Social Media Analytics for Twitter.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "%pylab inline\n",
    "import seaborn as sns\n",
    "from pandas import read_csv\n",
    "from matplotlib import pyplot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# About the dataset\n",
    "\n",
    "The dataset has 11 variables describing each person's Twiiter Activity. These variables are: \n",
    "- Follower Count: Number of followers of the person on Twitter\n",
    "- Follwoing Count: Number of Twitter users the person follows \n",
    "- Listed Count: Number of people that have added the person to a list\n",
    "- Mentions Received: Number of times people have mentioned (@) the person\n",
    "- Retweets Received: Number of times a tweet was retweeted\n",
    "- Mentions Sent: Number of times the person mentioned (@) someone\n",
    "- Retweets Sent: Number of times the person retweet a tweet\n",
    "- Posts: Number of posts \n",
    "- Network Feature 1, 2, 3: Local follower network of the person  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Choice</th>\n",
       "      <th>A_follower_count</th>\n",
       "      <th>A_following_count</th>\n",
       "      <th>A_listed_count</th>\n",
       "      <th>A_mentions_received</th>\n",
       "      <th>A_retweets_received</th>\n",
       "      <th>A_mentions_sent</th>\n",
       "      <th>A_retweets_sent</th>\n",
       "      <th>A_posts</th>\n",
       "      <th>A_network_feature_1</th>\n",
       "      <th>...</th>\n",
       "      <th>B_following_count</th>\n",
       "      <th>B_listed_count</th>\n",
       "      <th>B_mentions_received</th>\n",
       "      <th>B_retweets_received</th>\n",
       "      <th>B_mentions_sent</th>\n",
       "      <th>B_retweets_sent</th>\n",
       "      <th>B_posts</th>\n",
       "      <th>B_network_feature_1</th>\n",
       "      <th>B_network_feature_2</th>\n",
       "      <th>B_network_feature_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>228</td>\n",
       "      <td>302</td>\n",
       "      <td>3</td>\n",
       "      <td>0.583979</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.362150</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>29808</td>\n",
       "      <td>1689</td>\n",
       "      <td>15.430498</td>\n",
       "      <td>3.984029</td>\n",
       "      <td>8.204331</td>\n",
       "      <td>0.332423</td>\n",
       "      <td>6.988815</td>\n",
       "      <td>66</td>\n",
       "      <td>75.530303</td>\n",
       "      <td>1916.893939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>21591</td>\n",
       "      <td>1179</td>\n",
       "      <td>228</td>\n",
       "      <td>90.456506</td>\n",
       "      <td>25.798292</td>\n",
       "      <td>5.709329</td>\n",
       "      <td>1.111159</td>\n",
       "      <td>5.176620</td>\n",
       "      <td>369</td>\n",
       "      <td>...</td>\n",
       "      <td>848</td>\n",
       "      <td>1610</td>\n",
       "      <td>40.495021</td>\n",
       "      <td>8.943607</td>\n",
       "      <td>3.227677</td>\n",
       "      <td>0.564343</td>\n",
       "      <td>1.070321</td>\n",
       "      <td>163</td>\n",
       "      <td>132.030675</td>\n",
       "      <td>2931.515337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>7310</td>\n",
       "      <td>1215</td>\n",
       "      <td>101</td>\n",
       "      <td>25.503644</td>\n",
       "      <td>9.556347</td>\n",
       "      <td>5.361519</td>\n",
       "      <td>0.591206</td>\n",
       "      <td>3.589718</td>\n",
       "      <td>95</td>\n",
       "      <td>...</td>\n",
       "      <td>482</td>\n",
       "      <td>206</td>\n",
       "      <td>0.734696</td>\n",
       "      <td>0.354379</td>\n",
       "      <td>0.603202</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.603202</td>\n",
       "      <td>3</td>\n",
       "      <td>10.333333</td>\n",
       "      <td>277.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>7.690824</td>\n",
       "      <td>0.277306</td>\n",
       "      <td>1.331508</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>2.830627</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>17637</td>\n",
       "      <td>278</td>\n",
       "      <td>572.874856</td>\n",
       "      <td>390.293681</td>\n",
       "      <td>27.552040</td>\n",
       "      <td>7.167557</td>\n",
       "      <td>32.101906</td>\n",
       "      <td>1762</td>\n",
       "      <td>23.469296</td>\n",
       "      <td>1395.845634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>45589</td>\n",
       "      <td>862</td>\n",
       "      <td>2641</td>\n",
       "      <td>148.854279</td>\n",
       "      <td>36.998884</td>\n",
       "      <td>27.881768</td>\n",
       "      <td>3.333492</td>\n",
       "      <td>23.861282</td>\n",
       "      <td>551</td>\n",
       "      <td>...</td>\n",
       "      <td>1711</td>\n",
       "      <td>181</td>\n",
       "      <td>21.601866</td>\n",
       "      <td>3.581661</td>\n",
       "      <td>6.764657</td>\n",
       "      <td>1.119727</td>\n",
       "      <td>4.563246</td>\n",
       "      <td>85</td>\n",
       "      <td>48.500000</td>\n",
       "      <td>1993.627907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5495</th>\n",
       "      <td>0</td>\n",
       "      <td>41765</td>\n",
       "      <td>185</td>\n",
       "      <td>1356</td>\n",
       "      <td>1529.643058</td>\n",
       "      <td>282.858500</td>\n",
       "      <td>76.809514</td>\n",
       "      <td>5.392171</td>\n",
       "      <td>104.438625</td>\n",
       "      <td>3307</td>\n",
       "      <td>...</td>\n",
       "      <td>1370</td>\n",
       "      <td>15143</td>\n",
       "      <td>6881.482996</td>\n",
       "      <td>5013.642295</td>\n",
       "      <td>22.683131</td>\n",
       "      <td>2.385538</td>\n",
       "      <td>18.420981</td>\n",
       "      <td>23958</td>\n",
       "      <td>21.324988</td>\n",
       "      <td>1413.825398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5496</th>\n",
       "      <td>1</td>\n",
