td id=\"T_78b9c_row1_col21\" class=\"data row1 col21\" >0.003269</td>\n",
       "      <td id=\"T_78b9c_row1_col22\" class=\"data row1 col22\" >0.015923</td>\n",
       "      <td id=\"T_78b9c_row1_col23\" class=\"data row1 col23\" >-0.004967</td>\n",
       "      <td id=\"T_78b9c_row1_col24\" class=\"data row1 col24\" >-0.003145</td>\n",
       "      <td id=\"T_78b9c_row1_col25\" class=\"data row1 col25\" >0.000796</td>\n",
       "      <td id=\"T_78b9c_row1_col26\" class=\"data row1 col26\" >0.005265</td>\n",
       "      <td id=\"T_78b9c_row1_col27\" class=\"data row1 col27\" >-0.056137</td>\n",
       "      <td id=\"T_78b9c_row1_col28\" class=\"data row1 col28\" >-0.530737</td>\n",
       "      <td id=\"T_78b9c_row1_col29\" class=\"data row1 col29\" >-0.008336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row2\" class=\"row_heading level0 row2\" >V3</th>\n",
       "      <td id=\"T_78b9c_row2_col0\" class=\"data row2 col0\" >-0.015559</td>\n",
       "      <td id=\"T_78b9c_row2_col1\" class=\"data row2 col1\" >0.015350</td>\n",
       "      <td id=\"T_78b9c_row2_col2\" class=\"data row2 col2\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row2_col3\" class=\"data row2 col3\" >0.014562</td>\n",
       "      <td id=\"T_78b9c_row2_col4\" class=\"data row2 col4\" >-0.021869</td>\n",
       "      <td id=\"T_78b9c_row2_col5\" class=\"data row2 col5\" >-0.000392</td>\n",
       "      <td id=\"T_78b9c_row2_col6\" class=\"data row2 col6\" >-0.023424</td>\n",
       "      <td id=\"T_78b9c_row2_col7\" class=\"data row2 col7\" >0.002836</td>\n",
       "      <td id=\"T_78b9c_row2_col8\" class=\"data row2 col8\" >-0.017386</td>\n",
       "      <td id=\"T_78b9c_row2_col9\" class=\"data row2 col9\" >-0.028125</td>\n",
       "      <td id=\"T_78b9c_row2_col10\" class=\"data row2 col10\" >0.014767</td>\n",
       "      <td id=\"T_78b9c_row2_col11\" class=\"data row2 col11\" >-0.031194</td>\n",
       "      <td id=\"T_78b9c_row2_col12\" class=\"data row2 col12\" >-0.000198</td>\n",
       "      <td id=\"T_78b9c_row2_col13\" class=\"data row2 col13\" >-0.027995</td>\n",
       "      <td id=\"T_78b9c_row2_col14\" class=\"data row2 col14\" >-0.000757</td>\n",
       "      <td id=\"T_78b9c_row2_col15\" class=\"data row2 col15\" >-0.025457</td>\n",
       "      <td id=\"T_78b9c_row2_col16\" class=\"data row2 col16\" >-0.043299</td>\n",
       "      <td id=\"T_78b9c_row2_col17\" class=\"data row2 col17\" >-0.014518</td>\n",
       "      <td id=\"T_78b9c_row2_col18\" class=\"data row2 col18\" >-0.000450</td>\n",
       "      <td id=\"T_78b9c_row2_col19\" class=\"data row2 col19\" >0.007484</td>\n",
       "      <td id=\"T_78b9c_row2_col20\" class=\"data row2 col20\" >-0.001319</td>\n",
       "      <td id=\"T_78b9c_row2_col21\" class=\"data row2 col21\" >0.004616</td>\n",
       "      <td id=\"T_78b9c_row2_col22\" class=\"data row2 col22\" >0.001575</td>\n",
       "      <td id=\"T_78b9c_row2_col23\" class=\"data row2 col23\" >0.003254</td>\n",
       "      <td id=\"T_78b9c_row2_col24\" class=\"data row2 col24\" >0.000031</td>\n",
       "      <td id=\"T_78b9c_row2_col25\" class=\"data row2 col25\" >0.000671</td>\n",
       "      <td id=\"T_78b9c_row2_col26\" class=\"data row2 col26\" >0.003680</td>\n",
       "      <td id=\"T_78b9c_row2_col27\" class=\"data row2 col27\" >-0.022925</td>\n",
       "      <td id=\"T_78b9c_row2_col28\" class=\"data row2 col28\" >-0.207185</td>\n",
       "      <td id=\"T_78b9c_row2_col29\" class=\"data row2 col29\" >-0.429787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row3\" class=\"row_heading level0 row3\" >V4</th>\n",
       "      <td id=\"T_78b9c_row3_col0\" class=\"data row3 col0\" >0.005334</td>\n",
       "      <td id=\"T_78b9c_row3_col1\" class=\"data row3 col1\" >-0.004632</td>\n",
       "      <td id=\"T_78b9c_row3_col2\" class=\"data row3 col2\" >0.014562</td>\n",
       "      <td id=\"T_78b9c_row3_col3\" class=\"data row3 col3\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row3_col4\" class=\"data row3 col4\" >0.009886</td>\n",
       "      <td id=\"T_78b9c_row3_col5\" class=\"data row3 col5\" >0.001191</td>\n",
       "      <td id=\"T_78b9c_row3_col6\" class=\"data row3 col6\" >0.011269</td>\n",
       "      <td id=\"T_78b9c_row3_col7\" class=\"data row3 col7\" >0.000271</td>\n",
       "      <td id=\"T_78b9c_row3_col8\" class=\"data row3 col8\" >0.008993</td>\n",
       "      <td id=\"T_78b9c_row3_col9\" class=\"data row3 col9\" >0.015209</td>\n",
       "      <td id=\"T_78b9c_row3_col10\" class=\"data row3 col10\" >-0.011847</td>\n",
       "      <td id=\"T_78b9c_row3_col11\" class=\"data row3 col11\" >0.019518</td>\n",
       "      <td id=\"T_78b9c_row3_col12\" class=\"data row3 col12\" >0.000823</td>\n",
       "      <td id=\"T_78b9c_row3_col13\" class=\"data row3 col13\" >0.017077</td>\n",
       "      <td id=\"T_78b9c_row3_col14\" class=\"data row3 col14\" >0.004899</td>\n",
       "      <td id=\"T_78b9c_row3_col15\" class=\"data row3 col15\" >0.016185</td>\n",
       "      <td id=\"T_78b9c_row3_col16\" class=\"data row3 col16\" >0.029483</td>\n",
       "      <td id=\"T_78b9c_row3_col17\" class=\"data row3 col17\" >0.008506</td>\n",
       "      <td id=\"T_78b9c_row3_col18\" class=\"data row3 col18\" >0.000459</td>\n",
       "      <td id=\"T_78b9c_row3_col19\" class=\"data row3 col19\" >-0.013619</td>\n",
       "      <td id=\"T_78b9c_row3_col20\" class=\"data row3 col20\" >-0.000008</td>\n",
       "      <td id=\"T_78b9c_row3_col21\" class=\"data row3 col21\" >-0.002732</td>\n",
       "      <td id=\"T_78b9c_row3_col22\" class=\"data row3 col22\" >-0.004866</td>\n",
       "      <td id=\"T_78b9c_row3_col23\" class=\"data row3 col23\" >0.001741</td>\n",
       "      <td id=\"T_78b9c_row3_col24\" class=\"data row3 col24\" >0.001838</td>\n",
       "      <td id=\"T_78b9c_row3_col25\" class=\"data row3 col25\" >0.002832</td>\n",
       "      <td id=\"T_78b9c_row3_col26\" class=\"data row3 col26\" >-0.007131</td>\n",
       "      <td id=\"T_78b9c_row3_col27\" class=\"data row3 col27\" >0.025828</td>\n",
       "      <td id=\"T_78b9c_row3_col28\" class=\"data row3 col28\" >0.091913</td>\n",
       "      <td id=\"T_78b9c_row3_col29\" class=\"data row3 col29\" >-0.103017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row4\" class=\"row_heading level0 row4\" >V5</th>\n",
       "      <td id=\"T_78b9c_row4_col0\" class=\"data row4 col0\" >-0.009902</td>\n",
       "      <td id=\"T_78b9c_row4_col1\" class=\"data row4 col1\" >0.005220</td>\n",
       "      <td id=\"T_78b9c_row4_col2\" class=\"data row4 col2\" >-0.021869</td>\n",
       "      <td id=\"T_78b9c_row4_col3\" class=\"data row4 col3\" >0.009886</td>\n",
       "      <td id=\"T_78b9c_row4_col4\" class=\"data row4 col4\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row4_col5\" class=\"data row4 col5\" >-0.000701</td>\n",
       "      <td id=\"T_78b9c_row4_col6\" class=\"data row4 col6\" >-0.012571</td>\n",
       "      <td id=\"T_78b9c_row4_col7\" class=\"data row4 col7\" >-0.000730</td>\n",
       "      <td id=\"T_78b9c_row4_col8\" class=\"data row4 col8\" >-0.006367</td>\n",
       "      <td id=\"T_78b9c_row4_col9\" class=\"data row4 col9\" >-0.015695</td>\n",
       "      <td id=\"T_78b9c_row4_col10\" class=\"data row4 col10\" >0.012430</td>\n",
       "      <td id=\"T_78b9c_row4_col11\" class=\"data row4 col11\" >-0.016962</td>\n",
       "      <td id=\"T_78b9c_row4_col12\" class=\"data row4 col12\" >0.000159</td>\n",
       "      <td id=\"T_78b9c_row4_col13\" class=\"data row4 col13\" >-0.013393</td>\n",
       "      <td id=\"T_78b9c_row4_col14\" class=\"data row4 col14\" >0.004254</td>\n",
       "      <td id=\"T_78b9c_row4_col15\" class=\"data row4 col15\" >-0.010199</td>\n",
       "      <td id=\"T_78b9c_row4_col16\" class=\"data row4 col16\" >-0.032848</td>\n",
       "      <td id=\"T_78b9c_row4_col17\" class=\"data row4 col17\" >-0.012700</td>\n",
       "      <td id=\"T_78b9c_row4_col18\" class=\"data row4 col18\" >0.003612</td>\n",
       "      <td id=\"T_78b9c_row4_col19\" class=\"data row4 col19\" >-0.013624</td>\n",
       "      <td id=\"T_78b9c_row4_col20\" class=\"data row4 col20\" >-0.013559</td>\n",
       "      <td id=\"T_78b9c_row4_col21\" class=\"data row4 col21\" >0.004033</td>\n",
       "      <td id=\"T_78b9c_row4_col22\" class=\"data row4 col22\" >-0.009856</td>\n",
       "      <td id=\"T_78b9c_row4_col23\" class=\"data row4 col23\" >0.004302</td>\n",
       "      <td id=\"T_78b9c_row4_col24\" class=\"data row4 col24\" >-0.001611</td>\n",
       "      <td id=\"T_78b9c_row4_col25\" class=\"data row4 col25\" >-0.000664</td>\n",
       "      <td id=\"T_78b9c_row4_col26\" class=\"data row4 col26\" >0.020649</td>\n",
       "      <td id=\"T_78b9c_row4_col27\" class=\"data row4 col27\" >-0.004484</td>\n",
       "      <td id=\"T_78b9c_row4_col28\" class=\"data row4 col28\" >-0.382747</td>\n",
       "      <td id=\"T_78b9c_row4_col29\" class=\"data row4 col29\" >0.175038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row5\" class=\"row_heading level0 row5\" >V6</th>\n",
       "      <td id=\"T_78b9c_row5_col0\" class=\"data row5 col0\" >-0.000177</td>\n",
       "      <td id=\"T_78b9c_row5_col1\" class=\"data row5 col1\" >0.004119</td>\n",
       "      <td id=\"T_78b9c_row5_col2\" class=\"data row5 col2\" >-0.000392</td>\n",
       "      <td id=\"T_78b9c_row5_col3\" class=\"data row5 col3\" >0.001191</td>\n",
       "      <td id=\"T_78b9c_row5_col4\" class=\"data row5 col4\" >-0.000701</td>\n",
       "      <td id=\"T_78b9c_row5_col5\" class=\"data row5 col5\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row5_col6\" class=\"data row5 col6\" >-0.010017</td>\n",
       "      <td id=\"T_78b9c_row5_col7\" class=\"data row5 col7\" >0.008153</td>\n",
       "      <td id=\"T_78b9c_row5_col8\" class=\"data row5 col8\" >-0.004149</td>\n",
       "      <td id=\"T_78b9c_row5_col9\" class=\"data row5 col9\" >-0.008792</td>\n",
       "      <td id=\"T_78b9c_row5_col10\" class=\"data row5 col10\" >0.002667</td>\n",
       "      <td id=\"T_78b9c_row5_col11\" class=\"data row5 col11\" >-0.008045</td>\n",
       "      <td id=\"T_78b9c_row5_col12\" class=\"data row5 col12\" >0.000016</td>\n",
       "      <td id=\"T_78b9c_row5_col13\" class=\"data row5 col13\" >-0.007086</td>\n",
       "      <td id=\"T_78b9c_row5_col14\" class=\"data row5 col14\" >0.003166</td>\n",
       "      <td id=\"T_78b9c_row5_col15\" class=\"data row5 col15\" >-0.007531</td>\n",
       "      <td id=\"T_78b9c_row5_col16\" class=\"data row5 col16\" >-0.007004</td>\n",
       "      <td id=\"T_78b9c_row5_col17\" class=\"data row5 col17\" >-0.000151</td>\n",
       "      <td id=\"T_78b9c_row5_col18\" class=\"data row5 col18\" >0.000606</td>\n",
       "      <td id=\"T_78b9c_row5_col19\" class=\"data row5 col19\" >0.014117</td>\n",
       "      <td id=\"T_78b9c_row5_col20\" class=\"data row5 col20\" >0.006419</td>\n",
       "      <td id=\"T_78b9c_row5_col21\" class=\"data row5 col21\" >-0.004982</td>\n",
       "      <td id=\"T_78b9c_row5_col22\" class=\"data row5 col22\" >0.004990</td>\n",
       "      <td id=\"T_78b9c_row5_col23\" class=\"data row5 col23\" >-0.001185</td>\n",
       "      <td id=\"T_78b9c_row5_col24\" class=\"data row5 col24\" >-0.000470</td>\n",
       "      <td id=\"T_78b9c_row5_col25\" class=\"data row5 col25\" >-0.003615</td>\n",
       "      <td id=\"T_78b9c_row5_col26\" class=\"data row5 col26\" >-0.011486</td>\n",
       "      <td id=\"T_78b9c_row5_col27\" class=\"data row5 col27\" >0.000046</td>\n",
       "      <td id=\"T_78b9c_row5_col28\" class=\"data row5 col28\" >0.210381</td>\n",
       "      <td id=\"T_78b9c_row5_col29\" class=\"data row5 col29\" >-0.061977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row6\" class=\"row_heading level0 row6\" >V7</th>\n",
       "      <td id=\"T_78b9c_row6_col0\" class=\"data row6 col0\" >-0.021839</td>\n",
       "      <td id=\"T_78b9c_row6_col1\" class=\"data row6 col1\" >0.020025</td>\n",
       "      <td id=\"T_78b9c_row6_col2\" class=\"data row6 col2\" >-0.023424</td>\n",
       "      <td id=\"T_78b9c_row6_col3\" class=\"data row6 col3\" >0.011269</td>\n",
       "      <td id=\"T_78b9c_row6_col4\" class=\"data row6 col4\" >-0.012571</td>\n",
       "      <td id=\"T_78b9c_row6_col5\" class=\"data row6 col5\" >-0.010017</td>\n",
       "      <td id=\"T_78b9c_row6_col6\" class=\"data row6 col6\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row6_col7\" class=\"data row6 col7\" >0.000813</td>\n",
       "      <td id=\"T_78b9c_row6_col8\" class=\"data row6 col8\" >-0.009075</td>\n",
       "      <td id=\"T_78b9c_row6_col9\" class=\"data row6 col9\" >-0.033170</td>\n",
       "      <td id=\"T_78b9c_row6_col10\" class=\"data row6 col10\" >0.017100</td>\n",
       "      <td id=\"T_78b9c_row6_col11\" class=\"data row6 col11\" >-0.027035</td>\n",
       "      <td id=\"T_78b9c_row6_col12\" class=\"data row6 col12\" >-0.002328</td>\n",
       "      <td id=\"T_78b9c_row6_col13\" class=\"data row6 col13\" >-0.026328</td>\n",
       "      <td id=\"T_78b9c_row6_col14\" class=\"data row6 col14\" >-0.005088</td>\n",
       "      <td id=\"T_78b9c_row6_col15\" class=\"data row6 col15\" >-0.030188</td>\n",