       "      <td>112</td>\n",
       "      <td>243</td>\n",
       "      <td>5</td>\n",
       "      <td>1.445174</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.603177</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>479</td>\n",
       "      <td>8</td>\n",
       "      <td>0.306073</td>\n",
       "      <td>0.306073</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>1.621981</td>\n",
       "      <td>1</td>\n",
       "      <td>49.000000</td>\n",
       "      <td>4969.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5497</th>\n",
       "      <td>0</td>\n",
       "      <td>15385</td>\n",
       "      <td>673</td>\n",
       "      <td>747</td>\n",
       "      <td>55.993546</td>\n",
       "      <td>22.321945</td>\n",
       "      <td>6.946233</td>\n",
       "      <td>0.341936</td>\n",
       "      <td>6.503977</td>\n",
       "      <td>202</td>\n",
       "      <td>...</td>\n",
       "      <td>1711</td>\n",
       "      <td>181</td>\n",
       "      <td>21.601866</td>\n",
       "      <td>3.581661</td>\n",
       "      <td>6.764657</td>\n",
       "      <td>1.119727</td>\n",
       "      <td>4.563246</td>\n",
       "      <td>85</td>\n",
       "      <td>48.500000</td>\n",
       "      <td>1993.627907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5498</th>\n",
       "      <td>0</td>\n",
       "      <td>265258</td>\n",
       "      <td>209</td>\n",
       "      <td>551</td>\n",
       "      <td>631.915946</td>\n",
       "      <td>457.648550</td>\n",
       "      <td>5.460985</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>7.498126</td>\n",
       "      <td>2603</td>\n",
       "      <td>...</td>\n",
       "      <td>37601</td>\n",
       "      <td>22232</td>\n",
       "      <td>176.533716</td>\n",
       "      <td>42.330221</td>\n",
       "      <td>6.849227</td>\n",
       "      <td>1.345073</td>\n",
       "      <td>4.359865</td>\n",
       "      <td>696</td>\n",
       "      <td>234.338068</td>\n",
       "      <td>7168.464489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5499</th>\n",
       "      <td>0</td>\n",
       "      <td>628</td>\n",
       "      <td>921</td>\n",
       "      <td>6</td>\n",
       "      <td>3.943848</td>\n",
       "      <td>0.618590</td>\n",
       "      <td>4.769930</td>\n",
       "      <td>0.870136</td>\n",
       "      <td>2.953237</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>3109</td>\n",
       "      <td>12772</td>\n",
       "      <td>311.800311</td>\n",
       "      <td>142.633229</td>\n",
       "      <td>4.486090</td>\n",
       "      <td>1.049408</td>\n",
       "      <td>4.806823</td>\n",
       "      <td>1280</td>\n",
       "      <td>48.532819</td>\n",
       "      <td>1554.489575</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5500 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      Choice  A_follower_count  A_following_count  A_listed_count  \\\n",
       "0          0               228                302               3   \n",
       "1          0             21591               1179             228   \n",
       "2          0              7310               1215             101   \n",
       "3          0                20                  7               2   \n",
       "4          1             45589                862            2641   \n",
       "...      ...               ...                ...             ...   \n",
       "5495       0             41765                185            1356   \n",
       "5496       1               112                243               5   \n",
       "5497       0             15385                673             747   \n",
       "5498       0            265258                209             551   \n",
       "5499       0               628                921               6   \n",
       "\n",
       "      A_mentions_received  A_retweets_received  A_mentions_sent  \\\n",
       "0                0.583979             0.100503         0.100503   \n",
       "1               90.456506            25.798292         5.709329   \n",
       "2               25.503644             9.556347         5.361519   \n",
       "3                7.690824             0.277306         1.331508   \n",
       "4              148.854279            36.998884        27.881768   \n",
       "...                   ...                  ...              ...   \n",
       "5495          1529.643058           282.858500        76.809514   \n",
       "5496             1.445174             0.100503         0.100503   \n",
       "5497            55.993546            22.321945         6.946233   \n",
       "5498           631.915946           457.648550         5.460985   \n",
       "5499             3.943848             0.618590         4.769930   \n",
       "\n",
       "      A_retweets_sent     A_posts  A_network_feature_1  ...  \\\n",
       "0            0.100503    0.362150                    2  ...   \n",
       "1            1.111159    5.176620                  369  ...   \n",
       "2            0.591206    3.589718                   95  ...   \n",
       "3            0.100503    2.830627                    6  ...   \n",
       "4            3.333492   23.861282                  551  ...   \n",
       "...               ...         ...                  ...  ...   \n",
       "5495         5.392171  104.438625                 3307  ...   \n",
       "5496         0.100503    0.603177                    5  ...   \n",
       "5497         0.341936    6.503977                  202  ...   \n",
       "5498         0.100503    7.498126                 2603  ...   \n",
       "5499         0.870136    2.953237                    8  ...   \n",
       "\n",
       "      B_following_count  B_listed_count  B_mentions_received  \\\n",
       "0                 29808            1689            15.430498   \n",
       "1                   848            1610            40.495021   \n",
       "2                   482             206             0.734696   \n",
       "3                 17637             278           572.874856   \n",