       "      <td id=\"T_78b9c_row6_col16\" class=\"data row6 col16\" >-0.049680</td>\n",
       "      <td id=\"T_78b9c_row6_col17\" class=\"data row6 col17\" >-0.022684</td>\n",
       "      <td id=\"T_78b9c_row6_col18\" class=\"data row6 col18\" >0.004158</td>\n",
       "      <td id=\"T_78b9c_row6_col19\" class=\"data row6 col19\" >0.027373</td>\n",
       "      <td id=\"T_78b9c_row6_col20\" class=\"data row6 col20\" >0.008721</td>\n",
       "      <td id=\"T_78b9c_row6_col21\" class=\"data row6 col21\" >-0.000123</td>\n",
       "      <td id=\"T_78b9c_row6_col22\" class=\"data row6 col22\" >0.011034</td>\n",
       "      <td id=\"T_78b9c_row6_col23\" class=\"data row6 col23\" >-0.001084</td>\n",
       "      <td id=\"T_78b9c_row6_col24\" class=\"data row6 col24\" >0.000420</td>\n",
       "      <td id=\"T_78b9c_row6_col25\" class=\"data row6 col25\" >0.004651</td>\n",
       "      <td id=\"T_78b9c_row6_col26\" class=\"data row6 col26\" >-0.033952</td>\n",
       "      <td id=\"T_78b9c_row6_col27\" class=\"data row6 col27\" >0.006749</td>\n",
       "      <td id=\"T_78b9c_row6_col28\" class=\"data row6 col28\" >0.391239</td>\n",
       "      <td id=\"T_78b9c_row6_col29\" class=\"data row6 col29\" >0.084646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row7\" class=\"row_heading level0 row7\" >V8</th>\n",
       "      <td id=\"T_78b9c_row7_col0\" class=\"data row7 col0\" >0.003626</td>\n",
       "      <td id=\"T_78b9c_row7_col1\" class=\"data row7 col1\" >0.002628</td>\n",
       "      <td id=\"T_78b9c_row7_col2\" class=\"data row7 col2\" >0.002836</td>\n",
       "      <td id=\"T_78b9c_row7_col3\" class=\"data row7 col3\" >0.000271</td>\n",
       "      <td id=\"T_78b9c_row7_col4\" class=\"data row7 col4\" >-0.000730</td>\n",
       "      <td id=\"T_78b9c_row7_col5\" class=\"data row7 col5\" >0.008153</td>\n",
       "      <td id=\"T_78b9c_row7_col6\" class=\"data row7 col6\" >0.000813</td>\n",
       "      <td id=\"T_78b9c_row7_col7\" class=\"data row7 col7\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row7_col8\" class=\"data row7 col8\" >-0.003052</td>\n",
       "      <td id=\"T_78b9c_row7_col9\" class=\"data row7 col9\" >-0.001326</td>\n",
       "      <td id=\"T_78b9c_row7_col10\" class=\"data row7 col10\" >-0.004210</td>\n",
       "      <td id=\"T_78b9c_row7_col11\" class=\"data row7 col11\" >0.008680</td>\n",
       "      <td id=\"T_78b9c_row7_col12\" class=\"data row7 col12\" >0.000216</td>\n",
       "      <td id=\"T_78b9c_row7_col13\" class=\"data row7 col13\" >0.010783</td>\n",
       "      <td id=\"T_78b9c_row7_col14\" class=\"data row7 col14\" >-0.001076</td>\n",
       "      <td id=\"T_78b9c_row7_col15\" class=\"data row7 col15\" >0.006039</td>\n",
       "      <td id=\"T_78b9c_row7_col16\" class=\"data row7 col16\" >0.008411</td>\n",
       "      <td id=\"T_78b9c_row7_col17\" class=\"data row7 col17\" >0.001526</td>\n",
       "      <td id=\"T_78b9c_row7_col18\" class=\"data row7 col18\" >-0.003837</td>\n",
       "      <td id=\"T_78b9c_row7_col19\" class=\"data row7 col19\" >-0.004896</td>\n",
       "      <td id=\"T_78b9c_row7_col20\" class=\"data row7 col20\" >0.001672</td>\n",
       "      <td id=\"T_78b9c_row7_col21\" class=\"data row7 col21\" >0.000508</td>\n",
       "      <td id=\"T_78b9c_row7_col22\" class=\"data row7 col22\" >0.013885</td>\n",
       "      <td id=\"T_78b9c_row7_col23\" class=\"data row7 col23\" >0.000094</td>\n",
       "      <td id=\"T_78b9c_row7_col24\" class=\"data row7 col24\" >-0.003231</td>\n",
       "      <td id=\"T_78b9c_row7_col25\" class=\"data row7 col25\" >0.003217</td>\n",
       "      <td id=\"T_78b9c_row7_col26\" class=\"data row7 col26\" >-0.006134</td>\n",
       "      <td id=\"T_78b9c_row7_col27\" class=\"data row7 col27\" >-0.014922</td>\n",
       "      <td id=\"T_78b9c_row7_col28\" class=\"data row7 col28\" >-0.102777</td>\n",
       "      <td id=\"T_78b9c_row7_col29\" class=\"data row7 col29\" >-0.036350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row8\" class=\"row_heading level0 row8\" >V9</th>\n",
       "      <td id=\"T_78b9c_row8_col0\" class=\"data row8 col0\" >-0.009599</td>\n",
       "      <td id=\"T_78b9c_row8_col1\" class=\"data row8 col1\" >0.001865</td>\n",
       "      <td id=\"T_78b9c_row8_col2\" class=\"data row8 col2\" >-0.017386</td>\n",
       "      <td id=\"T_78b9c_row8_col3\" class=\"data row8 col3\" >0.008993</td>\n",
       "      <td id=\"T_78b9c_row8_col4\" class=\"data row8 col4\" >-0.006367</td>\n",
       "      <td id=\"T_78b9c_row8_col5\" class=\"data row8 col5\" >-0.004149</td>\n",
       "      <td id=\"T_78b9c_row8_col6\" class=\"data row8 col6\" >-0.009075</td>\n",
       "      <td id=\"T_78b9c_row8_col7\" class=\"data row8 col7\" >-0.003052</td>\n",
       "      <td id=\"T_78b9c_row8_col8\" class=\"data row8 col8\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row8_col9\" class=\"data row8 col9\" >-0.009130</td>\n",
       "      <td id=\"T_78b9c_row8_col10\" class=\"data row8 col10\" >0.009347</td>\n",
       "      <td id=\"T_78b9c_row8_col11\" class=\"data row8 col11\" >-0.017163</td>\n",
       "      <td id=\"T_78b9c_row8_col12\" class=\"data row8 col12\" >-0.002462</td>\n",
       "      <td id=\"T_78b9c_row8_col13\" class=\"data row8 col13\" >-0.016008</td>\n",
       "      <td id=\"T_78b9c_row8_col14\" class=\"data row8 col14\" >-0.002699</td>\n",
       "      <td id=\"T_78b9c_row8_col15\" class=\"data row8 col15\" >-0.017783</td>\n",
       "      <td id=\"T_78b9c_row8_col16\" class=\"data row8 col16\" >-0.029087</td>\n",
       "      <td id=\"T_78b9c_row8_col17\" class=\"data row8 col17\" >-0.013444</td>\n",
       "      <td id=\"T_78b9c_row8_col18\" class=\"data row8 col18\" >0.002219</td>\n",
       "      <td id=\"T_78b9c_row8_col19\" class=\"data row8 col19\" >-0.000583</td>\n",
       "      <td id=\"T_78b9c_row8_col20\" class=\"data row8 col20\" >-0.007212</td>\n",
       "      <td id=\"T_78b9c_row8_col21\" class=\"data row8 col21\" >-0.002725</td>\n",
       "      <td id=\"T_78b9c_row8_col22\" class=\"data row8 col22\" >-0.008715</td>\n",
       "      <td id=\"T_78b9c_row8_col23\" class=\"data row8 col23\" >0.002114</td>\n",
       "      <td id=\"T_78b9c_row8_col24\" class=\"data row8 col24\" >-0.001449</td>\n",
       "      <td id=\"T_78b9c_row8_col25\" class=\"data row8 col25\" >0.001348</td>\n",
       "      <td id=\"T_78b9c_row8_col26\" class=\"data row8 col26\" >-0.001733</td>\n",
       "      <td id=\"T_78b9c_row8_col27\" class=\"data row8 col27\" >0.008454</td>\n",
       "      <td id=\"T_78b9c_row8_col28\" class=\"data row8 col28\" >-0.041503</td>\n",
       "      <td id=\"T_78b9c_row8_col29\" class=\"data row8 col29\" >-0.007580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row9\" class=\"row_heading level0 row9\" >V10</th>\n",
       "      <td id=\"T_78b9c_row9_col0\" class=\"data row9 col0\" >-0.019278</td>\n",
       "      <td id=\"T_78b9c_row9_col1\" class=\"data row9 col1\" >0.011013</td>\n",
       "      <td id=\"T_78b9c_row9_col2\" class=\"data row9 col2\" >-0.028125</td>\n",
       "      <td id=\"T_78b9c_row9_col3\" class=\"data row9 col3\" >0.015209</td>\n",
       "      <td id=\"T_78b9c_row9_col4\" class=\"data row9 col4\" >-0.015695</td>\n",
       "      <td id=\"T_78b9c_row9_col5\" class=\"data row9 col5\" >-0.008792</td>\n",
       "      <td id=\"T_78b9c_row9_col6\" class=\"data row9 col6\" >-0.033170</td>\n",
       "      <td id=\"T_78b9c_row9_col7\" class=\"data row9 col7\" >-0.001326</td>\n",
       "      <td id=\"T_78b9c_row9_col8\" class=\"data row9 col8\" >-0.009130</td>\n",
       "      <td id=\"T_78b9c_row9_col9\" class=\"data row9 col9\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row9_col10\" class=\"data row9 col10\" >0.019515</td>\n",
       "      <td id=\"T_78b9c_row9_col11\" class=\"data row9 col11\" >-0.043050</td>\n",
       "      <td id=\"T_78b9c_row9_col12\" class=\"data row9 col12\" >-0.001893</td>\n",
       "      <td id=\"T_78b9c_row9_col13\" class=\"data row9 col13\" >-0.037438</td>\n",
       "      <td id=\"T_78b9c_row9_col14\" class=\"data row9 col14\" >0.000347</td>\n",
       "      <td id=\"T_78b9c_row9_col15\" class=\"data row9 col15\" >-0.029659</td>\n",
       "      <td id=\"T_78b9c_row9_col16\" class=\"data row9 col16\" >-0.060935</td>\n",
       "      <td id=\"T_78b9c_row9_col17\" class=\"data row9 col17\" >-0.017309</td>\n",
       "      <td id=\"T_78b9c_row9_col18\" class=\"data row9 col18\" >0.007515</td>\n",
       "      <td id=\"T_78b9c_row9_col19\" class=\"data row9 col19\" >0.004115</td>\n",
       "      <td id=\"T_78b9c_row9_col20\" class=\"data row9 col20\" >-0.006280</td>\n",
       "      <td id=\"T_78b9c_row9_col21\" class=\"data row9 col21\" >0.001598</td>\n",
       "      <td id=\"T_78b9c_row9_col22\" class=\"data row9 col22\" >-0.015512</td>\n",
       "      <td id=\"T_78b9c_row9_col23\" class=\"data row9 col23\" >0.004706</td>\n",
       "      <td id=\"T_78b9c_row9_col24\" class=\"data row9 col24\" >-0.000694</td>\n",
       "      <td id=\"T_78b9c_row9_col25\" class=\"data row9 col25\" >-0.000581</td>\n",
       "      <td id=\"T_78b9c_row9_col26\" class=\"data row9 col26\" >-0.005133</td>\n",
       "      <td id=\"T_78b9c_row9_col27\" class=\"data row9 col27\" >0.002669</td>\n",
       "      <td id=\"T_78b9c_row9_col28\" class=\"data row9 col28\" >-0.103121</td>\n",
       "      <td id=\"T_78b9c_row9_col29\" class=\"data row9 col29\" >0.029717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row10\" class=\"row_heading level0 row10\" >V11</th>\n",
       "      <td id=\"T_78b9c_row10_col0\" class=\"data row10 col0\" >0.003578</td>\n",
       "      <td id=\"T_78b9c_row10_col1\" class=\"data row10 col1\" >-0.006297</td>\n",
       "      <td id=\"T_78b9c_row10_col2\" class=\"data row10 col2\" >0.014767</td>\n",
       "      <td id=\"T_78b9c_row10_col3\" class=\"data row10 col3\" >-0.011847</td>\n",
       "      <td id=\"T_78b9c_row10_col4\" class=\"data row10 col4\" >0.012430</td>\n",
       "      <td id=\"T_78b9c_row10_col5\" class=\"data row10 col5\" >0.002667</td>\n",
       "      <td id=\"T_78b9c_row10_col6\" class=\"data row10 col6\" >0.017100</td>\n",
       "      <td id=\"T_78b9c_row10_col7\" class=\"data row10 col7\" >-0.004210</td>\n",
       "      <td id=\"T_78b9c_row10_col8\" class=\"data row10 col8\" >0.009347</td>\n",
       "      <td id=\"T_78b9c_row10_col9\" class=\"data row10 col9\" >0.019515</td>\n",
       "      <td id=\"T_78b9c_row10_col10\" class=\"data row10 col10\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row10_col11\" class=\"data row10 col11\" >0.023207</td>\n",
       "      <td id=\"T_78b9c_row10_col12\" class=\"data row10 col12\" >-0.000230</td>\n",
       "      <td id=\"T_78b9c_row10_col13\" class=\"data row10 col13\" >0.027573</td>\n",
       "      <td id=\"T_78b9c_row10_col14\" class=\"data row10 col14\" >0.000902</td>\n",
       "      <td id=\"T_78b9c_row10_col15\" class=\"data row10 col15\" >0.015528</td>\n",
       "      <td id=\"T_78b9c_row10_col16\" class=\"data row10 col16\" >0.033550</td>\n",
       "      <td id=\"T_78b9c_row10_col17\" class=\"data row10 col17\" >0.011801</td>\n",
       "      <td id=\"T_78b9c_row10_col18\" class=\"data row10 col18\" >-0.003847</td>\n",
       "      <td id=\"T_78b9c_row10_col19\" class=\"data row10 col19\" >-0.002357</td>\n",
       "      <td id=\"T_78b9c_row10_col20\" class=\"data row10 col20\" >-0.006372</td>\n",
       "      <td id=\"T_78b9c_row10_col21\" class=\"data row10 col21\" >0.000032</td>\n",
       "      <td id=\"T_78b9c_row10_col22\" class=\"data row10 col22\" >-0.001902</td>\n",
       "      <td id=\"T_78b9c_row10_col23\" class=\"data row10 col23\" >-0.003079</td>\n",
       "      <td id=\"T_78b9c_row10_col24\" class=\"data row10 col24\" >-0.003555</td>\n",
       "      <td id=\"T_78b9c_row10_col25\" class=\"data row10 col25\" >0.001169</td>\n",
       "      <td id=\"T_78b9c_row10_col26\" class=\"data row10 col26\" >-0.001217</td>\n",
       "      <td id=\"T_78b9c_row10_col27\" class=\"data row10 col27\" >-0.001694</td>\n",
       "      <td id=\"T_78b9c_row10_col28\" class=\"data row10 col28\" >-0.002312</td>\n",
       "      <td id=\"T_78b9c_row10_col29\" class=\"data row10 col29\" >-0.247051</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row11\" class=\"row_heading level0 row11\" >V12</th>\n",
       "      <td id=\"T_78b9c_row11_col0\" class=\"data row11 col0\" >-0.011240</td>\n",
       "      <td id=\"T_78b9c_row11_col1\" class=\"data row11 col1\" >0.013002</td>\n",
       "      <td id=\"T_78b9c_row11_col2\" class=\"data row11 col2\" >-0.031194</td>\n",
       "      <td id=\"T_78b9c_row11_col3\" class=\"data row11 col3\" >0.019518</td>\n",
       "      <td id=\"T_78b9c_row11_col4\" class=\"data row11 col4\" >-0.016962</td>\n",
       "      <td id=\"T_78b9c_row11_col5\" class=\"data row11 col5\" >-0.008045</td>\n",
       "      <td id=\"T_78b9c_row11_col6\" class=\"data row11 col6\" >-0.027035</td>\n",
       "      <td id=\"T_78b9c_row11_col7\" class=\"data row11 col7\" >0.008680</td>\n",
       "      <td id=\"T_78b9c_row11_col8\" class=\"data row11 col8\" >-0.017163</td>\n",
       "      <td id=\"T_78b9c_row11_col9\" class=\"data row11 col9\" >-0.043050</td>\n",