       "4                  1711             181            21.601866   \n",
       "...                 ...             ...                  ...   \n",
       "5495               1370           15143          6881.482996   \n",
       "5496                479               8             0.306073   \n",
       "5497               1711             181            21.601866   \n",
       "5498              37601           22232           176.533716   \n",
       "5499               3109           12772           311.800311   \n",
       "\n",
       "      B_retweets_received  B_mentions_sent  B_retweets_sent    B_posts  \\\n",
       "0                3.984029         8.204331         0.332423   6.988815   \n",
       "1                8.943607         3.227677         0.564343   1.070321   \n",
       "2                0.354379         0.603202         0.100503   0.603202   \n",
       "3              390.293681        27.552040         7.167557  32.101906   \n",
       "4                3.581661         6.764657         1.119727   4.563246   \n",
       "...                   ...              ...              ...        ...   \n",
       "5495          5013.642295        22.683131         2.385538  18.420981   \n",
       "5496             0.306073         0.100503         0.100503   1.621981   \n",
       "5497             3.581661         6.764657         1.119727   4.563246   \n",
       "5498            42.330221         6.849227         1.345073   4.359865   \n",
       "5499           142.633229         4.486090         1.049408   4.806823   \n",
       "\n",
       "      B_network_feature_1  B_network_feature_2  B_network_feature_3  \n",
       "0                      66            75.530303          1916.893939  \n",
       "1                     163           132.030675          2931.515337  \n",
       "2                       3            10.333333           277.333333  \n",
       "3                    1762            23.469296          1395.845634  \n",
       "4                      85            48.500000          1993.627907  \n",
       "...                   ...                  ...                  ...  \n",
       "5495                23958            21.324988          1413.825398  \n",
       "5496                    1            49.000000          4969.000000  \n",
       "5497                   85            48.500000          1993.627907  \n",
       "5498                  696           234.338068          7168.464489  \n",
       "5499                 1280            48.532819          1554.489575  \n",
       "\n",
       "[5500 rows x 23 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load data \n",
    "Twitter_Train = pd.read_csv(\"/Users/shimonyagrawal/Desktop/Social Media Analytics/train.csv\")\n",
    "Twitter_Train "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 5500 entries, 0 to 5499\n",
      "Data columns (total 23 columns):\n",
      " #   Column               Non-Null Count  Dtype  \n",
      "---  ------               --------------  -----  \n",
      " 0   Choice               5500 non-null   int64  \n",
      " 1   A_follower_count     5500 non-null   int64  \n",
      " 2   A_following_count    5500 non-null   int64  \n",
      " 3   A_listed_count       5500 non-null   int64  \n",
      " 4   A_mentions_received  5500 non-null   float64\n",
      " 5   A_retweets_received  5500 non-null   float64\n",
      " 6   A_mentions_sent      5500 non-null   float64\n",
      " 7   A_retweets_sent      5500 non-null   float64\n",
      " 8   A_posts              5500 non-null   float64\n",
      " 9   A_network_feature_1  5500 non-null   int64  \n",
      " 10  A_network_feature_2  5500 non-null   float64\n",
      " 11  A_network_feature_3  5500 non-null   float64\n",
      " 12  B_follower_count     5500 non-null   int64  \n",
      " 13  B_following_count    5500 non-null   int64  \n",
      " 14  B_listed_count       5500 non-null   int64  \n",
      " 15  B_mentions_received  5500 non-null   float64\n",
      " 16  B_retweets_received  5500 non-null   float64\n",
      " 17  B_mentions_sent      5500 non-null   float64\n",
      " 18  B_retweets_sent      5500 non-null   float64\n",
      " 19  B_posts              5500 non-null   float64\n",
      " 20  B_network_feature_1  5500 non-null   int64  \n",
      " 21  B_network_feature_2  5500 non-null   float64\n",
      " 22  B_network_feature_3  5500 non-null   float64\n",
      "dtypes: float64(14), int64(9)\n",
      "memory usage: 988.4 KB\n"
     ]
    }
   ],
   "source": [
    "Twitter_Train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Choice</th>\n",
       "      <th>A_follower_count</th>\n",
       "      <th>A_following_count</th>\n",
       "      <th>A_listed_count</th>\n",
       "      <th>A_mentions_received</th>\n",
       "      <th>A_retweets_received</th>\n",
       "      <th>A_mentions_sent</th>\n",
       "      <th>A_retweets_sent</th>\n",
       "      <th>A_posts</th>\n",
       "      <th>A_network_feature_1</th>\n",
       "      <th>...</th>\n",
       "      <th>B_following_count</th>\n",
       "      <th>B_listed_count</th>\n",
       "      <th>B_mentions_received</th>\n",
       "      <th>B_retweets_received</th>\n",
       "      <th>B_mentions_sent</th>\n",
       "      <th>B_retweets_sent</th>\n",
       "      <th>B_posts</th>\n",
       "      <th>B_network_feature_1</th>\n",
       "      <th>B_network_feature_2</th>\n",
       "      <th>B_network_feature_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5.500000e+03</td>\n",
       "      <td>5.500000e+03</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5.500000e+03</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5.500000e+03</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "      <td>5500.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.509455</td>\n",
       "      <td>6.498840e+05</td>\n",
       "      <td>1.265895e+04</td>\n",
       "      <td>5952.453273</td>\n",
       "      <td>2.666032e+03</td>\n",
       "      <td>1032.371839</td>\n",
       "      <td>6.011873</td>\n",
       "      <td>1.109924</td>\n",