       "      <td id=\"T_78b9c_row11_col10\" class=\"data row11 col10\" >0.023207</td>\n",
       "      <td id=\"T_78b9c_row11_col11\" class=\"data row11 col11\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row11_col12\" class=\"data row11 col12\" >-0.000627</td>\n",
       "      <td id=\"T_78b9c_row11_col13\" class=\"data row11 col13\" >-0.047175</td>\n",
       "      <td id=\"T_78b9c_row11_col14\" class=\"data row11 col14\" >-0.002212</td>\n",
       "      <td id=\"T_78b9c_row11_col15\" class=\"data row11 col15\" >-0.037526</td>\n",
       "      <td id=\"T_78b9c_row11_col16\" class=\"data row11 col16\" >-0.063247</td>\n",
       "      <td id=\"T_78b9c_row11_col17\" class=\"data row11 col17\" >-0.019409</td>\n",
       "      <td id=\"T_78b9c_row11_col18\" class=\"data row11 col18\" >0.008660</td>\n",
       "      <td id=\"T_78b9c_row11_col19\" class=\"data row11 col19\" >-0.006171</td>\n",
       "      <td id=\"T_78b9c_row11_col20\" class=\"data row11 col20\" >0.004447</td>\n",
       "      <td id=\"T_78b9c_row11_col21\" class=\"data row11 col21\" >0.006262</td>\n",
       "      <td id=\"T_78b9c_row11_col22\" class=\"data row11 col22\" >-0.006749</td>\n",
       "      <td id=\"T_78b9c_row11_col23\" class=\"data row11 col23\" >0.000183</td>\n",
       "      <td id=\"T_78b9c_row11_col24\" class=\"data row11 col24\" >-0.000906</td>\n",
       "      <td id=\"T_78b9c_row11_col25\" class=\"data row11 col25\" >0.000006</td>\n",
       "      <td id=\"T_78b9c_row11_col26\" class=\"data row11 col26\" >-0.000261</td>\n",
       "      <td id=\"T_78b9c_row11_col27\" class=\"data row11 col27\" >0.007021</td>\n",
       "      <td id=\"T_78b9c_row11_col28\" class=\"data row11 col28\" >-0.007580</td>\n",
       "      <td id=\"T_78b9c_row11_col29\" class=\"data row11 col29\" >0.126405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row12\" class=\"row_heading level0 row12\" >V13</th>\n",
       "      <td id=\"T_78b9c_row12_col0\" class=\"data row12 col0\" >0.003362</td>\n",
       "      <td id=\"T_78b9c_row12_col1\" class=\"data row12 col1\" >0.001395</td>\n",
       "      <td id=\"T_78b9c_row12_col2\" class=\"data row12 col2\" >-0.000198</td>\n",
       "      <td id=\"T_78b9c_row12_col3\" class=\"data row12 col3\" >0.000823</td>\n",
       "      <td id=\"T_78b9c_row12_col4\" class=\"data row12 col4\" >0.000159</td>\n",
       "      <td id=\"T_78b9c_row12_col5\" class=\"data row12 col5\" >0.000016</td>\n",
       "      <td id=\"T_78b9c_row12_col6\" class=\"data row12 col6\" >-0.002328</td>\n",
       "      <td id=\"T_78b9c_row12_col7\" class=\"data row12 col7\" >0.000216</td>\n",
       "      <td id=\"T_78b9c_row12_col8\" class=\"data row12 col8\" >-0.002462</td>\n",
       "      <td id=\"T_78b9c_row12_col9\" class=\"data row12 col9\" >-0.001893</td>\n",
       "      <td id=\"T_78b9c_row12_col10\" class=\"data row12 col10\" >-0.000230</td>\n",
       "      <td id=\"T_78b9c_row12_col11\" class=\"data row12 col11\" >-0.000627</td>\n",
       "      <td id=\"T_78b9c_row12_col12\" class=\"data row12 col12\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row12_col13\" class=\"data row12 col13\" >-0.001228</td>\n",
       "      <td id=\"T_78b9c_row12_col14\" class=\"data row12 col14\" >0.001389</td>\n",
       "      <td id=\"T_78b9c_row12_col15\" class=\"data row12 col15\" >0.002060</td>\n",
       "      <td id=\"T_78b9c_row12_col16\" class=\"data row12 col16\" >0.002138</td>\n",
       "      <td id=\"T_78b9c_row12_col17\" class=\"data row12 col17\" >0.000218</td>\n",
       "      <td id=\"T_78b9c_row12_col18\" class=\"data row12 col18\" >-0.000276</td>\n",
       "      <td id=\"T_78b9c_row12_col19\" class=\"data row12 col19\" >-0.005098</td>\n",
       "      <td id=\"T_78b9c_row12_col20\" class=\"data row12 col20\" >-0.001357</td>\n",
       "      <td id=\"T_78b9c_row12_col21\" class=\"data row12 col21\" >0.000038</td>\n",
       "      <td id=\"T_78b9c_row12_col22\" class=\"data row12 col22\" >-0.008278</td>\n",
       "      <td id=\"T_78b9c_row12_col23\" class=\"data row12 col23\" >-0.001882</td>\n",
       "      <td id=\"T_78b9c_row12_col24\" class=\"data row12 col24\" >-0.002112</td>\n",
       "      <td id=\"T_78b9c_row12_col25\" class=\"data row12 col25\" >-0.001993</td>\n",
       "      <td id=\"T_78b9c_row12_col26\" class=\"data row12 col26\" >-0.001064</td>\n",
       "      <td id=\"T_78b9c_row12_col27\" class=\"data row12 col27\" >0.007712</td>\n",
       "      <td id=\"T_78b9c_row12_col28\" class=\"data row12 col28\" >0.001035</td>\n",
       "      <td id=\"T_78b9c_row12_col29\" class=\"data row12 col29\" >-0.063595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row13\" class=\"row_heading level0 row13\" >V14</th>\n",
       "      <td id=\"T_78b9c_row13_col0\" class=\"data row13 col0\" >-0.010310</td>\n",
       "      <td id=\"T_78b9c_row13_col1\" class=\"data row13 col1\" >0.018952</td>\n",
       "      <td id=\"T_78b9c_row13_col2\" class=\"data row13 col2\" >-0.027995</td>\n",
       "      <td id=\"T_78b9c_row13_col3\" class=\"data row13 col3\" >0.017077</td>\n",
       "      <td id=\"T_78b9c_row13_col4\" class=\"data row13 col4\" >-0.013393</td>\n",
       "      <td id=\"T_78b9c_row13_col5\" class=\"data row13 col5\" >-0.007086</td>\n",
       "      <td id=\"T_78b9c_row13_col6\" class=\"data row13 col6\" >-0.026328</td>\n",
       "      <td id=\"T_78b9c_row13_col7\" class=\"data row13 col7\" >0.010783</td>\n",
       "      <td id=\"T_78b9c_row13_col8\" class=\"data row13 col8\" >-0.016008</td>\n",
       "      <td id=\"T_78b9c_row13_col9\" class=\"data row13 col9\" >-0.037438</td>\n",
       "      <td id=\"T_78b9c_row13_col10\" class=\"data row13 col10\" >0.027573</td>\n",
       "      <td id=\"T_78b9c_row13_col11\" class=\"data row13 col11\" >-0.047175</td>\n",
       "      <td id=\"T_78b9c_row13_col12\" class=\"data row13 col12\" >-0.001228</td>\n",
       "      <td id=\"T_78b9c_row13_col13\" class=\"data row13 col13\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row13_col14\" class=\"data row13 col14\" >0.002867</td>\n",
       "      <td id=\"T_78b9c_row13_col15\" class=\"data row13 col15\" >-0.037006</td>\n",
       "      <td id=\"T_78b9c_row13_col16\" class=\"data row13 col16\" >-0.060543</td>\n",
       "      <td id=\"T_78b9c_row13_col17\" class=\"data row13 col17\" >-0.021840</td>\n",
       "      <td id=\"T_78b9c_row13_col18\" class=\"data row13 col18\" >0.008848</td>\n",
       "      <td id=\"T_78b9c_row13_col19\" class=\"data row13 col19\" >-0.005431</td>\n",
       "      <td id=\"T_78b9c_row13_col20\" class=\"data row13 col20\" >0.010242</td>\n",
       "      <td id=\"T_78b9c_row13_col21\" class=\"data row13 col21\" >-0.000861</td>\n",
       "      <td id=\"T_78b9c_row13_col22\" class=\"data row13 col22\" >0.002720</td>\n",
       "      <td id=\"T_78b9c_row13_col23\" class=\"data row13 col23\" >-0.004254</td>\n",
       "      <td id=\"T_78b9c_row13_col24\" class=\"data row13 col24\" >0.004171</td>\n",
       "      <td id=\"T_78b9c_row13_col25\" class=\"data row13 col25\" >-0.001902</td>\n",
       "      <td id=\"T_78b9c_row13_col26\" class=\"data row13 col26\" >0.001947</td>\n",
       "      <td id=\"T_78b9c_row13_col27\" class=\"data row13 col27\" >0.001494</td>\n",
       "      <td id=\"T_78b9c_row13_col28\" class=\"data row13 col28\" >0.033107</td>\n",
       "      <td id=\"T_78b9c_row13_col29\" class=\"data row13 col29\" >-0.104093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row14\" class=\"row_heading level0 row14\" >V15</th>\n",
       "      <td id=\"T_78b9c_row14_col0\" class=\"data row14 col0\" >-0.002944</td>\n",
       "      <td id=\"T_78b9c_row14_col1\" class=\"data row14 col1\" >-0.003989</td>\n",
       "      <td id=\"T_78b9c_row14_col2\" class=\"data row14 col2\" >-0.000757</td>\n",
       "      <td id=\"T_78b9c_row14_col3\" class=\"data row14 col3\" >0.004899</td>\n",
       "      <td id=\"T_78b9c_row14_col4\" class=\"data row14 col4\" >0.004254</td>\n",
       "      <td id=\"T_78b9c_row14_col5\" class=\"data row14 col5\" >0.003166</td>\n",
       "      <td id=\"T_78b9c_row14_col6\" class=\"data row14 col6\" >-0.005088</td>\n",
       "      <td id=\"T_78b9c_row14_col7\" class=\"data row14 col7\" >-0.001076</td>\n",
       "      <td id=\"T_78b9c_row14_col8\" class=\"data row14 col8\" >-0.002699</td>\n",
       "      <td id=\"T_78b9c_row14_col9\" class=\"data row14 col9\" >0.000347</td>\n",
       "      <td id=\"T_78b9c_row14_col10\" class=\"data row14 col10\" >0.000902</td>\n",
       "      <td id=\"T_78b9c_row14_col11\" class=\"data row14 col11\" >-0.002212</td>\n",
       "      <td id=\"T_78b9c_row14_col12\" class=\"data row14 col12\" >0.001389</td>\n",
       "      <td id=\"T_78b9c_row14_col13\" class=\"data row14 col13\" >0.002867</td>\n",
       "      <td id=\"T_78b9c_row14_col14\" class=\"data row14 col14\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row14_col15\" class=\"data row14 col15\" >-0.000723</td>\n",
       "      <td id=\"T_78b9c_row14_col16\" class=\"data row14 col16\" >-0.003153</td>\n",
       "      <td id=\"T_78b9c_row14_col17\" class=\"data row14 col17\" >0.002868</td>\n",
       "      <td id=\"T_78b9c_row14_col18\" class=\"data row14 col18\" >0.000333</td>\n",
       "      <td id=\"T_78b9c_row14_col19\" class=\"data row14 col19\" >-0.003305</td>\n",
       "      <td id=\"T_78b9c_row14_col20\" class=\"data row14 col20\" >0.001312</td>\n",
       "      <td id=\"T_78b9c_row14_col21\" class=\"data row14 col21\" >-0.001683</td>\n",
       "      <td id=\"T_78b9c_row14_col22\" class=\"data row14 col22\" >-0.007571</td>\n",
       "      <td id=\"T_78b9c_row14_col23\" class=\"data row14 col23\" >0.000642</td>\n",
       "      <td id=\"T_78b9c_row14_col24\" class=\"data row14 col24\" >0.000280</td>\n",
       "      <td id=\"T_78b9c_row14_col25\" class=\"data row14 col25\" >0.001081</td>\n",
       "      <td id=\"T_78b9c_row14_col26\" class=\"data row14 col26\" >-0.002872</td>\n",
       "      <td id=\"T_78b9c_row14_col27\" class=\"data row14 col27\" >0.010331</td>\n",
       "      <td id=\"T_78b9c_row14_col28\" class=\"data row14 col28\" >-0.003503</td>\n",
       "      <td id=\"T_78b9c_row14_col29\" class=\"data row14 col29\" >-0.181832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row15\" class=\"row_heading level0 row15\" >V16</th>\n",
       "      <td id=\"T_78b9c_row15_col0\" class=\"data row15 col0\" >-0.014766</td>\n",
       "      <td id=\"T_78b9c_row15_col1\" class=\"data row15 col1\" >0.005743</td>\n",
       "      <td id=\"T_78b9c_row15_col2\" class=\"data row15 col2\" >-0.025457</td>\n",
       "      <td id=\"T_78b9c_row15_col3\" class=\"data row15 col3\" >0.016185</td>\n",
       "      <td id=\"T_78b9c_row15_col4\" class=\"data row15 col4\" >-0.010199</td>\n",
       "      <td id=\"T_78b9c_row15_col5\" class=\"data row15 col5\" >-0.007531</td>\n",
       "      <td id=\"T_78b9c_row15_col6\" class=\"data row15 col6\" >-0.030188</td>\n",
       "      <td id=\"T_78b9c_row15_col7\" class=\"data row15 col7\" >0.006039</td>\n",
       "      <td id=\"T_78b9c_row15_col8\" class=\"data row15 col8\" >-0.017783</td>\n",
       "      <td id=\"T_78b9c_row15_col9\" class=\"data row15 col9\" >-0.029659</td>\n",
       "      <td id=\"T_78b9c_row15_col10\" class=\"data row15 col10\" >0.015528</td>\n",
       "      <td id=\"T_78b9c_row15_col11\" class=\"data row15 col11\" >-0.037526</td>\n",
       "      <td id=\"T_78b9c_row15_col12\" class=\"data row15 col12\" >0.002060</td>\n",
       "      <td id=\"T_78b9c_row15_col13\" class=\"data row15 col13\" >-0.037006</td>\n",
       "      <td id=\"T_78b9c_row15_col14\" class=\"data row15 col14\" >-0.000723</td>\n",
       "      <td id=\"T_78b9c_row15_col15\" class=\"data row15 col15\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row15_col16\" class=\"data row15 col16\" >-0.055428</td>\n",
       "      <td id=\"T_78b9c_row15_col17\" class=\"data row15 col17\" >-0.022144</td>\n",
       "      <td id=\"T_78b9c_row15_col18\" class=\"data row15 col18\" >0.007966</td>\n",
       "      <td id=\"T_78b9c_row15_col19\" class=\"data row15 col19\" >0.002607</td>\n",
       "      <td id=\"T_78b9c_row15_col20\" class=\"data row15 col20\" >0.012689</td>\n",
       "      <td id=\"T_78b9c_row15_col21\" class=\"data row15 col21\" >-0.000198</td>\n",
       "      <td id=\"T_78b9c_row15_col22\" class=\"data row15 col22\" >-0.005070</td>\n",
       "      <td id=\"T_78b9c_row15_col23\" class=\"data row15 col23\" >0.000735</td>\n",
       "      <td id=\"T_78b9c_row15_col24\" class=\"data row15 col24\" >-0.004086</td>\n",
       "      <td id=\"T_78b9c_row15_col25\" class=\"data row15 col25\" >0.001122</td>\n",
       "      <td id=\"T_78b9c_row15_col26\" class=\"data row15 col26\" >-0.008073</td>\n",
       "      <td id=\"T_78b9c_row15_col27\" class=\"data row15 col27\" >0.016978</td>\n",
       "      <td id=\"T_78b9c_row15_col28\" class=\"data row15 col28\" >-0.004419</td>\n",
       "      <td id=\"T_78b9c_row15_col29\" class=\"data row15 col29\" >0.012150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row16\" class=\"row_heading level0 row16\" >V17</th>\n",
       "      <td id=\"T_78b9c_row16_col0\" class=\"data row16 col0\" >-0.023218</td>\n",
       "      <td id=\"T_78b9c_row16_col1\" class=\"data row16 col1\" >0.021561</td>\n",
       "      <td id=\"T_78b9c_row16_col2\" class=\"data row16 col2\" >-0.043299</td>\n",
       "      <td id=\"T_78b9c_row16_col3\" class=\"data row16 col3\" >0.029483</td>\n",
       "      <td id=\"T_78b9c_row16_col4\" class=\"data row16 col4\" >-0.032848</td>\n",
       "      <td id=\"T_78b9c_row16_col5\" class=\"data row16 col5\" >-0.007004</td>\n",
       "      <td id=\"T_78b9c_row16_col6\" class=\"data row16 col6\" >-0.049680</td>\n",
       "      <td id=\"T_78b9c_row16_col7\" class=\"data row16 col7\" >0.008411</td>\n",
       "      <td id=\"T_78b9c_row16_col8\" class=\"data row16 col8\" >-0.029087</td>\n",
       "      <td id=\"T_78b9c_row16_col9\" class=\"data row16 col9\" >-0.060935</td>\n",
       "      <td id=\"T_78b9c_row16_col10\" class=\"data row16 col10\" >0.033550</td>\n",
       "      <td id=\"T_78b9c_row16_col11\" class=\"data row16 col11\" >-0.063247</td>\n",
       "      <td id=\"T_78b9c_row16_col12\" class=\"data row16 col12\" >0.002138</td>\n",
       "      <td id=\"T_78b9c_row16_col13\" class=\"data row16 col13\" >-0.060543</td>\n",
       "      <td id=\"T_78b9c_row16_col14\" class=\"data row16 col14\" >-0.003153</td>\n",
       "      <td id=\"T_78b9c_row16_col15\" class=\"data row16 col15\" >-0.055428</td>\n",
       "      <td id=\"T_78b9c_row16_col16\" class=\"data row16 col16\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row16_col17\" class=\"data row16 col17\" >-0.039021</td>\n",
       "      <td id=\"T_78b9c_row16_col18\" class=\"data row16 col18\" >0.009197</td>\n",
       "      <td id=\"T_78b9c_row16_col19\" class=\"data row16 col19\" >-0.001493</td>\n",
       "      <td id=\"T_78b9c_row16_col20\" class=\"data row16 col20\" >0.003763</td>\n",
       "      <td id=\"T_78b9c_row16_col21\" class=\"data row16 col21\" >0.002498</td>\n",
       "      <td id=\"T_78b9c_row16_col22\" class=\"data row16 col22\" >0.005391</td>\n",
       "      <td id=\"T_78b9c_row16_col23\" class=\"data row16 col23\" >-0.002260</td>\n",
       "      <td id=\"T_78b9c_row16_col24\" class=\"data row16 col24\" >-0.004748</td>\n",
       "      <td id=\"T_78b9c_row16_col25\" class=\"data row16 col25\" >0.001256</td>\n",
       "      <td id=\"T_78b9c_row16_col26\" class=\"data row16 col26\" >-0.002218</td>\n",
       "      <td id=\"T_78b9c_row16_col27\" class=\"data row16 col27\" >-0.003540</td>\n",
       "      <td id=\"T_78b9c_row16_col28\" class=\"data row16 col28\" >0.011952</td>\n",
       "      <td id=\"T_78b9c_row16_col29\" class=\"data row16 col29\" >-0.080489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row17\" class=\"row_heading level0 row17\" >V18</th>\n",
       "      <td id=\"T_78b9c_row17_col0\" class=\"data row17 col0\" >-0.007048</td>\n",
       "      <td id=\"T_78b9c_row17_col1\" class=\"data row17 col1\" >0.011934</td>\n",
       "      <td id=\"T_78b9c_row17_col2\" class=\"data row17 col2\" >-0.014518</td>\n",
       "      <td id=\"T_78b9c_row17_col3\" class=\"data row17 col3\" >0.008506</td>\n",
       "      <td id=\"T_78b9c_row17_col4\" class=\"data row17 col4\" >-0.012700</td>\n",
       "      <td id=\"T_78b9c_row17_col5\" class=\"data row17 col5\" >-0.000151</td>\n",
       "      <td id=\"T_78b9c_row17_col6\" class=\"data row17 col6\" >-0.022684</td>\n",
       "      <td id=\"T_78b9c_row17_col7\" class=\"data row17 col7\" >0.001526</td>\n",
       "      <td id=\"T_78b9c_row17_col8\" class=\"data row17 col8\" >-0.013444</td>\n",
       "      <td id=\"T_78b9c_row17_col9\" class=\"data row17 col9\" >-0.017309</td>\n",
       "      <td id=\"T_78b9c_row17_col10\" class=\"data row17 col10\" >0.011801</td>\n",
       "      <td id=\"T_78b9c_row17_col11\" class=\"data row17 col11\" >-0.019409</td>\n",
       "      <td id=\"T_78b9c_row17_col12\" class=\"data row17 col12\" >0.000218</td>\n",
       "      <td id=\"T_78b9c_row17_col13\" class=\"data row17 col13\" >-0.021840</td>\n",
       "      <td id=\"T_78b9c_row17_col14\" class=\"data row17 col14\" >0.002868</td>\n",
       "      <td id=\"T_78b9c_row17_col15\" class=\"data row17 col15\" >-0.022144</td>\n",
       "      <td id=\"T_78b9c_row17_col16\" class=\"data row17 col16\" >-0.039021</td>\n",
       "      <td id=\"T_78b9c_row17_col17\" class=\"data row17 col17\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row17_col18\" class=\"data row17 col18\" >0.003743</td>\n",
       "      <td id=\"T_78b9c_row17_col19\" class=\"data row17 col19\" >0.001532</td>\n",
       "      <td id=\"T_78b9c_row17_col20\" class=\"data row17 col20\" >0.000372</td>\n",
       "      <td id=\"T_78b9c_row17_col21\" class=\"data row17 col21\" >-0.001103</td>\n",
       "      <td id=\"T_78b9c_row17_col22\" class=\"data row17 col22\" >0.007771</td>\n",
       "      <td id=\"T_78b9c_row17_col23\" class=\"data row17 col23\" >-0.001614</td>\n",
       "      <td id=\"T_78b9c_row17_col24\" class=\"data row17 col24\" >0.001884</td>\n",
       "      <td id=\"T_78b9c_row17_col25\" class=\"data row17 col25\" >-0.000001</td>\n",
       "      <td id=\"T_78b9c_row17_col26\" class=\"data row17 col26\" >-0.004221</td>\n",
       "      <td id=\"T_78b9c_row17_col27\" class=\"data row17 col27\" >0.000518</td>\n",
       "      <td id=\"T_78b9c_row17_col28\" class=\"data row17 col28\" >0.033585</td>\n",
       "      <td id=\"T_78b9c_row17_col29\" class=\"data row17 col29\" >0.089812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row18\" class=\"row_heading level0 row18\" >V19</th>\n",
       "      <td id=\"T_78b9c_row18_col0\" class=\"data row18 col0\" >-0.000303</td>\n",
       "      <td id=\"T_78b9c_row18_col1\" class=\"data row18 col1\" >-0.007469</td>\n",
       "      <td id=\"T_78b9c_row18_col2\" class=\"data row18 col2\" >-0.000450</td>\n",
       "      <td id=\"T_78b9c_row18_col3\" class=\"data row18 col3\" >0.000459</td>\n",
       "      <td id=\"T_78b9c_row18_col4\" class=\"data row18 col4\" >0.003612</td>\n",
       "      <td id=\"T_78b9c_row18_col5\" class=\"data row18 col5\" >0.000606</td>\n",
       "      <td id=\"T_78b9c_row18_col6\" class=\"data row18 col6\" >0.004158</td>\n",
       "      <td id=\"T_78b9c_row18_col7\" class=\"data row18 col7\" >-0.003837</td>\n",
       "      <td id=\"T_78b9c_row18_col8\" class=\"data row18 col8\" >0.002219</td>\n",
       "      <td id=\"T_78b9c_row18_col9\" class=\"data row18 col9\" >0.007515</td>\n",
       "      <td id=\"T_78b9c_row18_col10\" class=\"data row18 col10\" >-0.003847</td>\n",
       "      <td id=\"T_78b9c_row18_col11\" class=\"data row18 col11\" >0.008660</td>\n",
       "      <td id=\"T_78b9c_row18_col12\" class=\"data row18 col12\" >-0.000276</td>\n",
       "      <td id=\"T_78b9c_row18_col13\" class=\"data row18 col13\" >0.008848</td>\n",
       "      <td id=\"T_78b9c_row18_col14\" class=\"data row18 col14\" >0.000333</td>\n",
       "      <td id=\"T_78b9c_row18_col15\" class=\"data row18 col15\" >0.007966</td>\n",
       "      <td id=\"T_78b9c_row18_col16\" class=\"data row18 col16\" >0.009197</td>\n",
       "      <td id=\"T_78b9c_row18_col17\" class=\"data row18 col17\" >0.003743</td>\n",
       "      <td id=\"T_78b9c_row18_col18\" class=\"data row18 col18\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row18_col19\" class=\"data row18 col19\" >0.005959</td>\n",
       "      <td id=\"T_78b9c_row18_col20\" class=\"data row18 col20\" >0.000546</td>\n",
       "      <td id=\"T_78b9c_row18_col21\" class=\"data row18 col21\" >-0.006349</td>\n",
       "      <td id=\"T_78b9c_row18_col22\" class=\"data row18 col22\" >-0.006434</td>\n",
       "      <td id=\"T_78b9c_row18_col23\" class=\"data row18 col23\" >0.001900</td>\n",
       "      <td id=\"T_78b9c_row18_col24\" class=\"data row18 col24\" >0.000290</td>\n",
       "      <td id=\"T_78b9c_row18_col25\" class=\"data row18 col25\" >0.001928</td>\n",
       "      <td id=\"T_78b9c_row18_col26\" class=\"data row18 col26\" >-0.006113</td>\n",
       "      <td id=\"T_78b9c_row18_col27\" class=\"data row18 col27\" >0.013287</td>\n",
       "      <td id=\"T_78b9c_row18_col28\" class=\"data row18 col28\" >-0.051163</td>\n",
       "      <td id=\"T_78b9c_row18_col29\" class=\"data row18 col29\" >0.030244</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row19\" class=\"row_heading level0 row19\" >V20</th>\n",
       "      <td id=\"T_78b9c_row19_col0\" class=\"data row19 col0\" >0.016057</td>\n",
       "      <td id=\"T_78b9c_row19_col1\" class=\"data row19 col1\" >0.022579</td>\n",
       "      <td id=\"T_78b9c_row19_col2\" class=\"data row19 col2\" >0.007484</td>\n",
       "      <td id=\"T_78b9c_row19_col3\" class=\"data row19 col3\" >-0.013619</td>\n",
       "      <td id=\"T_78b9c_row19_col4\" class=\"data row19 col4\" >-0.013624</td>\n",
       "      <td id=\"T_78b9c_row19_col5\" class=\"data row19 col5\" >0.014117</td>\n",
       "      <td id=\"T_78b9c_row19_col6\" class=\"data row19 col6\" >0.027373</td>\n",
       "      <td id=\"T_78b9c_row19_col7\" class=\"data row19 col7\" >-0.004896</td>\n",
       "      <td id=\"T_78b9c_row19_col8\" class=\"data row19 col8\" >-0.000583</td>\n",
       "      <td id=\"T_78b9c_row19_col9\" class=\"data row19 col9\" >0.004115</td>\n",
       "      <td id=\"T_78b9c_row19_col10\" class=\"data row19 col10\" >-0.002357</td>\n",
       "      <td id=\"T_78b9c_row19_col11\" class=\"data row19 col11\" >-0.006171</td>\n",
       "      <td id=\"T_78b9c_row19_col12\" class=\"data row19 col12\" >-0.005098</td>\n",
       "      <td id=\"T_78b9c_row19_col13\" class=\"data row19 col13\" >-0.005431</td>\n",
       "      <td id=\"T_78b9c_row19_col14\" class=\"data row19 col14\" >-0.003305</td>\n",
       "      <td id=\"T_78b9c_row19_col15\" class=\"data row19 col15\" >0.002607</td>\n",
       "      <td id=\"T_78b9c_row19_col16\" class=\"data row19 col16\" >-0.001493</td>\n",
       "      <td id=\"T_78b9c_row19_col17\" class=\"data row19 col17\" >0.001532</td>\n",
       "      <td id=\"T_78b9c_row19_col18\" class=\"data row19 col18\" >0.005959</td>\n",
       "      <td id=\"T_78b9c_row19_col19\" class=\"data row19 col19\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row19_col20\" class=\"data row19 col20\" >-0.008272</td>\n",
       "      <td id=\"T_78b9c_row19_col21\" class=\"data row19 col21\" >0.017365</td>\n",
       "      <td id=\"T_78b9c_row19_col22\" class=\"data row19 col22\" >0.054189</td>\n",
       "      <td id=\"T_78b9c_row19_col23\" class=\"data row19 col23\" >-0.002105</td>\n",
       "      <td id=\"T_78b9c_row19_col24\" class=\"data row19 col24\" >0.013497</td>\n",
       "      <td id=\"T_78b9c_row19_col25\" class=\"data row19 col25\" >-0.000066</td>\n",
       "      <td id=\"T_78b9c_row19_col26\" class=\"data row19 col26\" >0.030510</td>\n",
       "      <td id=\"T_78b9c_row19_col27\" class=\"data row19 col27\" >-0.065891</td>\n",
       "      <td id=\"T_78b9c_row19_col28\" class=\"data row19 col28\" >0.348219</td>\n",
       "      <td id=\"T_78b9c_row19_col29\" class=\"data row19 col29\" >-0.049359</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row20\" class=\"row_heading level0 row20\" >V21</th>\n",
       "      <td id=\"T_78b9c_row20_col0\" class=\"data row20 col0\" >0.000829</td>\n",
       "      <td id=\"T_78b9c_row20_col1\" class=\"data row20 col1\" >0.010194</td>\n",
       "      <td id=\"T_78b9c_row20_col2\" class=\"data row20 col2\" >-0.001319</td>\n",
       "      <td id=\"T_78b9c_row20_col3\" class=\"data row20 col3\" >-0.000008</td>\n",
       "      <td id=\"T_78b9c_row20_col4\" class=\"data row20 col4\" >-0.013559</td>\n",
       "      <td id=\"T_78b9c_row20_col5\" class=\"data row20 col5\" >0.006419</td>\n",
       "      <td id=\"T_78b9c_row20_col6\" class=\"data row20 col6\" >0.008721</td>\n",
       "      <td id=\"T_78b9c_row20_col7\" class=\"data row20 col7\" >0.001672</td>\n",
       "      <td id=\"T_78b9c_row20_col8\" class=\"data row20 col8\" >-0.007212</td>\n",
       "      <td id=\"T_78b9c_row20_col9\" class=\"data row20 col9\" >-0.006280</td>\n",
       "      <td id=\"T_78b9c_row20_col10\" class=\"data row20 col10\" >-0.006372</td>\n",
       "      <td id=\"T_78b9c_row20_col11\" class=\"data row20 col11\" >0.004447</td>\n",
       "      <td id=\"T_78b9c_row20_col12\" class=\"data row20 col12\" >-0.001357</td>\n",
       "      <td id=\"T_78b9c_row20_col13\" class=\"data row20 col13\" >0.010242</td>\n",
       "      <td id=\"T_78b9c_row20_col14\" class=\"data row20 col14\" >0.001312</td>\n",
       "      <td id=\"T_78b9c_row20_col15\" class=\"data row20 col15\" >0.012689</td>\n",
       "      <td id=\"T_78b9c_row20_col16\" class=\"data row20 col16\" >0.003763</td>\n",
       "      <td id=\"T_78b9c_row20_col17\" class=\"data row20 col17\" >0.000372</td>\n",
       "      <td id=\"T_78b9c_row20_col18\" class=\"data row20 col18\" >0.000546</td>\n",
       "      <td id=\"T_78b9c_row20_col19\" class=\"data row20 col19\" >-0.008272</td>\n",
       "      <td id=\"T_78b9c_row20_col20\" class=\"data row20 col20\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row20_col21\" class=\"data row20 col21\" >0.011044</td>\n",
       "      <td id=\"T_78b9c_row20_col22\" class=\"data row20 col22\" >0.015318</td>\n",
       "      <td id=\"T_78b9c_row20_col23\" class=\"data row20 col23\" >0.003217</td>\n",