       "      <td>9.090730</td>\n",
       "      <td>5267.768000</td>\n",
       "      <td>...</td>\n",
       "      <td>12738.259818</td>\n",
       "      <td>5903.148364</td>\n",
       "      <td>2.554598e+03</td>\n",
       "      <td>997.149954</td>\n",
       "      <td>6.099658</td>\n",
       "      <td>1.106236</td>\n",
       "      <td>9.505821</td>\n",
       "      <td>5254.933636</td>\n",
       "      <td>85.024196</td>\n",
       "      <td>3745.175480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.499956</td>\n",
       "      <td>2.028787e+06</td>\n",
       "      <td>4.900867e+04</td>\n",
       "      <td>17339.141191</td>\n",
       "      <td>2.916543e+04</td>\n",
       "      <td>10954.953223</td>\n",
       "      <td>9.519797</td>\n",
       "      <td>1.910104</td>\n",
       "      <td>18.311060</td>\n",
       "      <td>28946.777345</td>\n",
       "      <td>...</td>\n",
       "      <td>50054.520874</td>\n",
       "      <td>16298.462018</td>\n",
       "      <td>2.508873e+04</td>\n",
       "      <td>9342.006880</td>\n",
       "      <td>9.729557</td>\n",
       "      <td>1.939914</td>\n",
       "      <td>19.424680</td>\n",
       "      <td>26778.820125</td>\n",
       "      <td>106.689384</td>\n",
       "      <td>5518.399590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.600000e+01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.005034e-01</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.005034e-01</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.663750e+03</td>\n",
       "      <td>3.220000e+02</td>\n",
       "      <td>85.000000</td>\n",
       "      <td>3.453649e+00</td>\n",
       "      <td>0.716816</td>\n",
       "      <td>0.359534</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.632440</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>322.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>3.260473e+00</td>\n",
       "      <td>0.714556</td>\n",
       "      <td>0.356943</td>\n",
       "      <td>0.100503</td>\n",
       "      <td>0.822584</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>15.176568</td>\n",
       "      <td>1206.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>4.558900e+04</td>\n",
       "      <td>7.780000e+02</td>\n",
       "      <td>932.000000</td>\n",
       "      <td>4.876542e+01</td>\n",
       "      <td>14.029113</td>\n",
       "      <td>2.299666</td>\n",
       "      <td>0.341936</td>\n",
       "      <td>3.555194</td>\n",
       "      <td>195.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>773.000000</td>\n",
       "      <td>890.000000</td>\n",
       "      <td>4.876542e+01</td>\n",
       "      <td>14.029113</td>\n",
       "      <td>2.251398</td>\n",
       "      <td>0.341936</td>\n",
       "      <td>3.342999</td>\n",
       "      <td>190.000000</td>\n",
       "      <td>54.925234</td>\n",
       "      <td>2206.420734</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.927380e+05</td>\n",
       "      <td>2.838000e+03</td>\n",
       "      <td>6734.000000</td>\n",
       "      <td>3.498196e+02</td>\n",
       "      <td>118.704407</td>\n",
       "      <td>7.198330</td>\n",
       "      <td>1.320681</td>\n",
       "      <td>10.691878</td>\n",
       "      <td>1323.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2838.000000</td>\n",
       "      <td>6734.000000</td>\n",
       "      <td>3.743699e+02</td>\n",
       "      <td>107.081021</td>\n",
       "      <td>6.866840</td>\n",
       "      <td>1.320681</td>\n",
       "      <td>10.600502</td>\n",
       "      <td>1323.000000</td>\n",
       "      <td>112.191489</td>\n",
       "      <td>4349.908608</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.654319e+07</td>\n",
       "      <td>1.165830e+06</td>\n",
       "      <td>549144.000000</td>\n",
       "      <td>1.145219e+06</td>\n",
       "      <td>435825.874241</td>\n",
       "      <td>76.809514</td>\n",
       "      <td>16.290540</td>\n",
       "      <td>193.072418</td>\n",
       "      <td>920838.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>664324.000000</td>\n",
       "      <td>549144.000000</td>\n",
       "      <td>1.145219e+06</td>\n",
       "      <td>435825.874241</td>\n",
       "      <td>76.809514</td>\n",
       "      <td>16.290540</td>\n",
       "      <td>193.072418</td>\n",
       "      <td>920838.000000</td>\n",
       "      <td>1861.583333</td>\n",
       "      <td>75526.083333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Choice  A_follower_count  A_following_count  A_listed_count  \\\n",
       "count  5500.000000      5.500000e+03       5.500000e+03     5500.000000   \n",
       "mean      0.509455      6.498840e+05       1.265895e+04     5952.453273   \n",
       "std       0.499956      2.028787e+06       4.900867e+04    17339.141191   \n",
       "min       0.000000      1.600000e+01       0.000000e+00        0.000000   \n",
       "25%       0.000000      2.663750e+03       3.220000e+02       85.000000   \n",
       "50%       1.000000      4.558900e+04       7.780000e+02      932.000000   \n",
       "75%       1.000000      3.927380e+05       2.838000e+03     6734.000000   \n",
       "max       1.000000      3.654319e+07       1.165830e+06   549144.000000   \n",
       "\n",
       "       A_mentions_received  A_retweets_received  A_mentions_sent  \\\n",
       "count         5.500000e+03          5500.000000      5500.000000   \n",
       "mean          2.666032e+03          1032.371839         6.011873   \n",
       "std           2.916543e+04         10954.953223         9.519797   \n",
       "min           1.005034e-01             0.100503         0.100503   \n",
       "25%           3.453649e+00             0.716816         0.359534   \n",
       "50%           4.876542e+01            14.029113         2.299666   \n",
       "75%           3.498196e+02           118.704407         7.198330   \n",
       "max           1.145219e+06        435825.874241        76.809514   \n",
       "\n",
       "       A_retweets_sent      A_posts  A_network_feature_1  ...  \\\n",
       "count      5500.000000  5500.000000          5500.000000  ...   \n",
       "mean          1.109924     9.090730          5267.768000  ...   \n",
       "std           1.910104    18.311060         28946.777345  ...   \n",
       "min           0.100503     0.100503             0.000000  ...   \n",
       "25%           0.100503     0.632440            12.000000  ...   \n",
       "50%           0.341936     3.555194           195.000000  ...   \n",
       "75%           1.320681    10.691878          1323.000000  ...   \n",
       "max          16.290540   193.072418        920838.000000  ...   \n",
       "\n",
       "       B_following_count  B_listed_count  B_mentions_received  \\\n",
       "count        5500.000000     5500.000000         5.500000e+03   \n",
       "mean        12738.259818     5903.148364         2.554598e+03   \n",
       "std         50054.520874    16298.462018         2.508873e+04   \n",
       "min             0.000000        0.000000         1.005034e-01   \n",
       "25%           322.000000       75.000000         3.260473e+00   \n",
       "50%           773.000000      890.000000         4.876542e+01   \n",
       "75%          2838.000000     6734.000000         3.743699e+02   \n",
       "max        664324.000000   549144.000000         1.145219e+06   \n",
       "\n",
       "       B_retweets_received  B_mentions_sent  B_retweets_sent      B_posts  \\\n",
       "count          5500.000000      5500.000000      5500.000000  5500.000000   \n",
       "mean            997.149954         6.099658         1.106236     9.505821   \n",
       "std            9342.006880         9.729557         1.939914    19.424680   \n",
       "min               0.100503         0.100503         0.100503     0.100503   \n",
       "25%               0.714556         0.356943         0.100503     0.822584   \n",
       "50%              14.029113         2.251398         0.341936     3.342999   \n",
       "75%             107.081021         6.866840         1.320681    10.600502   \n",
       "max          435825.874241        76.809514        16.290540   193.072418   \n",
       "\n",
       "       B_network_feature_1  B_network_feature_2  B_network_feature_3  \n",
       "count          5500.000000          5500.000000          5500.000000  \n",
       "mean           5254.933636            85.024196          3745.175480  \n",
       "std           26778.820125           106.689384          5518.399590  \n",
       "min               0.000000             0.000000             0.000000  \n",
       "25%              11.000000            15.176568          1206.500000  \n",
       "50%             190.000000            54.925234          2206.420734  \n",
       "75%            1323.000000           112.191489          4349.908608  \n",
       "max          920838.000000          1861.583333         75526.083333  \n",
       "\n",
       "[8 rows x 23 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Twitter_Train.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Any missing values in the dataset: False\n"
     ]
    }
   ],
   "source": [
    "print(\"Any missing values in the dataset:\", Twitter_Train.isnull().values.any())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0.5, 0, 'A_follower_count'),\n",
       " Text(1.5, 0, 'A_following_count'),\n",
       " Text(2.5, 0, 'A_listed_count'),\n",
       " Text(3.5, 0, 'A_mentions_received'),\n",
       " Text(4.5, 0, 'A_retweets_received'),\n",
       " Text(5.5, 0, 'A_mentions_sent'),\n",
       " Text(6.5, 0, 'A_retweets_sent'),\n",
       " Text(7.5, 0, 'A_posts'),\n",
       " Text(8.5, 0, 'A_network_feature_1'),\n",
       " Text(9.5, 0, 'A_network_feature_2'),\n",
       " Text(10.5, 0, 'A_network_feature_3')]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZoAAAFFCAYAAADVbiuZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOydebhVZdn/P99zGFVUnNEQzEhUUBREEQdUJLMSLUyRzDGzV6XJgfJNTbMwSy3TlEzRQlFAi1cNxIHBAQQVGcwpsCL9qQUZKiHD/fvjebYsNntY+5x1OHsf7s91rWuv9axn3c+99j5n3/sZ1v2VmeE4juM4TUVdczvgOI7jtGw80DiO4zhNigcax3Ecp0nxQOM4juM0KR5oHMdxnCbFA43jOI7TpHigcRzHaWFIul3SO5IWFDkvSb+U9LqkeZL2T5w7TdJrcTstC3880DiO47Q8RgPHlDj/WaBb3M4Bfg0gaRvgcuBAoC9wuaSOjXXGA43jOE4Lw8ymA0tLVBkM3GWBmcDWkjoBnwGmmNlSM1sGTKF0wEpFq8YacKqfAVf8KrP0D3cNPzUrUwBcPvbhTO317dYlM1trM86a8dV++2Rq7z+rMzVH+7atM7N106QZmdkC+Havzpnas+13ydRehw4d1FgblfyfTvvhBV8n9ERyjDKzURU0twvw98TxklhWrLxReKBxHMepMWJQqSSw5FMoMFqJ8kbhQ2eO4zibHkuAZDfxE8CbJcobhQcax3GcKqC+ri71lgETga/G1WcHAe+Z2VvAZGCQpI5xEcCgWNYofOjMcRynClCjZ3mStnQPMADYTtISwkqy1gBmdgvwMHAs8DrwIXBGPLdU0lXA7GjqSjMrtaggFR5oHMdxWhhmNrTMeQPOK3LuduD2LP3xQOM4jlMFZDQkVpV4oHEcx6kClOXYWZXRrCFU0gmSTFL3MvWulbRQ0rUl6lwh6cK4P1rSkKz9rRYkdZV0SnP74ThOdrSqr0u91RrN7fFQ4Eng5DL1vg7sb2YXNb1L6ZHUXD3CroAHGsdxaoJmCzSStgD6A2dRItBImghsDsySdJKkLpIei4ngHpO0a5l2jpL0gqT5MdFcW0l9Jd0fzw+WtEJSG0ntJC2K5btLmiTpOUkzcr2u2Fu6TtITwDXF7k3SHbHNeZK+FMuHxrIFkq5J1H8/sT9E0uhEW7+U9LSkRYle2kjgUElzJX27iA/nSJojac6bzz1V6i1yHKcKqJNSb7VGc/ZojgcmmdmrwNJk9tAkZnYcsMLMepnZvcCvCDl69gHGAL8s1oCkdoTkcieZWU/CnNQ3gOeB/WK1Q4EFwAGERHKzYvko4AIz6w1cCNycMP1pYKCZfbdI0z8grEvvGf18XNLOhMB0JNALOEDS8cV8T9AJOAT4PCHAAIwAZsT35PpCF5nZKDPrY2Z9du7dP0UzjuM0Jx5omoahwNi4PzYep6EfcHfc/x3hS7gYewCLYzADuBM4zMxWA69L2pOQofQ64DBC0JkRe1sHA+MkzQVuJXzh5xhnZmtKtDsQuCl3EJPTHQBMNbN3Y/tjYpvl+IOZrTWzl4AdU9R3HKcGkZR6qzWaZY5B0raEX/Y9JBlQD5iki+P67kooVb/UJzKDkCp7FfAooedTT+i91AH/NrNeRa79oIxPKuBXKV+SddvlnVuZ0objOE5V0lw9miGE4a8uZtbVzDoDiyndO8nxNOvmdIYRFhMU42Wgq6RPxeNTgWlxfzrwLeAZM3sX2BboDiw0s/8AiyWdCB+LBO2b/vZ4BDg/dxBTOcwCDpe0naR6Qg8u58vbkvaUVAeckML+cqBDBf44jlPl+Kqz7BkKPJBXNoF0K6mGA2dImkcIHN8sVtHM/ktIrTBO0nxgLXBLPD2LMBQ1PR7PA+YlelTDgLMkvQgsJOg3pOVHQMc46f8icETMI/Q94AngReB5M/tjrD8CeBB4HHgrhf15wGpJLxZbDOA4Tm3RkudommXozMwGFCgrOqlvZlsk9t8gDLvl17kisX96Yv8x1k38J+uvANomjs/JO7+YAoI/Sdsl/H0f2EAC1czuZt38UrJ8PDC+XFu598HMVgFHlfPDcZzaoU6111NJS8u9M8dxHKcqqJoUNJJ6ElaRJVlpZgc2hz9pkHQGGw7dPWVmBZPVOY7jFKOu9kbEUlM1gcbM5hOeL6kZzOwO4I7m9sNxnNqnFpctp6VqAo3TdNw1/NTMbH31l/mdzsYx5psbTGU1ijatq/dPetlHH2Vqb4dV75evVAH/WLVZZrbOHZRmAWl6Psr45/6KlasytZfFEtDW9fUZWKlOfI7GcRzHaVKq9+ef4zjOJoQPnTmO4zhNSkteDOBDZ47jOFVAXV1d6q0cko6R9Iqk1yWNKHD++pj9fa6kVyX9O3FuTeLcxCzuzXs0juM4VUBWT/zHFFc3AUcDS4DZkibGxLwAmNm3E/UvYP2H2leUyPPYILxH4ziOUwVkmL25L/C6mS0ys48I2fFLpdAaCtyT0W0UxAON4zhOFSCl38qwC/D3xPGSWFagTXUBdiPkWczRLoomzkypmVWWjRZoJJ0gyXJKlSXqXStpoaRrS9S5QtKFcX90Qnmyob5dKWlgY2xsbCR9S1J2Dz44jtOstKqrS70lFXTjlszVWCgUFZNTORkYn6evtauZ9SEkOb5B0u6NvrfGGqiAoYSU/icDV5So93VgezNbWaJOppjZZRurrQz5FvB74MPmdsRxnI2LmY0iqAAXYgnQOXH8CeDNInVPBtZLmWVmb8bXRZKmEuZv/tIYfzdKjyYqVvYHzmKdlkyhehOBzYFZkk6S1EXSY5Lmxdddy7RzlKQXJM2XdLuktpL6Sro/nh8saYWkNpLaSVoUyz/uFUl6Q9IPJT0f7XSP5dtLmhLLb5X0V0nblfDlq9HvFyX9LpYVvJ/8Xpmk9+PrAElTJY2X9LKkMVEbZziwM/CEpCeKtP/xL5677xxd6m1zHKcKyHCOZjbQTdJuktoQvnM3WD0maQ+gI/BMoqyjpLZxfzvC9/ZL+ddWysYaOjsemBQllZdK2r9QJTM7jrjiwczuBX5FEEjbhyB9XFRKQFI7gkrmSWbWk9Bb+wbwPOtWVBwKLCDIKh9I0KQpxD/NbH/g1wTFTYDLgcdj+QNA0aAnaW/gUuBIM9uXdYk3U99Pgv0IvZe9gE8C/aOkwpsEnZsjCl1kZqPMrI+Z9TnltNNTNOM4TnPSqr4+9VaKKBV/PjAZ+DNwn5ktjFMExyWqDgXG5qka7wnMUdDRegIYmVyt1uB7a6yBlAwFboj7Y+Px8ymu6wd8Me7/Dvhpibp7AItjMAO4EzjPzG5QWEu+J2E1xnXAYQTZ5hlFbN0fX59LtH8IUf3SzCZJWlbClyMJ457/jPWXNuB+cjxrZksAJM0FulJaVdRxnBokywc2zexh4OG8ssvyjq8ocN3TQM/sPAk0eaCRtC3hi7eHJCN8wZuki/MiaRpK1S/1Mc0APgusAh4l9HzqWddbySc3P7SGde9RJX8GorSvOXJ1VhN7lwr94jYFfMn3x3EcpybYGENnQwjDRV3MrKuZdQYWE3oI5XiadXM6wyj9S/5loKukT8XjU4FpcX86YfjpGTN7F9gW6E6QaE7Lk8CXASQNIoxtFuMx4MsxyCJpmzL38wbQO+4PBlqn8Gc52SSNdRynCshwjqbq2BiBZihhTiPJBMLSuXIMB86QNI8QOPJFxj7GzP4LnAGMkzQfWAvcEk/PAnYkBByAecC8CntUPwQGSXqe0Dt6i/BlX8iXhcDVwLQ41nldmfv5DXC4pGcJc0cfpPBnFPCnYosBHMepLVpyoFHlo1ebJnElxhozWy2pH/DrrNM0NBV/W/peZh+y69E0nBXVrkeT4WNZW27WLjNbAPVVrkezS8cOjXbw0rEPp/4/vfrkY2sq2lTvf2X1sStwn6Q64CPga83sj+M4LYha7KmkpVkCjaSehFVXSVaa2YHN4U8azOw11k88l1vo8FiB6keZ2b82imOO47QIWnCcaZ5AY2bzgZoYdipFDCZVfx+Xj324fKWUZD3UNewXd2Zqb58uO2dqL0t+cOJnM7X3t1XZDk91zHC466d/mJKZLYCrju5dvlIFtN5sy0ztZUF9ivT/tYoPnTmO41QBWckEVCMtN4Q6juM4VYH3aBzHcaqAltyj8UDjOI5TBbTkOZqWe2eO4zhOVeA9GsdxnCqgLuOHUqsJDzSO4zhVgD+w6TiO4zQp9S040NTUHI2kEyRZTvWySJ2ukhbE/T6SSomldZWUJrln/nXrKWJubCQdL2mv5mrfcZzsqZNSb7VGTQUaQiboJykhB53EzOaY2fASVbqSLot0tXE8QXHTcZwWQkvO3lwzgUbSFgT96rNIGWgkDZD0YNw/XNLcuL0gqQMwEjg0ln1bUr2kayXNljRP0tfjtZL0K0kvSXoI2KFMuwdIelrSi5KeldRBUjtJd0iaH9s/ItY9XdKvEtc+KGlA3H9f0tXRzkxJO0o6GDgOuDb6vXsRH86RNEfSnFeeLJSOzXGcasIDTXVwPDApSjUvlbR/hddfSJB27gUcCqwARgAzzKyXmV1PCGLvmdkBwAHA1yTtRpBw3oMgcfo14OBijUhqA9wLfNPM9gUGxrbOAzCznoSe2Z2SyiWX2hyYGe1MB74WpVYnAhdFv/9S6EIzG2Vmfcyszx6HHFX2zXEcp3mpr1PqrdaopUAzFBgb98fG40p4CrhO0nBgazNbXaDOIOCrkuYSxNK2BboBhwH3mNkaM3sTeLxEO3sAb5nZbAAz+09s6xBixmozexn4K/DpMj5/BDwY958jDPU5jtMCybJHI+kYSa9Iel3SiALnT5f0bmKU5+zEudMkvRa3TLLo1sSqs5iO/0ighyQD6gGTdHFalUwzGxmHvY4FZkoaWKgp4AIzm5zX/rFAWlEiFalb7K9jNesH/GQvZ1Xi/tZQI5+X4zjNh6R64CbgaGAJMFvSRDN7Ka/qvWZ2ft612wCXA30I32PPxWuXNcanWunRDAHuMrMuZtbVzDoDiwm9hFRI2t3M5pvZNcAcoDtBirlDotpk4BuSWsdrPi1pc8Kw1clxDqcTcESJpl4GdpZ0QLTRQVKraGNYzi5BSO0V4A2gl6Q6SZ2BviluJ99vx3FqnNb19am3MvQFXjezRWb2EWEEaHBKNz4DTDGzpTG4TAGOafBNRWol0AwFHsgrm0BlK8a+JWmBpBcJcyZ/AuYBq+Nk+7eB24CXgOcVlkjfSuhFPAC8BswHfg1MK9ZI/GBPAm6MbU0h9FJuBuolzSfM4ZxuZisJQ3qLo+2fAc+nuJexwEVxUUHBxQCO49QWlSxvTi72ids5CVO7AH9PHC+JZfl8KS56Gh9/5FZybUXUxFCMmQ0oUFbw+RgzewPoEfenAlPj/gVFzOfPlH8/bvmcX6CsIHF+5qACp04vUNeIPZ0C57ZI7I8Hxsf9p/DlzY6zyWJmo4BRRU4XGqbPH87/P8K880pJ5wJ3EqYn0lxbMbXSo3Ecx2nRZLgYYAnQOXH8CeDNZAUz+1ccUQH4DdA77bUNoSZ6NIWQ1JO4iivBSjM7cCP68ACwW17xJfmLCRzHccqR4eMxs4Fu8dGMfxCeO1xvmkFSJzN7Kx4eB/w57k8GfiypYzweBHyvsQ7VbKAxs/lAr2b24YTmbD8tfbt1ycxWm9bZ/sns02XnTO3N+2ujf3w1Ge0KrqhvOB033yxTe1n69+d/vJ2ZLQC1ap2pvWqkVflJ/lSY2WpJ5xOCRj1wu5ktlHQlMMfMJgLDJR1HWPW6lDisb2ZLJV1FCFYAV5rZ0sb6VLOBxnEcxymMmT0MPJxXdlli/3sU6amY2e3A7Vn644HGcRynCqjBB/5T44HGcRynCmjJUs4eaBzHcaqAWkz/nxYPNI7jOFVALWZlTosHGsdxnCrAA43jOI7TpNRi+v+0pJp9UgoJ5aZC0taS/idxvLOk8Rvbj6yQdJsykGHOF0xzHKe2ceGzCiWUM2Zr4ONAY2ZvmtmQjelAzL6cCWZ2doF03Y7jOC2WsoFGKSWUFWSTp0m6T9KrkkZKGqYgZTw/l2VY0vaSJijIJc+W1D+WXyHpdklTJS1SECiDILe8exTnuVZS15hZGZWWR75f0qQo3vPTWF4vaXTM4jw/Zmwudj9TJf1Y0jTgmyX83iLhwzxJX4rlgyQ9I+l5SePi+5iz20fSN3J+JXy+Me5/Jb5vcyXdqqAvgaQz4ns7LX4mjuO0EOpQ6q3WSNOjqURCeV/gmwTJ41OBT5tZX0L6/Vz25F8A10e55C/Fczm6E/QQ+gKXK+jCjAD+EmWLL8prr5Q8ci9Cuv6ewEkxDXYvYBcz6xGvuaPMvW9tZoeb2c9L+P0DgvxzTzPbB3hc0nbA/wIDzWx/gv7Nd/Jsjwe+mDg+CbhX0p5xv3+UnV4DDFPQwfkhIcAcTZnszUqkEX/yoT+WuU3HcZqbVq3qU2+1RpohoaHADXE/J6FcTDNldi5Rm6S/AI/E8vmsEwsbCOyVGGfcUlJOxOuhmFF0paR3gB3L+HYIcCMEeWRJSXnkx8zsvejLS0AXYCHwydhzeCjhXzHuTewX83sgiZ6emS2T9HlCIHgq1m8DPJM0bGbvxp7bQQStmz0I2jTnETKpzo7XtgfeAQ4EpprZu/Ge7qWEFHQyjfivpzzd6DTfjuM0LS14LUDpQKPKJZRXJvbXJo7XJtqqA/qZ2Yq8tvKvTyNdXOqj2cBWDAL7EnpN5wFfBs4sYeODxH4xvwtJN4ugUje0jP/3Rh9eBh4wM4v27oy5iJLtHF+gHcdxnKqn3NBZoyWUC/AICRExSeUyMJeSLS4mj1yQOKRVZ2YTCENepYYB8ynmd355R2Am0F/Sp2LZZtG/fO4nDE0OZV3v6TFgiKQd4rXbSOoCzAIGSNo2DimeWIHvjuNUOYVWlxXbao1ygSYLCeV8hgN94sT5S8C5pSqb2b8IQ1ALJF2bd7qYPHIxdgGmSpoLjKYynYVifv8I6Kh1MtFHxOGt04F7JM0jBJ4NloZHTe6XgC5m9mwse4kwv/NIvHYKkNOOuIIwBPco6SSfHcepEepVl3qrNVR4BMxpSWQ5RzOk335ZmQLgqnF/ytReNevR/N93T8vU3n+zW3UPZKtHM/TmsZnZAphw9uBM7X3Upn2m9jp06NDobsajC15P/X86sMenaqpb45kBHMdxqoBNdjFAIVQFEspZIukmNnwm5RdmVm7ps+M4TmbUuUzAOqpBQjlLzOy85vbBcRzHZQKcmmatz8O1SDJPwrgmW3OZYmszNdd2dak1Qw2h2MLY9GQZaCQdQ3jIvB64zcxG5p3/DnA2sBp4FzjTzP4az60hPPsI8DczO66x/nigcRzHaUHElFU3ETKILCE8/D0xL8fiC0AfM/tQ0jeAnxIykgCsiFlJMqPlDgo6juPUEBk+R9MXeN3MFpnZR4SMLust2zOzJ8zsw3g4E/hE5jeUwAON4zhOFdCqvi71lsxlGLdzEqZ2Af6eOF4Sy4pxFpB8zqBdtDkzZiRp/L1lYcRxHMdpHJXM0SRzGRagkKGCE7WSvgL0AQ5PFO9qZm9K+iQhSfB8M/tLaucK4D0ax3GclsUSoHPi+BPABk8ySxoIXAocl8yoYmZvxtdFwFSg0U9pe6BxHMepAurqlHorw2ygm6TdJLUhZJefmKwgaT/gVkKQeSdR3lFS27i/HeEZw0YLNfrQmeM4ThWQ1fJmM1st6XxgMmF58+1mtlDSlcAcM5sIXAtsAYyLiwtyy5j3BG6VtJbQERmZhSJwQzIDnEDIOrynmb3cWAckfd/MftxYOyXsnw48kusONjfxw55uZo820s4A4EIz+3wmjjmO06xkmZXZzB4GHs4ruyyxP7DIdU8TxCIzpSFDZ0OBJykh65wkJ0Ncgu83wIdKOB3YuTEGpOyyF5rZZY0NMo7jtDzqpNRbrVFRoFHQve9PWA5XNNBIGiDpCUl3E58wlfQVSc9KmivpVkn1kkYC7WPZGEkXSxoe618v6fG4f5Sk38f9QZKekfS8pHHRJyT1ljRN0nOSJkvqJGkIYUXFmNhGe0kjJb0U0/3/rMQ9jJZ0naQngGskbS7pdkmzJb0gaXCsVy/pZ5LmR5sXFPMnYXeIpM9Kui/vPfu/Mvd4jKSXJT3J+jLQhfxfJ+X8sEs5O061U19Xl3qrNSr1+Hhgkpm9CiyVVEo4rC9wqZntJWlPwlOn/eMTp2uAYWY2gvgUqpkNIwiZHRqv7wNsoSDydQgwI05O/S8w0Mz2B+YA34l1bgSGmFlv4HbgajMbH+sMi+22B04A9jazfQhaMqX4dGzru4TVGY+b2QEEWeprJW0OnAPsBuwXbY4p5k+e7SnAQdEG8f25t8Q9tgN+A3whvkc7lXLczEaZWR8z63PIsdmmWHccJ3syfGCz6qh0SGgocEPcHxuPiwlwPWtmi+P+UUBvQioECF/47xS45jmgt6QOBCnm5wkB51CC8NhBwF4EITSANgQhsD2AHsCUWF4PvFXA/n+A/wK3SXoIeLDM/Y4zs1wGqEHAcZIujMftCIqeA4FbzIKYh5ktldSjnD9xwm4S8AVJ44