       "      <td id=\"T_78b9c_row20_col24\" class=\"data row20 col24\" >0.008225</td>\n",
       "      <td id=\"T_78b9c_row20_col25\" class=\"data row20 col25\" >0.001941</td>\n",
       "      <td id=\"T_78b9c_row20_col26\" class=\"data row20 col26\" >0.006291</td>\n",
       "      <td id=\"T_78b9c_row20_col27\" class=\"data row20 col27\" >-0.031159</td>\n",
       "      <td id=\"T_78b9c_row20_col28\" class=\"data row20 col28\" >0.114076</td>\n",
       "      <td id=\"T_78b9c_row20_col29\" class=\"data row20 col29\" >0.046886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row21\" class=\"row_heading level0 row21\" >V22</th>\n",
       "      <td id=\"T_78b9c_row21_col0\" class=\"data row21 col0\" >-0.001097</td>\n",
       "      <td id=\"T_78b9c_row21_col1\" class=\"data row21 col1\" >0.003269</td>\n",
       "      <td id=\"T_78b9c_row21_col2\" class=\"data row21 col2\" >0.004616</td>\n",
       "      <td id=\"T_78b9c_row21_col3\" class=\"data row21 col3\" >-0.002732</td>\n",
       "      <td id=\"T_78b9c_row21_col4\" class=\"data row21 col4\" >0.004033</td>\n",
       "      <td id=\"T_78b9c_row21_col5\" class=\"data row21 col5\" >-0.004982</td>\n",
       "      <td id=\"T_78b9c_row21_col6\" class=\"data row21 col6\" >-0.000123</td>\n",
       "      <td id=\"T_78b9c_row21_col7\" class=\"data row21 col7\" >0.000508</td>\n",
       "      <td id=\"T_78b9c_row21_col8\" class=\"data row21 col8\" >-0.002725</td>\n",
       "      <td id=\"T_78b9c_row21_col9\" class=\"data row21 col9\" >0.001598</td>\n",
       "      <td id=\"T_78b9c_row21_col10\" class=\"data row21 col10\" >0.000032</td>\n",
       "      <td id=\"T_78b9c_row21_col11\" class=\"data row21 col11\" >0.006262</td>\n",
       "      <td id=\"T_78b9c_row21_col12\" class=\"data row21 col12\" >0.000038</td>\n",
       "      <td id=\"T_78b9c_row21_col13\" class=\"data row21 col13\" >-0.000861</td>\n",
       "      <td id=\"T_78b9c_row21_col14\" class=\"data row21 col14\" >-0.001683</td>\n",
       "      <td id=\"T_78b9c_row21_col15\" class=\"data row21 col15\" >-0.000198</td>\n",
       "      <td id=\"T_78b9c_row21_col16\" class=\"data row21 col16\" >0.002498</td>\n",
       "      <td id=\"T_78b9c_row21_col17\" class=\"data row21 col17\" >-0.001103</td>\n",
       "      <td id=\"T_78b9c_row21_col18\" class=\"data row21 col18\" >-0.006349</td>\n",
       "      <td id=\"T_78b9c_row21_col19\" class=\"data row21 col19\" >0.017365</td>\n",
       "      <td id=\"T_78b9c_row21_col20\" class=\"data row21 col20\" >0.011044</td>\n",
       "      <td id=\"T_78b9c_row21_col21\" class=\"data row21 col21\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row21_col22\" class=\"data row21 col22\" >-0.002295</td>\n",
       "      <td id=\"T_78b9c_row21_col23\" class=\"data row21 col23\" >0.001556</td>\n",
       "      <td id=\"T_78b9c_row21_col24\" class=\"data row21 col24\" >0.002036</td>\n",
       "      <td id=\"T_78b9c_row21_col25\" class=\"data row21 col25\" >-0.003430</td>\n",
       "      <td id=\"T_78b9c_row21_col26\" class=\"data row21 col26\" >-0.006681</td>\n",
       "      <td id=\"T_78b9c_row21_col27\" class=\"data row21 col27\" >0.002792</td>\n",
       "      <td id=\"T_78b9c_row21_col28\" class=\"data row21 col28\" >-0.065640</td>\n",
       "      <td id=\"T_78b9c_row21_col29\" class=\"data row21 col29\" >0.144250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row22\" class=\"row_heading level0 row22\" >V23</th>\n",
       "      <td id=\"T_78b9c_row22_col0\" class=\"data row22 col0\" >0.017551</td>\n",
       "      <td id=\"T_78b9c_row22_col1\" class=\"data row22 col1\" >0.015923</td>\n",
       "      <td id=\"T_78b9c_row22_col2\" class=\"data row22 col2\" >0.001575</td>\n",
       "      <td id=\"T_78b9c_row22_col3\" class=\"data row22 col3\" >-0.004866</td>\n",
       "      <td id=\"T_78b9c_row22_col4\" class=\"data row22 col4\" >-0.009856</td>\n",
       "      <td id=\"T_78b9c_row22_col5\" class=\"data row22 col5\" >0.004990</td>\n",
       "      <td id=\"T_78b9c_row22_col6\" class=\"data row22 col6\" >0.011034</td>\n",
       "      <td id=\"T_78b9c_row22_col7\" class=\"data row22 col7\" >0.013885</td>\n",
       "      <td id=\"T_78b9c_row22_col8\" class=\"data row22 col8\" >-0.008715</td>\n",
       "      <td id=\"T_78b9c_row22_col9\" class=\"data row22 col9\" >-0.015512</td>\n",
       "      <td id=\"T_78b9c_row22_col10\" class=\"data row22 col10\" >-0.001902</td>\n",
       "      <td id=\"T_78b9c_row22_col11\" class=\"data row22 col11\" >-0.006749</td>\n",
       "      <td id=\"T_78b9c_row22_col12\" class=\"data row22 col12\" >-0.008278</td>\n",
       "      <td id=\"T_78b9c_row22_col13\" class=\"data row22 col13\" >0.002720</td>\n",
       "      <td id=\"T_78b9c_row22_col14\" class=\"data row22 col14\" >-0.007571</td>\n",
       "      <td id=\"T_78b9c_row22_col15\" class=\"data row22 col15\" >-0.005070</td>\n",
       "      <td id=\"T_78b9c_row22_col16\" class=\"data row22 col16\" >0.005391</td>\n",
       "      <td id=\"T_78b9c_row22_col17\" class=\"data row22 col17\" >0.007771</td>\n",
       "      <td id=\"T_78b9c_row22_col18\" class=\"data row22 col18\" >-0.006434</td>\n",
       "      <td id=\"T_78b9c_row22_col19\" class=\"data row22 col19\" >0.054189</td>\n",
       "      <td id=\"T_78b9c_row22_col20\" class=\"data row22 col20\" >0.015318</td>\n",
       "      <td id=\"T_78b9c_row22_col21\" class=\"data row22 col21\" >-0.002295</td>\n",
       "      <td id=\"T_78b9c_row22_col22\" class=\"data row22 col22\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row22_col23\" class=\"data row22 col23\" >-0.007131</td>\n",
       "      <td id=\"T_78b9c_row22_col24\" class=\"data row22 col24\" >0.007478</td>\n",
       "      <td id=\"T_78b9c_row22_col25\" class=\"data row22 col25\" >0.001616</td>\n",
       "      <td id=\"T_78b9c_row22_col26\" class=\"data row22 col26\" >-0.033673</td>\n",
       "      <td id=\"T_78b9c_row22_col27\" class=\"data row22 col27\" >0.002355</td>\n",
       "      <td id=\"T_78b9c_row22_col28\" class=\"data row22 col28\" >-0.109790</td>\n",
       "      <td id=\"T_78b9c_row22_col29\" class=\"data row22 col29\" >0.050357</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row23\" class=\"row_heading level0 row23\" >V24</th>\n",
       "      <td id=\"T_78b9c_row23_col0\" class=\"data row23 col0\" >-0.001289</td>\n",
       "      <td id=\"T_78b9c_row23_col1\" class=\"data row23 col1\" >-0.004967</td>\n",
       "      <td id=\"T_78b9c_row23_col2\" class=\"data row23 col2\" >0.003254</td>\n",
       "      <td id=\"T_78b9c_row23_col3\" class=\"data row23 col3\" >0.001741</td>\n",
       "      <td id=\"T_78b9c_row23_col4\" class=\"data row23 col4\" >0.004302</td>\n",
       "      <td id=\"T_78b9c_row23_col5\" class=\"data row23 col5\" >-0.001185</td>\n",
       "      <td id=\"T_78b9c_row23_col6\" class=\"data row23 col6\" >-0.001084</td>\n",
       "      <td id=\"T_78b9c_row23_col7\" class=\"data row23 col7\" >0.000094</td>\n",
       "      <td id=\"T_78b9c_row23_col8\" class=\"data row23 col8\" >0.002114</td>\n",
       "      <td id=\"T_78b9c_row23_col9\" class=\"data row23 col9\" >0.004706</td>\n",
       "      <td id=\"T_78b9c_row23_col10\" class=\"data row23 col10\" >-0.003079</td>\n",
       "      <td id=\"T_78b9c_row23_col11\" class=\"data row23 col11\" >0.000183</td>\n",
       "      <td id=\"T_78b9c_row23_col12\" class=\"data row23 col12\" >-0.001882</td>\n",
       "      <td id=\"T_78b9c_row23_col13\" class=\"data row23 col13\" >-0.004254</td>\n",
       "      <td id=\"T_78b9c_row23_col14\" class=\"data row23 col14\" >0.000642</td>\n",
       "      <td id=\"T_78b9c_row23_col15\" class=\"data row23 col15\" >0.000735</td>\n",
       "      <td id=\"T_78b9c_row23_col16\" class=\"data row23 col16\" >-0.002260</td>\n",
       "      <td id=\"T_78b9c_row23_col17\" class=\"data row23 col17\" >-0.001614</td>\n",
       "      <td id=\"T_78b9c_row23_col18\" class=\"data row23 col18\" >0.001900</td>\n",
       "      <td id=\"T_78b9c_row23_col19\" class=\"data row23 col19\" >-0.002105</td>\n",
       "      <td id=\"T_78b9c_row23_col20\" class=\"data row23 col20\" >0.003217</td>\n",
       "      <td id=\"T_78b9c_row23_col21\" class=\"data row23 col21\" >0.001556</td>\n",
       "      <td id=\"T_78b9c_row23_col22\" class=\"data row23 col22\" >-0.007131</td>\n",
       "      <td id=\"T_78b9c_row23_col23\" class=\"data row23 col23\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row23_col24\" class=\"data row23 col24\" >-0.000471</td>\n",
       "      <td id=\"T_78b9c_row23_col25\" class=\"data row23 col25\" >0.003388</td>\n",
       "      <td id=\"T_78b9c_row23_col26\" class=\"data row23 col26\" >-0.000067</td>\n",
       "      <td id=\"T_78b9c_row23_col27\" class=\"data row23 col27\" >0.008187</td>\n",
       "      <td id=\"T_78b9c_row23_col28\" class=\"data row23 col28\" >0.004653</td>\n",
       "      <td id=\"T_78b9c_row23_col29\" class=\"data row23 col29\" >-0.015530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row24\" class=\"row_heading level0 row24\" >V25</th>\n",
       "      <td id=\"T_78b9c_row24_col0\" class=\"data row24 col0\" >0.002329</td>\n",
       "      <td id=\"T_78b9c_row24_col1\" class=\"data row24 col1\" >-0.003145</td>\n",
       "      <td id=\"T_78b9c_row24_col2\" class=\"data row24 col2\" >0.000031</td>\n",
       "      <td id=\"T_78b9c_row24_col3\" class=\"data row24 col3\" >0.001838</td>\n",
       "      <td id=\"T_78b9c_row24_col4\" class=\"data row24 col4\" >-0.001611</td>\n",
       "      <td id=\"T_78b9c_row24_col5\" class=\"data row24 col5\" >-0.000470</td>\n",
       "      <td id=\"T_78b9c_row24_col6\" class=\"data row24 col6\" >0.000420</td>\n",
       "      <td id=\"T_78b9c_row24_col7\" class=\"data row24 col7\" >-0.003231</td>\n",
       "      <td id=\"T_78b9c_row24_col8\" class=\"data row24 col8\" >-0.001449</td>\n",
       "      <td id=\"T_78b9c_row24_col9\" class=\"data row24 col9\" >-0.000694</td>\n",
       "      <td id=\"T_78b9c_row24_col10\" class=\"data row24 col10\" >-0.003555</td>\n",
       "      <td id=\"T_78b9c_row24_col11\" class=\"data row24 col11\" >-0.000906</td>\n",
       "      <td id=\"T_78b9c_row24_col12\" class=\"data row24 col12\" >-0.002112</td>\n",
       "      <td id=\"T_78b9c_row24_col13\" class=\"data row24 col13\" >0.004171</td>\n",
       "      <td id=\"T_78b9c_row24_col14\" class=\"data row24 col14\" >0.000280</td>\n",
       "      <td id=\"T_78b9c_row24_col15\" class=\"data row24 col15\" >-0.004086</td>\n",
       "      <td id=\"T_78b9c_row24_col16\" class=\"data row24 col16\" >-0.004748</td>\n",
       "      <td id=\"T_78b9c_row24_col17\" class=\"data row24 col17\" >0.001884</td>\n",
       "      <td id=\"T_78b9c_row24_col18\" class=\"data row24 col18\" >0.000290</td>\n",
       "      <td id=\"T_78b9c_row24_col19\" class=\"data row24 col19\" >0.013497</td>\n",
       "      <td id=\"T_78b9c_row24_col20\" class=\"data row24 col20\" >0.008225</td>\n",
       "      <td id=\"T_78b9c_row24_col21\" class=\"data row24 col21\" >0.002036</td>\n",
       "      <td id=\"T_78b9c_row24_col22\" class=\"data row24 col22\" >0.007478</td>\n",
       "      <td id=\"T_78b9c_row24_col23\" class=\"data row24 col23\" >-0.000471</td>\n",
       "      <td id=\"T_78b9c_row24_col24\" class=\"data row24 col24\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row24_col25\" class=\"data row24 col25\" >0.002175</td>\n",
       "      <td id=\"T_78b9c_row24_col26\" class=\"data row24 col26\" >-0.016789</td>\n",
       "      <td id=\"T_78b9c_row24_col27\" class=\"data row24 col27\" >0.015340</td>\n",
       "      <td id=\"T_78b9c_row24_col28\" class=\"data row24 col28\" >-0.048192</td>\n",
       "      <td id=\"T_78b9c_row24_col29\" class=\"data row24 col29\" >-0.231473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row25\" class=\"row_heading level0 row25\" >V26</th>\n",
       "      <td id=\"T_78b9c_row25_col0\" class=\"data row25 col0\" >-0.002013</td>\n",
       "      <td id=\"T_78b9c_row25_col1\" class=\"data row25 col1\" >0.000796</td>\n",
       "      <td id=\"T_78b9c_row25_col2\" class=\"data row25 col2\" >0.000671</td>\n",
       "      <td id=\"T_78b9c_row25_col3\" class=\"data row25 col3\" >0.002832</td>\n",
       "      <td id=\"T_78b9c_row25_col4\" class=\"data row25 col4\" >-0.000664</td>\n",
       "      <td id=\"T_78b9c_row25_col5\" class=\"data row25 col5\" >-0.003615</td>\n",
       "      <td id=\"T_78b9c_row25_col6\" class=\"data row25 col6\" >0.004651</td>\n",
       "      <td id=\"T_78b9c_row25_col7\" class=\"data row25 col7\" >0.003217</td>\n",
       "      <td id=\"T_78b9c_row25_col8\" class=\"data row25 col8\" >0.001348</td>\n",
       "      <td id=\"T_78b9c_row25_col9\" class=\"data row25 col9\" >-0.000581</td>\n",
       "      <td id=\"T_78b9c_row25_col10\" class=\"data row25 col10\" >0.001169</td>\n",
       "      <td id=\"T_78b9c_row25_col11\" class=\"data row25 col11\" >0.000006</td>\n",
       "      <td id=\"T_78b9c_row25_col12\" class=\"data row25 col12\" >-0.001993</td>\n",
       "      <td id=\"T_78b9c_row25_col13\" class=\"data row25 col13\" >-0.001902</td>\n",
       "      <td id=\"T_78b9c_row25_col14\" class=\"data row25 col14\" >0.001081</td>\n",
       "      <td id=\"T_78b9c_row25_col15\" class=\"data row25 col15\" >0.001122</td>\n",
       "      <td id=\"T_78b9c_row25_col16\" class=\"data row25 col16\" >0.001256</td>\n",
       "      <td id=\"T_78b9c_row25_col17\" class=\"data row25 col17\" >-0.000001</td>\n",
       "      <td id=\"T_78b9c_row25_col18\" class=\"data row25 col18\" >0.001928</td>\n",
       "      <td id=\"T_78b9c_row25_col19\" class=\"data row25 col19\" >-0.000066</td>\n",
       "      <td id=\"T_78b9c_row25_col20\" class=\"data row25 col20\" >0.001941</td>\n",
       "      <td id=\"T_78b9c_row25_col21\" class=\"data row25 col21\" >-0.003430</td>\n",
       "      <td id=\"T_78b9c_row25_col22\" class=\"data row25 col22\" >0.001616</td>\n",
       "      <td id=\"T_78b9c_row25_col23\" class=\"data row25 col23\" >0.003388</td>\n",
       "      <td id=\"T_78b9c_row25_col24\" class=\"data row25 col24\" >0.002175</td>\n",
       "      <td id=\"T_78b9c_row25_col25\" class=\"data row25 col25\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row25_col26\" class=\"data row25 col26\" >-0.007480</td>\n",
       "      <td id=\"T_78b9c_row25_col27\" class=\"data row25 col27\" >-0.001379</td>\n",
       "      <td id=\"T_78b9c_row25_col28\" class=\"data row25 col28\" >-0.004199</td>\n",
       "      <td id=\"T_78b9c_row25_col29\" class=\"data row25 col29\" >-0.042956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row26\" class=\"row_heading level0 row26\" >V27</th>\n",
       "      <td id=\"T_78b9c_row26_col0\" class=\"data row26 col0\" >0.004085</td>\n",
       "      <td id=\"T_78b9c_row26_col1\" class=\"data row26 col1\" >0.005265</td>\n",
       "      <td id=\"T_78b9c_row26_col2\" class=\"data row26 col2\" >0.003680</td>\n",
       "      <td id=\"T_78b9c_row26_col3\" class=\"data row26 col3\" >-0.007131</td>\n",
       "      <td id=\"T_78b9c_row26_col4\" class=\"data row26 col4\" >0.020649</td>\n",
       "      <td id=\"T_78b9c_row26_col5\" class=\"data row26 col5\" >-0.011486</td>\n",
       "      <td id=\"T_78b9c_row26_col6\" class=\"data row26 col6\" >-0.033952</td>\n",
       "      <td id=\"T_78b9c_row26_col7\" class=\"data row26 col7\" >-0.006134</td>\n",
       "      <td id=\"T_78b9c_row26_col8\" class=\"data row26 col8\" >-0.001733</td>\n",
       "      <td id=\"T_78b9c_row26_col9\" class=\"data row26 col9\" >-0.005133</td>\n",
       "      <td id=\"T_78b9c_row26_col10\" class=\"data row26 col10\" >-0.001217</td>\n",
       "      <td id=\"T_78b9c_row26_col11\" class=\"data row26 col11\" >-0.000261</td>\n",
       "      <td id=\"T_78b9c_row26_col12\" class=\"data row26 col12\" >-0.001064</td>\n",
       "      <td id=\"T_78b9c_row26_col13\" class=\"data row26 col13\" >0.001947</td>\n",
       "      <td id=\"T_78b9c_row26_col14\" class=\"data row26 col14\" >-0.002872</td>\n",
       "      <td id=\"T_78b9c_row26_col15\" class=\"data row26 col15\" >-0.008073</td>\n",
       "      <td id=\"T_78b9c_row26_col16\" class=\"data row26 col16\" >-0.002218</td>\n",
       "      <td id=\"T_78b9c_row26_col17\" class=\"data row26 col17\" >-0.004221</td>\n",
       "      <td id=\"T_78b9c_row26_col18\" class=\"data row26 col18\" >-0.006113</td>\n",
       "      <td id=\"T_78b9c_row26_col19\" class=\"data row26 col19\" >0.030510</td>\n",
       "      <td id=\"T_78b9c_row26_col20\" class=\"data row26 col20\" >0.006291</td>\n",
       "      <td id=\"T_78b9c_row26_col21\" class=\"data row26 col21\" >-0.006681</td>\n",
       "      <td id=\"T_78b9c_row26_col22\" class=\"data row26 col22\" >-0.033673</td>\n",
       "      <td id=\"T_78b9c_row26_col23\" class=\"data row26 col23\" >-0.000067</td>\n",
       "      <td id=\"T_78b9c_row26_col24\" class=\"data row26 col24\" >-0.016789</td>\n",
       "      <td id=\"T_78b9c_row26_col25\" class=\"data row26 col25\" >-0.007480</td>\n",
       "      <td id=\"T_78b9c_row26_col26\" class=\"data row26 col26\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row26_col27\" class=\"data row26 col27\" >-0.021681</td>\n",
       "      <td id=\"T_78b9c_row26_col28\" class=\"data row26 col28\" >0.013899</td>\n",
       "      <td id=\"T_78b9c_row26_col29\" class=\"data row26 col29\" >-0.004600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row27\" class=\"row_heading level0 row27\" >V28</th>\n",
       "      <td id=\"T_78b9c_row27_col0\" class=\"data row27 col0\" >-0.044459</td>\n",
       "      <td id=\"T_78b9c_row27_col1\" class=\"data row27 col1\" >-0.056137</td>\n",
       "      <td id=\"T_78b9c_row27_col2\" class=\"data row27 col2\" >-0.022925</td>\n",
       "      <td id=\"T_78b9c_row27_col3\" class=\"data row27 col3\" >0.025828</td>\n",
       "      <td id=\"T_78b9c_row27_col4\" class=\"data row27 col4\" >-0.004484</td>\n",
       "      <td id=\"T_78b9c_row27_col5\" class=\"data row27 col5\" >0.000046</td>\n",
       "      <td id=\"T_78b9c_row27_col6\" class=\"data row27 col6\" >0.006749</td>\n",
       "      <td id=\"T_78b9c_row27_col7\" class=\"data row27 col7\" >-0.014922</td>\n",
       "      <td id=\"T_78b9c_row27_col8\" class=\"data row27 col8\" >0.008454</td>\n",
       "      <td id=\"T_78b9c_row27_col9\" class=\"data row27 col9\" >0.002669</td>\n",
       "      <td id=\"T_78b9c_row27_col10\" class=\"data row27 col10\" >-0.001694</td>\n",
       "      <td id=\"T_78b9c_row27_col11\" class=\"data row27 col11\" >0.007021</td>\n",
       "      <td id=\"T_78b9c_row27_col12\" class=\"data row27 col12\" >0.007712</td>\n",
       "      <td id=\"T_78b9c_row27_col13\" class=\"data row27 col13\" >0.001494</td>\n",
       "      <td id=\"T_78b9c_row27_col14\" class=\"data row27 col14\" >0.010331</td>\n",
       "      <td id=\"T_78b9c_row27_col15\" class=\"data row27 col15\" >0.016978</td>\n",
       "      <td id=\"T_78b9c_row27_col16\" class=\"data row27 col16\" >-0.003540</td>\n",
       "      <td id=\"T_78b9c_row27_col17\" class=\"data row27 col17\" >0.000518</td>\n",
       "      <td id=\"T_78b9c_row27_col18\" class=\"data row27 col18\" >0.013287</td>\n",
       "      <td id=\"T_78b9c_row27_col19\" class=\"data row27 col19\" >-0.065891</td>\n",
       "      <td id=\"T_78b9c_row27_col20\" class=\"data row27 col20\" >-0.031159</td>\n",
       "      <td id=\"T_78b9c_row27_col21\" class=\"data row27 col21\" >0.002792</td>\n",
       "      <td id=\"T_78b9c_row27_col22\" class=\"data row27 col22\" >0.002355</td>\n",
       "      <td id=\"T_78b9c_row27_col23\" class=\"data row27 col23\" >0.008187</td>\n",
       "      <td id=\"T_78b9c_row27_col24\" class=\"data row27 col24\" >0.015340</td>\n",
       "      <td id=\"T_78b9c_row27_col25\" class=\"data row27 col25\" >-0.001379</td>\n",
       "      <td id=\"T_78b9c_row27_col26\" class=\"data row27 col26\" >-0.021681</td>\n",
       "      <td id=\"T_78b9c_row27_col27\" class=\"data row27 col27\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row27_col28\" class=\"data row27 col28\" >0.021446</td>\n",
       "      <td id=\"T_78b9c_row27_col29\" class=\"data row27 col29\" >-0.007932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row28\" class=\"row_heading level0 row28\" >V29</th>\n",
       "      <td id=\"T_78b9c_row28_col0\" class=\"data row28 col0\" >-0.231744</td>\n",
       "      <td id=\"T_78b9c_row28_col1\" class=\"data row28 col1\" >-0.530737</td>\n",
       "      <td id=\"T_78b9c_row28_col2\" class=\"data row28 col2\" >-0.207185</td>\n",
       "      <td id=\"T_78b9c_row28_col3\" class=\"data row28 col3\" >0.091913</td>\n",
       "      <td id=\"T_78b9c_row28_col4\" class=\"data row28 col4\" >-0.382747</td>\n",
       "      <td id=\"T_78b9c_row28_col5\" class=\"data row28 col5\" >0.210381</td>\n",
       "      <td id=\"T_78b9c_row28_col6\" class=\"data row28 col6\" >0.391239</td>\n",
       "      <td id=\"T_78b9c_row28_col7\" class=\"data row28 col7\" >-0.102777</td>\n",
       "      <td id=\"T_78b9c_row28_col8\" class=\"data row28 col8\" >-0.041503</td>\n",
       "      <td id=\"T_78b9c_row28_col9\" class=\"data row28 col9\" >-0.103121</td>\n",
       "      <td id=\"T_78b9c_row28_col10\" class=\"data row28 col10\" >-0.002312</td>\n",
       "      <td id=\"T_78b9c_row28_col11\" class=\"data row28 col11\" >-0.007580</td>\n",
       "      <td id=\"T_78b9c_row28_col12\" class=\"data row28 col12\" >0.001035</td>\n",
       "      <td id=\"T_78b9c_row28_col13\" class=\"data row28 col13\" >0.033107</td>\n",
       "      <td id=\"T_78b9c_row28_col14\" class=\"data row28 col14\" >-0.003503</td>\n",
       "      <td id=\"T_78b9c_row28_col15\" class=\"data row28 col15\" >-0.004419</td>\n",
       "      <td id=\"T_78b9c_row28_col16\" class=\"data row28 col16\" >0.011952</td>\n",
       "      <td id=\"T_78b9c_row28_col17\" class=\"data row28 col17\" >0.033585</td>\n",
       "      <td id=\"T_78b9c_row28_col18\" class=\"data row28 col18\" >-0.051163</td>\n",
       "      <td id=\"T_78b9c_row28_col19\" class=\"data row28 col19\" >0.348219</td>\n",
       "      <td id=\"T_78b9c_row28_col20\" class=\"data row28 col20\" >0.114076</td>\n",
       "      <td id=\"T_78b9c_row28_col21\" class=\"data row28 col21\" >-0.065640</td>\n",
       "      <td id=\"T_78b9c_row28_col22\" class=\"data row28 col22\" >-0.109790</td>\n",
       "      <td id=\"T_78b9c_row28_col23\" class=\"data row28 col23\" >0.004653</td>\n",
       "      <td id=\"T_78b9c_row28_col24\" class=\"data row28 col24\" >-0.048192</td>\n",
       "      <td id=\"T_78b9c_row28_col25\" class=\"data row28 col25\" >-0.004199</td>\n",
       "      <td id=\"T_78b9c_row28_col26\" class=\"data row28 col26\" >0.013899</td>\n",
       "      <td id=\"T_78b9c_row28_col27\" class=\"data row28 col27\" >0.021446</td>\n",
       "      <td id=\"T_78b9c_row28_col28\" class=\"data row28 col28\" >1.000000</td>\n",
       "      <td id=\"T_78b9c_row28_col29\" class=\"data row28 col29\" >-0.012268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_78b9c_level0_row29\" class=\"row_heading level0 row29\" >V30</th>\n",
       "      <td id=\"T_78b9c_row29_col0\" class=\"data row29 col0\" >0.118519</td>\n",
       "      <td id=\"T_78b9c_row29_col1\" class=\"data row29 col1\" >-0.008336</td>\n",
       "      <td id=\"T_78b9c_row29_col2\" class=\"data row29 col2\" >-0.429787</td>\n",
       "      <td id=\"T_78b9c_row29_col3\" class=\"data row29 col3\" >-0.103017</td>\n",
       "      <td id=\"T_78b9c_row29_col4\" class=\"data row29 col4\" >0.175038</td>\n",
       "      <td id=\"T_78b9c_row29_col5\" class=\"data row29 col5\" >-0.061977</td>\n",
       "      <td id=\"T_78b9c_row29_col6\" class=\"data row29 col6\" >0.084646</td>\n",
       "      <td id=\"T_78b9c_row29_col7\" class=\"data row29 col7\" >-0.036350</td>\n",
       "      <td id=\"T_78b9c_row29_col8\" class=\"data row29 col8\" >-0.007580</td>\n",
       "      <td id=\"T_78b9c_row29_col9\" class=\"data row29 col9\" >0.029717</td>\n",
       "      <td id=\"T_78b9c_row29_col10\" class=\"data row29 col10\" >-0.247051</td>\n",
       "      <td id=\"T_78b9c_row29_col11\" class=\"data row29 col11\" >0.126405</td>\n",
       "      <td id=\"T_78b9c_row29_col12\" class=\"data row29 col12\" >-0.063595</td>\n",
       "      <td id=\"T_78b9c_row29_col13\" class=\"data row29 col13\" >-0.104093</td>\n",
       "      <td id=\"T_78b9c_row29_col14\" class=\"data row29 col14\" >-0.181832</td>\n",
       "      <td id=\"T_78b9c_row29_col15\" class=\"data row29 col15\" >0.012150</td>\n",
       "      <td id=\"T_78b9c_row29_col16\" class=\"data row29 col16\" >-0.080489</td>\n",
       "      <td id=\"T_78b9c_row29_col17\" class=\"data row29 col17\" >0.089812</td>\n",
       "      <td id=\"T_78b9c_row29_col18\" class=\"data row29 col18\" >0.030244</td>\n",
       "      <td id=\"T_78b9c_row29_col19\" class=\"data row29 col19\" >-0.049359</td>\n",
       "      <td id=\"T_78b9c_row29_col20\" class=\"data row29 col20\" >0.046886</td>\n",
       "      <td id=\"T_78b9c_row29_col21\" class=\"data row29 col21\" >0.144250</td>\n",
       "      <td id=\"T_78b9c_row29_col22\" class=\"data row29 col22\" >0.050357</td>\n",
       "      <td id=\"T_78b9c_row29_col23\" class=\"data row29 col23\" >-0.015530</td>\n",
       "      <td id=\"T_78b9c_row29_col24\" class=\"data row29 col24\" >-0.231473</td>\n",
       "      <td id=\"T_78b9c_row29_col25\" class=\"data row29 col25\" >-0.042956</td>\n",
       "      <td id=\"T_78b9c_row29_col26\" class=\"data row29 col26\" >-0.004600</td>\n",
       "      <td id=\"T_78b9c_row29_col27\" class=\"data row29 col27\" >-0.007932</td>\n",
       "      <td id=\"T_78b9c_row29_col28\" class=\"data row29 col28\" >-0.012268</td>\n",
       "      <td id=\"T_78b9c_row29_col29\" class=\"data row29 col29\" >1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f468e0c46a0>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Correlation table color-coded\n",
    "def color_(value):\n",
    "    if value < 0 :\n",
    "        color = 'red'\n",
    "    elif value == 1 :\n",
    "        color = 'blue'\n",
    "    else:\n",
    "        color = 'green'\n",
    "    return 'color: %s' % color\n",
    "train_df.corr().style.applymap(color_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "QVCFv333tgC2"
   },
   "outputs": [],
   "source": [
    "# Excluding V29 since it contains outliers \n",
    "train_df.drop('V29', axis=1, inplace = True)\n",
    "test_df.drop('V29', axis=1, inplace = True)\n",
    "val_df.drop('V29', axis=1, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 550
    },
    "id": "cThYjnp0hcw2",
    "outputId": "fbdffbff-98ff-4dff-a266-a3c2f7c1fb11"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 540x540 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = sns.PairGrid(train_df.iloc[:,:3].sample(600, random_state=42))\n",
    "plt.subplots_adjust(top=0.9)\n",
    "g.fig.suptitle('Before:')\n",
    "g.map_diag(sns.kdeplot)\n",
    "g.map_offdiag(sns.kdeplot);\n",
    "# source: https://www.kaggle.com/code/robinteuwens/anomaly-detection-with-auto-encoders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "RdsrVfJmgaHf",
    "outputId": "f2803b99-8262-463f-c3a7-cba5486994f0"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(steps=[('normalizer', Normalizer()), ('scaler', MinMaxScaler())])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Normalizing & Standardizing\n",
    "pipeline = Pipeline([('normalizer', Normalizer()),\n",
    "                     ('scaler', MinMaxScaler())])\n",
    "\n",
    "# Get normalization parameters by fitting to the training data\n",
    "pipeline.fit(train_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "HwU0Ti9zgiqI"
   },
   "outputs": [],
   "source": [
    "# Apply\n",
    "train_df.iloc[:,:] = pipeline.transform(train_df)\n",
    "test_df.iloc[:,:] = pipeline.transform(test_df)\n",
    "val_df.iloc[:,:] = pipeline.transform(val_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 550
    },
    "id": "0iViyFHghoqY",
    "outputId": "1a9f3bec-69c0-4bf2-f58a-41e5214568fb"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 540x540 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = sns.PairGrid(train_df.iloc[:,:3].sample(600, random_state=42))\n",
    "plt.subplots_adjust(top=0.9)\n",
    "g.fig.suptitle('After:')\n",
    "g.map_diag(sns.kdeplot)\n",
    "g.map_offdiag(sns.kdeplot);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Cvgm9zxfhKHb"
   },
   "source": [
    "## Models\n",
    "* Tested 4 models commonly used for anomaly detection\n",
    "* Based on the f1 score from the validation dataset, the best performing model out of the 4 was chosen for final prediction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "XdHs87cnmQM_"
   },
   "outputs": [],
   "source": [
    "# Function for confusion matrix visualization\n",
    "def plot_confusion_matrix(cm,target_names,title='Confusion matrix',cmap=None,\n",
    "                          normalize=False):\n",
    "    import itertools\n",
    "    accuracy = np.trace(cm) / float(np.sum(cm))\n",
    "    misclass = 1 - accuracy\n",
    "\n",
    "    if cmap is None:\n",
    "        cmap = plt.get_cmap('Blues')\n",
    "\n",
    "    plt.figure(figsize=(8, 6))\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "\n",
    "    if target_names is not None:\n",
    "        tick_marks = np.arange(len(target_names))\n",
    "        plt.xticks(tick_marks, target_names, rotation=45)\n",
    "        plt.yticks(tick_marks, target_names)\n",
    "\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        \n",
    "    thresh = cm.max() / 1.5 if normalize else cm.max() / 2\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        if normalize:\n",
    "            plt.text(j, i, \"{:0.4f}\".format(cm[i, j]),\n",
    "                     horizontalalignment=\"center\",\n",
    "                     color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "        else:\n",
    "            plt.text(j, i, \"{:,}\".format(cm[i, j]),\n",
    "                     horizontalalignment=\"center\",\n",
    "                     color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label\\naccuracy={:0.4f}; misclass={:0.4f}'.format(accuracy, misclass))\n",
    "    plt.show()\n",
    "# Source: https://www.kaggle.com/code/drscarlat/compare-6-unsupervised-anomaly-detection-models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "j10XLOMQnF4Y",
    "outputId": "1e412fe3-004f-4623-e7c7-dd94fe993a39"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    28432\n",
       "1       30\n",
       "Name: Class, dtype: int64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val['Class'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NnvBrnQ2q7N4"
   },
   "source": [
    "### Isolation Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "EagLVno7L4sE",
    "outputId": "c3962f4e-9e8e-4626-9fad-914f17a4806f"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but IsolationForest was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro f1 score: 0.49900\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     28432\n",
      "           1       0.00      0.00      0.00        30\n",
      "\n",
      "    accuracy                           1.00     28462\n",
      "   macro avg       0.50      0.50      0.50     28462\n",
      "weighted avg       1.00      1.00      1.00     28462\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Test inital model\n",
    "IF = IsolationForest(n_estimators=150, max_samples=len(train_df), contamination = val_contamination, max_features = 1, random_state=77)\n",
    "IF.fit(train_df)\n",
    "\n",
    "pred = IF.predict(val_df)\n",
    "y = val['Class']\n",
    "\n",
    "pred_if = [1 if i==-1 else 0 for i in pred]\n",
    "val_score = f1_score(y, pred_if, average='macro')\n",
    "\n",
    "print(\"Macro f1 score: %.5f\" %(val_score))\n",
    "print(classification_report(y, pred_if))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 467
    },
    "id": "wpWldY0ZmwPn",
    "outputId": "831a6fce-3b0f-4087-95e8-a89757c93810"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "CM=confusion_matrix(y, pred_if)\n",
    "plot_confusion_matrix(CM, \n",
    "                      normalize    = False,\n",
    "                      target_names = ['Normal', 'Fraud'],\n",
    "                      title        = \"IsolationForest Confusion Matrix\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "S6mXVHPiqvwe"
   },
   "source": [
    "### Local Outlier Factor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "TkDdVp2-hdir",
    "outputId": "6b057c85-7b4d-455b-b720-611e5453e228"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro f1 score: 0.49951\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     28432\n",
      "           1       0.00      0.00      0.00        30\n",
      "\n",
      "    accuracy                           1.00     28462\n",
      "   macro avg       0.50      0.50      0.50     28462\n",
      "weighted avg       1.00      1.00      1.00     28462\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Test inital model\n",
    "LOF = LocalOutlierFactor(contamination = val_contamination, novelty=True)\n",
    "LOF.fit(train_df)\n",
    "\n",
    "pred = LOF.predict(val_df)\n",
    "y = val['Class'] \n",
    "\n",
    "pred_lof = [1 if i==-1 else 0 for i in pred]\n",
    "val_score = f1_score(y, pred_lof, average='macro')\n",
    "\n",
    "print(\"Macro f1 score: %.5f\" %(val_score))\n",
    "print(classification_report(y, pred_lof))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 467
    },
    "id": "Ikx20Ck7mqDI",
    "outputId": "4a90f299-6575-482b-b394-3344ba171098"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "CM=confusion_matrix(y, pred_lof)\n",
    "plot_confusion_matrix(CM, \n",
    "                      normalize    = False,\n",
    "                      target_names = ['Normal', 'Fraud'],\n",
    "                      title        = \"LocalOutlierFactor Confusion Matrix\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 557
    },
    "id": "3MKMGJ59qAKZ",
    "outputId": "7d522aad-c667-45da-c21a-caf78da48c71"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.7537  with num neighbors  1000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.7537  with num neighbors  1100\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.7537  with num neighbors  1200\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.7537  with num neighbors  1300\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.7578  with num neighbors  1400\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/sklearn/base.py:450: UserWarning: X does not have valid feature names, but LocalOutlierFactor was fitted with feature names\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.7537  with num neighbors  1500\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-dd2c4d25-61b6-49eb-a6f5-7ea8fe66b154\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>NumNeighb</th>\n",
       "      <th>F1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1000</td>\n",
       "      <td>0.7537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1100</td>\n",
       "      <td>0.7537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1200</td>\n",
       "      <td>0.7537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1300</td>\n",
       "      <td>0.7537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1400</td>\n",
       "      <td>0.7578</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-dd2c4d25-61b6-49eb-a6f5-7ea8fe66b154')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-dd2c4d25-61b6-49eb-a6f5-7ea8fe66b154 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-dd2c4d25-61b6-49eb-a6f5-7ea8fe66b154');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "   NumNeighb      F1\n",
       "0       1000  0.7537\n",
       "1       1100  0.7537\n",
       "2       1200  0.7537\n",
       "3       1300  0.7537\n",
       "4       1400  0.7578"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Optimize LOF hyperparameter n_neighbors for best f1\n",
    "\n",
    "minRE = 1000\n",
    "maxRE = 1600\n",
    "    \n",
    "LofF1 = []\n",
    "\n",
    "for TryRE in np.arange(minRE, maxRE, 100):\n",
    "  y = val['Class'] \n",
    "  clf = LocalOutlierFactor(n_neighbors=TryRE, contamination = val_contamination, novelty=True)\n",
    "  clf.fit(train_df)\n",
    "  y_pred = clf.predict(val_df)\n",
    "  \n",
    "  y_pred_lof = [1 if i==-1 else 0 for i in y_pred]\n",
    "  macro_f1 = f1_score(y, y_pred_lof, average='macro')\n",
    "  \n",
    "  print(\"F1 score on validation\", round(macro_f1,5), \" with num neighbors \", TryRE)\n",
    "  LofF1.append([TryRE, round(macro_f1,5)])\n",
    "    \n",
    "LofF1df = pd.DataFrame(LofF1, columns = ['NumNeighb', 'F1'])\n",
    "LofF1df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "3JmHV7DPrD4s",
    "outputId": "fc6865d5-835a-43c1-a0ae-0737f8b712b0"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "LofF1df.plot.line(\"NumNeighb\",\"F1\")\n",
    "plt.xlim(1000, 1600)\n",
    "plt.title(\"F1 vs NumNeighb\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hyEa-UxJq0fI"
   },
   "source": [
    "### Robust Covariance (Elliptic Envelope)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "15HSDzNCkWfO",
    "outputId": "fc06d733-d808-4392-bfe1-a37a80418f61"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro f1 score: 0.91658\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     28432\n",
      "           1       0.83      0.83      0.83        30\n",
      "\n",
      "    accuracy                           1.00     28462\n",
      "   macro avg       0.92      0.92      0.92     28462\n",
      "weighted avg       1.00      1.00      1.00     28462\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Test inital model\n",
    "EE = EllipticEnvelope(support_fraction = 0.994, contamination = val_contamination, random_state = 42)\n",
    "EE.fit(train_df)\n",
    "\n",
    "pred = EE.predict(val_df)\n",
    "y = val['Class'] \n",
    "\n",
    "pred_ee = [1 if i==-1 else 0 for i in pred]\n",
    "val_score_ee = f1_score(y, pred_ee, average='macro')\n",
    "\n",
    "print(\"Macro f1 score: %.5f\" %(val_score_ee))\n",
    "print(classification_report(y, pred_ee))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 467
    },
    "id": "9A6HKsQWmBf-",
    "outputId": "f96c5596-9d01-4a97-b7b4-f6b1c161f06d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "CM=confusion_matrix(y, pred_ee)\n",
    "plot_confusion_matrix(CM, \n",
    "                      normalize    = False,\n",
    "                      target_names = ['Normal', 'Fraud'],\n",
    "                      title        = \"EllipticEnvelope Confusion Matrix\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 298
    },
    "id": "Wymks4EInd3O",
    "outputId": "ea214dba-df00-4bc3-ac3f-3c4f8cfaa611"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score on validation 0.90668  with support_fraction  0.95\n",
      "F1 score on validation 0.90668  with support_fraction  0.96\n",
      "F1 score on validation 0.91658  with support_fraction  0.97\n",
      "F1 score on validation 0.91658  with support_fraction  0.98\n",
      "F1 score on validation 0.91658  with support_fraction  0.99\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-577ce939-c2f8-42e0-b7db-1f7f3c174a17\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>SupFrac</th>\n",
       "      <th>F1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.95</td>\n",
       "      <td>0.90668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.96</td>\n",
       "      <td>0.90668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.97</td>\n",
       "      <td>0.91658</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.98</td>\n",
       "      <td>0.91658</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.99</td>\n",
       "      <td>0.91658</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-577ce939-c2f8-42e0-b7db-1f7f3c174a17')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-577ce939-c2f8-42e0-b7db-1f7f3c174a17 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-577ce939-c2f8-42e0-b7db-1f7f3c174a17');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "   SupFrac       F1\n",
       "0     0.95  0.90668\n",
       "1     0.96  0.90668\n",
       "2     0.97  0.91658\n",
       "3     0.98  0.91658\n",
       "4     0.99  0.91658"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Optimize EE hyperparameter for best f1\n",
    "\n",
    "minRE = 0.95\n",
    "maxRE = 0.99\n",
    "    \n",
    "EpsF1 = []\n",
    "\n",
    "for TryRE in np.arange(minRE, maxRE, 0.01):\n",
    "  y = val['Class'] \n",
    "  cov = EllipticEnvelope(support_fraction = TryRE, contamination = val_contamination)\n",
    "  cov.fit(train_df)\n",
    "  y_pred = cov.predict(val_df)\n",
    "  y_pred_ee = [1 if i==-1 else 0 for i in y_pred]\n",
    "  macro_f1 = f1_score(y, y_pred_ee, average='macro')\n",
    "  \n",
    "  print(\"F1 score on validation\", round(macro_f1,5), \" with support_fraction \", TryRE)\n",
    "  EpsF1.append([TryRE, round(macro_f1,5)])\n",
    "    \n",
    "EpsF1df = pd.DataFrame(EpsF1, columns = ['SupFrac', 'F1'])\n",
    "EpsF1df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "YgdUK7k_k6vD",
    "outputId": "f5adf827-4759-4ffb-87af-8ca0c279de31"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "EpsF1df.plot.line(\"SupFrac\",\"F1\")\n",
    "plt.xlim(minRE, maxRE)\n",
    "plt.title(\"F1 vs SupFrac\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* The EE model no longer improves significantly with support fraction values larger than 0.99. After several experiments, I concluded that the model is best with support fraction of 0.994.\n",
    "* Setting `assume_cenetered=True` worsens the performance of the model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "PLvMtei1ebXA"
   },
   "source": [
    "### AutoEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "tmilYozqLlN6",
    "outputId": "004e75f3-e9ba-4852-a1ca-c9045635f0e0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense_14 (Dense)            (None, 29)                870       \n",
      "                                                                 \n",
      " dropout_4 (Dropout)         (None, 29)                0         \n",
      "                                                                 \n",
      " dense_15 (Dense)            (None, 29)                870       \n",
      "                                                                 \n",
      " dropout_5 (Dropout)         (None, 29)                0         \n",
      "                                                                 \n",
      " dense_16 (Dense)            (None, 128)               3840      \n",
      "                                                                 \n",
      " dropout_6 (Dropout)         (None, 128)               0         \n",
      "                                                                 \n",
      " dense_17 (Dense)            (None, 64)                8256      \n",
      "                                                                 \n",
      " dropout_7 (Dropout)         (None, 64)                0         \n",
      "                                                                 \n",
      " dense_18 (Dense)            (None, 32)                2080      \n",
      "                                                                 \n",
      " dropout_8 (Dropout)         (None, 32)                0         \n",
      "                                                                 \n",
      " dense_19 (Dense)            (None, 16)                528       \n",
      "                                                                 \n",
      " dropout_9 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_20 (Dense)            (None, 16)                272       \n",
      "                                                                 \n",
      " dropout_10 (Dropout)        (None, 16)                0         \n",
      "                                                                 \n",
      " dense_21 (Dense)            (None, 32)                544       \n",
      "                                                                 \n",
      " dropout_11 (Dropout)        (None, 32)                0         \n",
      "                                                                 \n",
      " dense_22 (Dense)            (None, 64)                2112      \n",
      "                                                                 \n",
      " dropout_12 (Dropout)        (None, 64)                0         \n",
      "                                                                 \n",
      " dense_23 (Dense)            (None, 128)               8320      \n",
      "                                                                 \n",
      " dropout_13 (Dropout)        (None, 128)               0         \n",
      "                                                                 \n",
      " dense_24 (Dense)            (None, 29)                3741      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 31,433\n",
      "Trainable params: 31,433\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "Epoch 1/20\n",
      "2847/2847 [==============================] - 10s 3ms/step - loss: 1.1712 - val_loss: 1.0192\n",
      "Epoch 2/20\n",
      "2847/2847 [==============================] - 9s 3ms/step - loss: 1.0145 - val_loss: 1.0060\n",
      "Epoch 3/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0061 - val_loss: 1.0012\n",
      "Epoch 4/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0029 - val_loss: 0.9994\n",
      "Epoch 5/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0017 - val_loss: 0.9986\n",
      "Epoch 6/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0012 - val_loss: 0.9983\n",
      "Epoch 7/20\n",
      "2847/2847 [==============================] - 13s 5ms/step - loss: 1.0009 - val_loss: 0.9981\n",
      "Epoch 8/20\n",
      "2847/2847 [==============================] - 10s 4ms/step - loss: 1.0008 - val_loss: 0.9981\n",
      "Epoch 9/20\n",
      "2847/2847 [==============================] - 9s 3ms/step - loss: 1.0007 - val_loss: 0.9980\n",
      "Epoch 10/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0006 - val_loss: 0.9979\n",
      "Epoch 11/20\n",
      "2847/2847 [==============================] - 9s 3ms/step - loss: 1.0006 - val_loss: 0.9979\n",
      "Epoch 12/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0006 - val_loss: 0.9979\n",
      "Epoch 13/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0006 - val_loss: 0.9979\n",
      "Epoch 14/20\n",
      "2847/2847 [==============================] - 10s 3ms/step - loss: 1.0006 - val_loss: 0.9979\n",
      "Epoch 15/20\n",
      "2847/2847 [==============================] - 9s 3ms/step - loss: 1.0006 - val_loss: 0.9979\n",
      "Epoch 16/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0005 - val_loss: 0.9979\n",
      "Epoch 17/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0005 - val_loss: 0.9979\n",
      "Epoch 18/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0005 - val_loss: 0.9979\n",
      "Epoch 19/20\n",
      "2847/2847 [==============================] - 8s 3ms/step - loss: 1.0005 - val_loss: 0.9979\n",
      "Epoch 20/20\n",
      "2847/2847 [==============================] - 10s 4ms/step - loss: 1.0005 - val_loss: 0.9979\n",
      "3558/3558 [==============================] - 4s 1ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "AutoEncoder(batch_size=32, contamination=0.0010551491277433877,\n",
       "      dropout_rate=0.2, epochs=20, hidden_activation='relu',\n",
       "      hidden_neurons=[128, 64, 32, 16, 16, 32, 64, 128],\n",
       "      l2_regularizer=0.1,\n",
       "      loss=<function mean_squared_error at 0x7f2032d70ca0>,\n",
       "      optimizer='adam', output_activation='sigmoid', preprocessing=True,\n",
       "      random_state=None, validation_size=0.2, verbose=1)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "atcdr = AutoEncoder(contamination=val_contamination,\n",
    "                    hidden_neurons =[128,64,32,16,16,32,64,128],\n",
    "                    validation_size = 0.2,\n",
    "                    epochs = 20)\n",
    "atcdr.fit(train_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "72grbVlfLsf7",
    "outputId": "c0dc595a-240a-4e36-c60f-ba87d3009663"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "890/890 [==============================] - 1s 1ms/step\n",
      "Macro f1 score: 0.49954\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     28432\n",
      "           1       0.00      0.00      0.00        30\n",
      "\n",
      "    accuracy                           1.00     28462\n",
      "   macro avg       0.50      0.50      0.50     28462\n",
      "weighted avg       1.00      1.00      1.00     28462\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y = val['Class']\n",
    "pred_ae = atcdr.predict(val_df)\n",
    "ae_score=f1_score(y, pred_ae, average='macro')\n",
    "\n",
    "print(\"Macro f1 score: %.5f\" %(ae_score))\n",
    "print(classification_report(y, pred_ae))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "qM602d-KOjrc"
   },
   "source": [
    "* Out of the 4 models, Elliptic Envelope predicts with the best validation f1 score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hsgiMI2xdafv"
   },
   "source": [
    "## Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "YLT-CU1OzUH8"
   },
   "outputs": [],
   "source": [
    "# Final prediction with EE\n",
    "cov = EllipticEnvelope(support_fraction = 0.994, contamination = val_contamination)\n",
    "cov.fit(train_df)\n",
    "pred = cov.predict(test_df)\n",
    "prediction = [1 if i==-1 else 0 for i in pred]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 424
    },
    "id": "GoIlB-Wli0aw",
    "outputId": "591688c0-5223-4395-8114-3068f76d8fff"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-deb3bc1c-cbf2-4b2e-8c6d-78f446837d1e\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AAAA0x1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AAAA0x2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AAAA0x5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AAAA0x7</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AAAA0xc</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142498</th>\n",
       "      <td>0x4587f</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142499</th>\n",
       "      <td>0x45880</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142500</th>\n",
       "      <td>0x45884</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142501</th>\n",
       "      <td>0x45885</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142502</th>\n",
       "      <td>0x45887</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>142503 rows × 2 columns</p>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-deb3bc1c-cbf2-4b2e-8c6d-78f446837d1e')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-deb3bc1c-cbf2-4b2e-8c6d-78f446837d1e button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-deb3bc1c-cbf2-4b2e-8c6d-78f446837d1e');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "             ID  Class\n",
       "0       AAAA0x1      0\n",
       "1       AAAA0x2      0\n",
       "2       AAAA0x5      0\n",
       "3       AAAA0x7      0\n",
       "4       AAAA0xc      0\n",
       "...         ...    ...\n",
       "142498  0x4587f      0\n",
       "142499  0x45880      0\n",
       "142500  0x45884      0\n",
       "142501  0x45885      0\n",
       "142502  0x45887      0\n",
       "\n",
       "[142503 rows x 2 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission = pd.read_csv('sample_submission.csv')\n",
    "submission['Class'] = prediction\n",
    "submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "gfs0tVI6ikFJ",
    "outputId": "d7eb76c5-84a6-4167-8109-a27b8847ebbc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    142205\n",
       "1       298\n",
       "Name: Class, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission['Class'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "wPg7SYVQjaSN"
   },
   "outputs": [],
   "source": [
    "submission.to_csv('sub.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QuuVO7BQGJMV"
   },
   "source": [
    "#### FInal scores\n",
    "* public: 0.9248241501 / private: 0.9105133393\n",
    "\n",
    "#### Takeaways\n",
    "* Using the same EE model, the score is better off when all 30 features are used for prediction.\n",
    "* The score improves when the features are NOT standardized/normalized\n",
    "* For this problem, tuning and finding the right kind of model may be more important than data preprocessing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "HG2uHDAb9bmz"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "provenance": []
  },
  "gpuClass": "standard",
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}