HPARcT1rMXusfuwGIzew0g9vCSD2k5jlPDuEwAIGlb4EighyQjfHmapIutsHraB8nLgTvNrKSipZmtkvQGcAbwNDCP0HvYHfhzfJ1iZkPzfOsJLDSzfmXsr5bUlxD4TiZIMB9Z4pL8e/iSma0nFa0QDfLvX2n8IaiCngcsBWab2fJor9A99irQjuM4TtVTydDZEOAuM+tiZl3NrDOwmDCsVY7HgCGSdgCQtI2kLvHcqjjUlGM6cGF8nUGQTJ4bg9lMoL+kT0U7m0n6NPAKsL2kfrG8taS9o73lxNSqca5jq7gi41tUJncwGbggBoLcOnSAR4BzFRcMSNqmjD9JpgL7A18jBB1K3OPLwG6Sdo/1huI4TouhVV196q3WqCTQDAUeyCubAJxS7sK4Dvt/gUckzSPMT3SKp0cB8ySNiccz4rlnzOxtwlDXjGjnXcIqsnuinZlA95g4bghh0v5FYC5wcLQ3GrhF0lxCwHkwXjsN+HYF938V0Dr6uiAeA9wG/C2WvwicUsaf5PuyhjB899n4Wuoe/0sYKnsoLgb4awW+O45T5bTkORoVHvVyWhI3PfJUZh/ylw8utf6jcq4a96fylSpg3l+r4nGpgvzfd0/L1N6q+tblK1VA6zWrMrM19OaxmdkCmHDWFzK1p/psn1Vv03G7Rn/7L3p3Wer/009u37Gmok3trZNzHMdxaopGhfU4Cf+7vOKVZnZgY+xuTCRdCpyYVzzOzPKXIzuO4zQZrepb7u/+RgUaM5tPZRPqVUcMKB5UHMdpVmpx7iUtPkezCbB8+fLMPuRlq7LVbm/fpk2m9tqFx5mqki/8/M5M7R3bu9BCxoazZk12n+35R/XNzBbACrJdadWeNeUrVUCHDh0aHSX+sSz9/+kuHRvf3sbEszc7juNUASnS/9csHmgcx3GqgFpMlpkWDzSO4zhVQEueo/FA4ziOUwV4oHEcx3GalHqfo3Ecx3GaksoyM7RrMj+agpb7hJDjOM4mShRJfEXS65JGFDjfVtK98fwsSV0T574Xy1+R9Jks/GnWQCPpBEkmqXsztL21pP9JHO8cdWFqAklNLYHtOE4NIqkeuImQrHcvYKikvfKqnQUsM7NPAdcD18Rr9yJIqOwNHAPcHO01iubu0QwFnqSELHQTsjXwcaAxszfNbEgz+NFQPNA4jlOIvsDrZrYoZpIfC+TL7A4Gck8QjweOihIog4GxZrYyCle+Hu01imYLNFEbpj8hshYNNJIGSJom6T5Jr0oaKWmYpGclzc/ps0jaXtIESbPj1j+WXyHpdklTJS2SNDyaHgnsLmmupGsldY3p/5HUTtId0f4Lko6I5adLul/SJEmvSfppLK+XNFrSgnhNUfkBScMlvSRpnqSxsWzz6OPs2N7gMu2NBNpH38cUaeccSXMkzbnjjjvSfzCO41Q9yf/vuCXVdncB/p44XhLLKFQnqgO/B2yb8tqKac7FAMcDk8zsVUlLJe1vZsVkofcF9iQoUS4CbjOzvpK+CVxAEDH7BXC9mT0paVeCUNme8fruBKXODsArkn4NjAB6mFkvgOQYJUH1EjPrGYf1HlEQH4OQ220/gtT0K5JuBHYAdjGzHtHW1iXuewSwm5mtTNS7FHjczM6MZc9KerRYe2Y2QtL5Od8LYWajCFo/maagcRyn+Un+fxeg0PK1QirAheqkubZimnPobCihS0d8LaUYOdvM3jKzlcBfCKqWAPOBrnF/IPArBYGzicCWkjrEcw/FruA/gXeAHcv4dggxK7WZvUwQGcsFmsfM7L0oRPYS0IUQ/D4p6UZJxwD/KWF7HjBG0leAXGKuQcCI6PtUwpKSXUu05ziOU4wlQOfE8SeAfKGmj+soqANvRfghn+baimmWQCNpW+BI4DZJbwAXASep+BNLKxP7axPHa1nXK6sD+plZr7jtYmbLC1y/hvI9uVIL2jewZWbLCL2uqYTe0G0lrv8cYaKuN/Bc/JAFfCnh+65m9ucG+u44zqbNbKCbpN0ktSFMTUzMqzMRyCnxDSGMqFgsPzmuStsN6AY821iHmqtHMwS4y8y6mFlXM+sMLCb0JBrKI8D5uQNJ5eQLlhOG0goxHRgW7Xya0Lt4pZghSdsBdWY2AfgBUFCGUlId0NnMngAuJixI2IIwzHdBLtBK2q+M7wCrJGUrseg4Ts0T51zOJ3yv/Bm4z8wWSrpS0nGx2m+BbSW9DnyHMKSPmS0E7iOMnkwCzouS842iuX4dDyVMxieZAJwCzGigzeHATZLmEe5rOnBuscpm9i9JT8UFAH8i9DJy3AzcImk+YXjr9DinUszcLsAdMZAAfK9IvXrg95K2IvRirjezf0u6CrgBmBeDzRvA58vc76hY/3kzG1amruM4VU7b1SvLV/qYYr+RA2b2MPBwXtllif3/sqHgY+5c5hpdrkezCeB6NNWB69E0nE1Bj2blv95J/X/adtsdaipfjY/3O47jVAOW7Y+4aqJqAo2knsSVXglWmtmBzeFPY5F0E+E5oSS/MDN/qMVxnA2wtTTZ7fQAACAASURBVC13dKlqAo2ZzSc8M9IiMLPzmtuHHP/JcDRph1XvZ2cM+NuqbJMDdtx8s8xsZZ1NN+uhroefW5ipvW9+fkBmtt5fm+17t83qbP/uVJf1OqjScyap8B6N4ziO05TY2pYbaJo715njOI7TwvEejeM4ThVga7JdCVdNeKBxHMepAmxN9S7Nbyw+dOY4juM0Kd6jcRzHqQZa8MPzHmgcx3GqgJacpcUDjeM4TjXQgp+jqZo5GkknSLIoNJaFvSaVOo7qlzs3ZRtl2v+WpOyeTnQcp1mxNWtSb7VG1QQaQkbnJykh65xEUrkse00aaIDTgWYLNARVUQ80jtNSMEu/1RhVEWgkbUHIC3YWJQKNpAGSnpB0N0FdE0lfkfSspLmSbpVUL2kk0D6WjZF0saThsf71kh6P+0dJ+n3cHyTpGUnPSxoXfUJSb0nTJD0nabKkTpKGAH0ISplzJbWXNFLSS5LmSfpZiXs4UdICSS9Kmh7L6iVdK2l2vP7rifudKmm8pJfjvSjey87AE5KeKNLOx5rivx/t6dUcp9oxs9RbrVEtczTHA5PM7FVJSyXtb2bPF6nbF+hhZosl7QmcBPQ3s1WSbgaGmdkISeebWS8ASQcB3wV+SQgQbaNo2CHAjChc9r/AQDP7QNIlwHck/QS4ERhsZu9KOgm42szOlHQ+cKGZzZG0DXAC0N3MTNLWJe71MuAzZvaPRL2zgPfM7ABJbYGnJOXkqvcD9ibIqT4V7/WXkr4DHBHlqTcgqSn+j2XZyQQ4jtM02NraGxJLS7UEmqEE4S+AsfG4WKB51swWx/2jCJLIs6MoWXvgnQLXPAf0ltSBII38PCHgHEoQTDsI2IvwBQ/QBngG2APoAUyJ5fXAWwXs/wf4L0Ga+iHgwRL3+hQwWtJ9wP2xbBCwT+wpQdDv7gZ8FO93CYCkuUBXwhCj4zhOTdDsgUbStsCRQA9JRvgyN0kXW+E+4gfJy4E7zayYoiUAsbfzBnAG8DQwDzgC2J0gdbo7MMXMhub51hNYaGb9ythfLakvIfCdTJBRPbJI3XMlHQh8DpgbJacFXGBmk/PaH0AIjDnWUAWfmeM42eOZAZqWIcBdZtbFzLqaWWdgMWFYqxyPAUMk7QAgaRtJXeK5VXF4LMd04ML4OoMg8zw3BrOZQH9Jn4p2NpP0aeAVYHtJ/WJ5a0m5XO/LibnB43zOVlE+9VuUkDuQtLuZzYqyqv8EOhO0vb+R81fSpyVtXubeP27fcZwWwFpLvzWC+D05RdJr8bVjgTq94pz1wjhvfFLi3GhJi+P8dO7HckmqIdAMBR7IK5sAnFLuQjN7iTC38oikecAUoFM8PQqYJ2lMPJ4Rzz1jZm8ThrpmRDvvElaR3RPtzCTMt3xECITXSHoRmAscHO2NBm6Jw1kdgAfjtdOAb5dw+1pJ8yUtIAS9F4HbgJeA52P5rZTvuYwC/lRsMYDjOLWFrV2TemskI4DHzKwb4cf6iAJ1PgS+amZ7A8cAN+TNPV9kZr3iNrdcg6rFFQxOZWS5GGDbj/6TlSkA/samI3x222PPZGqvmoXPDurWNTNbANus/jBTe1kLn7XpuF2j/1iWzZ2Z+v+0Y6+DGtyepFeAAWb2lqROwFQz26PMNS8CQ8zsNUmjgQfNbHzaNquhR+M4juNU8BxN8vGFuJ1TQUs7mtlboUl7C9ihVOU4/9wG+Eui+Oo4pHZ9XClbkqqcWI6T8L/LK15pZgc2hz8NQdKlwIl5xePM7Orm8MdxnOrGKkhBk3x8oRCSHgV2KnDq0kp8ij2e3wGn2ToHvwf8P0LwGQVcAlxZyk5VBhozm0+JCfVaIAaUqggq7du2Ll8pJf9YlW0ygo6bZTt01s4yXLmT8WMNa9Zkm8sqy6EugF88ODUzWwd/54zMbAGZ5wFb2ap9pvbaZGDDVmf3t2tmA4udk/S2pE6JobNCj4QgaUvgIeB/zWxmwnbuEY+Vku4gLLIqiQ+dOY7jbFpMBE6L+6cBf8yvIKkNYZHWXWY2Lu9cp/gqwsP2C8o16IHGcRynCrC1a1NvjWQkcLSk14Cj4zGS+ki6Ldb5MnAYcHqBZcxjJM0npAHbDvhRuQarcujMcRxnk2MjrQA2s38RHi7PL58DnB33fw/8vsj1BR9GL4UHGsdxnCqgJec686Ezx3Ecp0nxHo3jOE4VUIuCZmnxQOM4jlMF+NCZsx7KWHa6TFvNKhntOM5GwhU2nTwqkp1uJKfTvJLRjuM4jcIDTYWoMtnp6ZIeUJB4vkVSXTw3NJfBWdI1saw+pt9eEM99W42QjHYcp7bYiM/RbHR8jqZyKpWd3gv4KzAJ+KKkp4FrCMqgywgSB8cDfwd2MbMeAJK2NrN/q4GS0THJ3jkA1934K04786ws7t1xnCYiyxQ01YYHmsqpVHZ6EYCkewhibqsIabnfjeVjCE/gXgV8UtKNhPxCjxSwl1oyOpl0b+mH/629QV3H2dTIOJ9bNeFDZxWQkJ2+LUpDXwScFHP+FCL/C94Iss0bVjRbBuwLTAXOI4ih5ddZTeglTSD2rCq+CcdxnI2MB5rKqFR2uq+k3eLczEmEBQSzgMMlbSepntAjmiZpO6DOzCYAPwD2jzYaJBntOE5tYWapt1rDh84qYygxAV2CnOz0jAL1n4n1exJkmx8ws7WSvgc8QejdPGxmf5S0L3BHbsEAQfMB1klGrwA+C/xRUrt4bSnJaMdxagh/YNMBwMwGFCj7ZYlLPjSzkwpcczdwd17Zi6zrxSTLJxCCWY6+af11HKeGqMGeSlo80DiO41QBlShs1hoeaBpJGdnpqRvfI8dxahLv0TjFaAmy047jND+2xp+jcWqYmyYVWqfQMM4dVGyBXcP46R+mZGrvz/94O1N7WXLP/2Sbsej9tcVW1TeMg79zRma2TrrujsxsATx0xjGZ2mu9NuPeQ4cOjTZRi6vJ0uLLmx3HcZwmxQON4zhONbDW0m+NQNI2kqZIei2+dixSb03MsThX0sRE+W6SZsXr75XUplybHmgcx3GqAFuzKvXWSEYAj5lZN+CxeFyIFWbWK27HJcqvAa6P1y8jJBguiQcax3GcKmAjZgYYDNwZ9+8kpLNKRUy3dSQwvpLrPdA4juPUGJLOkTQnsZ1TweU7mtlbAPF1hyL12kXbM2OGeYBtgX/HvIsAS4BdyjXoq84cx3GqgQrmXpLZ2Qsh6VFgpwKnLq3Ao13N7E1JnwQelzSfkEF+A3fKGfJA4ziOUwXY2uxynZnZwGLnJL0tqZOZvSWpE/BOERtvxtdFkqYC+xHSYW0tqVXs1XwCeLOcPxUNnUk6QZJJ6l7JdSnsni4pc7niqFg5JGXdQyUtzClZNqCt71fuYcVtdJf0jKSVki5s6vYcx9l4bESFzYnAaXH/NOCP+RUkdZTUNu5vR1AVfsnCBNEThEz2Ra/Pp9I5mqGEVPfZPnkGpwOZBhpJlfbWhgE/iyssVjSgyYoDTQN8XAoMB1zC2XFaGrY2/dY4RgJHS3oNODoeI6mPpJwO1p7AHEkvEgLLSDN7KZ67BPiOpNcJcza/Lddg6kATtVD6E5ayFQ00kgZImippvKSXJY3JCYNJ6i1pmqTnJE2W1Cn2OPoAY2Jv4nBJ98f6gyWtkNRGUjtJObXKXnGCap6kB3LrwGO7P5Y0Dfhmnl9XxR7OBvcs6Wzgy8BlUfESSRdJmh3b+GGi7h+i/wtzE3CSRgLto/9jJHWVtCBxzYWSrijko6TtJU2Ibc2W1L/Ye2tm75jZbIJKZ0mSk4VzHnmoXHXHcZqZjdWjMbN/mdlRZtYtvi6N5XPM7Oy4/7SZ9TSzfePrbxPXLzKzvmb2KTM70cxWlmuzkh7N8cAkM3sVWCppg5T2CfYjCHPtBXwS6C+pNXAjMMTMegO3A1eb2XhgDjDMzHoBT8XrAQ4FFgAHAAcSRMMA7gIuMbN9gPnA5Ym2tzazw83s57kCST8lrKw4wwqkSDWz2wjdyYvMbJikQUA3Qkr+XkBvSYfF6mdG//sAwyVta2YjWLfmfFjJd3FDH39BWJN+APAlCihrNgQzG2VmfcysT59Bn8vCpOM4Tcnatem3GqOSoZuhwA1xf2w8fr5I3WfNbAmApLlAV+DfQA9gSuzg1ANv5V9oZqslvS5pT8IX/XXAYbH+DElbEb6op8VL7gTGJUzcm2fyB8AsM6tk+d+guL0Qj7cgBJ7phOByQizvHMv/VYHtfB8HAntpnRr0lpI6mNnyCm06jlPDtORcZ6kCjaRtCQ/p9JBkhC99k3SxFX53kl2pNbEdAQvNrF+KJmcQ1CRXAY8SVCbrgTQT4B/kHc8m9Ei2yXURUyDgJ2Z263qF0gBCYOhnZh/GlRjtCly/mvV7i/l1kj7WRXsNmRdyHMepetIOnQ0B7jKzLmbW1cw6A4uBSlL5vgJsL6kfgKTWkvaO55YDyfSn0wlDb8+Y2buECafuhED1HrBM0qGx7qnANIoziTDZ9ZCktClWJwNnxnkpJO0iaQdgK2BZDDLdgYMS16yKw4MAbwM7SNo2rtz4fIm2HgHOzx1IcskBx9kEsTWrU2+1Rtqhs6HElQkJJgCnEHofZTGzj+LE/y/j8FcrwlDcQkKP5RZJK4B+hLmYHQkBB2Ae8E6i93RarL8ZsAgomd/czMbFIDNR0rHleg9m9kgcunsmDmm9D3yFELTOlTSPEDhnJi4bBcyT9Hyc57ky3sdi4OUSzQ0Hboo2W8V7PrdQRUk7EeaztgTWSvoWsJeZFXqIynGcWqIG517SopY8LugErrp/SmYfsuvRNJxq16OpU3b2ql2Ppr795pnaa7dDp0a/ea/dfHXq/9Nu/3Npth9+E+OZARzHcaqAAgtiWwwNDjSSegK/yyteaWYHNs6lpkXSA8BuecWXmNnk5vCnEJLOIO85IOApMzuvOfxxHKfpsTXZpaCpNhocaMxsPuEZk5rCzE4oX6t5MbM7gGzHHhzHqW5a8DSGD51tAny7V+fMbH1Ul+3Q8FVH987Unlq1Ll8pLRkPZbxHfab2tln9fqb2srzfrOdUPnfHpEztnXjwfuUrVcB5gzo12kZLni/3QOM4jlMN+ByN4ziO05T4HI3jOI7TpPjQmeM4jtO0eKBxHMdxmpJaTC2TFg80juM41YD3aBzHcZympCVnBqhIylnSCZIsZi7ODEmnS8pUyjnaHR0Teaape2hUzZwrqX0D2qpYyrkBbQyLip/zJD0tad+mbtNxHKexVBRoCFmcn6SElHMDOR3INNBIqrS3Ngz4WVTJbIg2TMWBpgE+LgYOj8qiVxEyRjuO0xJYa+m3RiBpG0lTJL0WXzsWqHNE/NGd2/4r6fh4brSkxYlzZTPEpA40UZulP3AWJQKNpAGSpkoaL+llSWMUc+1L6i1pmqTnJE2W1Cn2OPoAY6LTh0u6P9YfLGmFpDaS2klaFMt7SZoZf9k/kHujYrs/ljSNvFxhkq6Kb9AG9yzpbODLwGWSxsSyiyTNjm38MFH3D9H/hZLOiWUjgfbR/zGSukpakLjmQklXFPJR0vaSJsS2ZkvqX+y9jTrey+LhTOATJT6HcyTNkTTn9nvuK1bNcZwqwdauSb01khHAY2bWDXgsHq/vi9kT8Ud3L4Lo5YcE7awcF+XOm9nccg1W8ov6eGCSmb0qaamk/c2smJTzfsDewJvAU0B/SbOAG4HBZvaupJOAq83sTEnnAxea2Zz4K390tHMosAA4IPo6K5bfBVxgZtOi7svlBKE0CDLPh0OIvPH1pwTRsjMKKYKa2W2SDgEeNLPxkgYRJJr7EtQ2J0o6zMymA2ea2dI4vDZb0gQzGyHp/PihIKlrmfcy6ePdwPVm9qSkXQmia3uWuR5CwP9TsZNmNorY43l/0cstd5bRcVoIGQSQtAwGBsT9O4GpwCUl6g8B/mRmHza0wUoCzVCCUBnA2HhcLNA8a2ZLACTNBboC/wZ6AFNiB6ceeCv/QjNbLel1BeGxvsB1wGGx/gwF0bStzSynqnknMC5h4t48kz8AZpnZOelvlUFxeyEeb0EIPNOB4ZJyiTk7x/J/VWA738eBwF5apwWypaQOZra82MWSjiAEmmzFYRzH2RTY0czeAjCztxTUg0txMuF7OMnVki4j9ojMbGUpA6kCjaRtCd2nHpKM8KVvki4u1EMAko2uie2IIMXcL0WTM4DPAquARwk9nHrgwhTXfpB3PBvoLWkbM1ua4nqirz8xs1vXK5QGEAJDvyjnPBVoV+D61aw/LJlfJ+ljXbSXal5I0j7AbcBnzazSAOc4TpVSSQqaOGyf/PE8Ko5i5M4/CuxU4NJLK/FJUiegJ2GkJcf3gP8HtCGMmlwCXFnKTto5miHAXWbWxcy6mllnwsR0Jb+oXwG2l9Qv3kBrSXvHc8uBDom60wlDYc+Y2bvAtkB3QqB6D1gm6dBY91RgGsWZRJChfkhBzjkNk4Ez47wUknaJUX8rYFkMMt2BgxLXrJKUSx38NrCDpG0ltQU+X6KtR4DzcwelJtbi0Nr9wKlm9mrKe3EcpxZYuzb1ZmajzKxPYltvYZCZDTSzHgW2PwJvxwCSCyTvlPDqy8ADZrYqYfstC6wkyJn0LXdraQPNUOCBvLIJwCkpr8fMPiIErGskvQjMBQ6Op0cDt2jd0uJZwI6EgAMwD5iX6D2dBlwraR5BE6dkNDWzccBvCHMtZZcum9kjwN3AM5LmA+MJgXAS0Cq2exVhQj7HKGCepDHxQ7ky3seDwMslmhsO9ImLDl4Czi1R9zJC0L05vldzyt2L4zi1gZml3hrJRMJ3KPH1jyXqDgXuSRYkgpQIc/cLCly3HmrJidycQJaLAT7aqWtWpgDYbHm2o39VrUfTdstM7W31UfXq0az+IFvfql+Ppn+jhZrmjjgz9f9pr5G3N7i9OBVyH7Ar8DfgxLjAqQ9wrpmdHet1JSzm6myJp0klPQ5sT5himBuvKfmBe2YAx3GcKmBjrTqLc7tHFSifA5ydOH4D2KVAvSMrbbPBgUZST+B3ecUrzezAhtrcGEh6ANgtr/gSM5tcqH5zIOkM8p4DAp4ys/Oawx/HcZoeW9tyU9A0ONCY2XzC/EhNYWYnlK/VvJjZHYRJNsdxNhVa8DSGD51tAtj2G/R+G8yKlavKV6qA1ptlO2+RJW1Xl3w0oGLak+3QiOoqzSBVmpWtKk7xV5TWjUyTkk/Wcyrjnn6hfKUKOG9Q0YQeDh5oHMdxqoKNmBlgo+OBxnEcpwqo5IHNWsMDjeM4TjXgejSO4ziO0zC8R+M4jlMF2JqW26PxQOM4jlMFtGQpZw80juM41UALfo6mojkaSSdIspi5ODMknS4pUynnaHe0goJnmrqHRtXMuWkSbxa4vmIp5wa0MTgm35wb1TNdj8ZxWgi2Zk3qrdaodDHAUOBJSkg5N5DTgUwDTVTqrIRhwM+iNGkqbZg8Kg40DfDxMWDfqOR5JkGXxnGcloBZ+q3GSB1oojZLf4KyY9FAI2mApKmSxkt6WdKYmE4aSb0lTZP0nKTJkjrFHkcfYEz8pX64pPtj/cGSVkhqI6mdpEWxvJekmfHX/QOSOsbyqZJ+LGkaebnCJF0Vezgb3LOkswm6C5dJGhPLLpI0O7bxw0TdP0T/F0bxISSNBNpH/8dI6ippQeKaCyVdUchHSdtLmhDbmi2p6CPGZvZ+Qiphc6DoX5ykc2KvZ84dd3g2G8epdmzt2tRbrVHJL+rjgUlm9qqkpZL2N7NiUs77AXsDbxLSTPeXNAu4ERhsZu9KOgm42szOlHQ+cKGZzYm/8kdHO4cStA4OiL7OiuV3AReY2TRJVwKXE4TSIMg8Hw5h6Cy+/pQgWnZGIUVQM7stDkM9aGbjJQ0iSDT3JaTCnijpMDObDpwZU2q3B2ZLmmBmIySdH3saufTapUj6eDdwvZk9GYXNJgN7FrtQQUb6J8AOwOeK1YtCSKMAli9fXns/gRxnU8MXAwBh2OyGuD82HhcLNM+a2RIASXOBrsC/gR7AlNjBqQfeyr/QzFZLel3SnoQv+uuAw2L9GZK2InxR51Q17wTGJUzcm2fyB8AsMzuH9AyKWy4h0haEwDMdGB6/7AE6x/JKRVWSPg4E9orvCcCWkjqY2fJCF5rZA8ADkg4jiK8NrLBtx3GcjUqqQBOFco4Eekgywpe+Sbq4UA8BSGYjXBPbEUGKuV+KJmcAnwVWAY8Sejj1wIUprv0g73g20FvSNma2NMX1RF9/Yma3rlcoDSB8sfeLcs5TgXYFrl/N+sOS+XWSPtZFexXNC5nZdEm7S9rOzP5ZybWO41Qfa9esbm4Xmoy0czRDgLvMrIuZdTWzzsBioJJVT68A20vqByCptaS947nlBKnkHNMJQ2HPmNm7BPni7oRA9R6wTNKhse6pwDSKMwkYCTwkqUOJekkmA2fGeSkk7SJpB8Lw27IYZLoDByWuWSUpJ+/4NrCDpG0ltQU+X6KtR4DzcweSikovSPpUYr5rf6ANlfemHMepRtZa+q3GSDt0NpTwZZ1kAnAKofdRFjP7KE78/zIOf7UiDMUtJPRYbpG0AuhHmIvZkRBwAOYB7yR6T6fF+psBi4AzyrQ9LgaZiZKOLdd7MLNH4tDdM/F7/X3gK4Sgda6keYTAOTNx2ShgnqTnzWxYnDuaRQjIL5dobjhwU7TZKt7zuUXqfgn4qqRVwArgpCI9SsdxaoyWnL1Z/j3V8slyMcB/Mu7db1nFjwxnrUezslXbTO1Vs3+tV5SUkK+Y385dlKm9rPVopl5xvsrXKs1Tpxye+v+0/93TGt3exsSTajqO41QDtjb91ggknRgfz1grqU+JesdIeiUuzhqRKN9N0ixJr0m6V1Kbcm02ONBI6hmfG0lus8pf2bzE527y/f5Mc/uVRNIZBXy8qbn9chyn6TCz1FsjWQB8kXVTExsgqR64ibAoay9gqKS94ulrCI9kdAOWEZ6tLEmDBy7MbD5QdOK6WjGzE8rXal7M7A7An7J0nE2JjTTJb2Z/Bkg8UlGIvsDrZpZ7SH4sMFjSnwkrkE+J9e4ErgB+Xa5R33zDzADOqUZbm5q9avat2u1Vs29Z+wXMSWwV+wlMBfoUOTcEuC1xfCrwK2C7GIBy5Z2BBeXa8jkaJ0klD7VuTFubmr1q9q3a7VWzb5lhZqPMrE9iG5U8L+lRSQsKbINTNlGou2MlyktSxWt+HMdxnIZgZo3NGLKE0FvJ8QlCSrF/AltLamVmqxPlJfEejeM4jpPPbKBbXGHWhpBIeaKF8bInCENrEJ5p/GM5Yx5onCSjyldpFlubmr1q9q3a7VWzb1WBgq7YEsLD8Q9JmhzLd5b0MISck4SMJZOBPwP3mdnCaOIS4DuSXidkbflt2TbjhI7jOI7jNAneo3Ecx3GaFA80juM4TpPigcZxHMdpUjzQOA1GZR4tbi4SUgqZ/X1naStLkp9BtfroOP6H2cKRtFmUPEDSHpI6l7umjL2kiNtOjXKO7L8cJW0PjJC0tZmtbUwwlLSTpB8AZGBrC0n94/5BUU+owUR73c3MJB0iaWezbLWAs/ohkfUPkiz+ZhTJyp5TGn+DWz6dgK9LugG4mcY/pPsFSRdJGkSQlN6+oV8kktoD+8T9gyT1aKRvEATyugDfkrRl/CKu2L94TSegl6SrABpqK2LAjyT9EbgeaGxQ6EjQdvo5Ic9Up0baS/YEu0K430bY+kQusDbyfcv9QLpO0jWS+mcQ9PcEbgFulXRE1gHa2RAPNC2fRcB/gDOBmWa2GBr+K87MxgFnA/cB37WggNrQf/ptgOMl/Qa4lyDX3SjMbAZBcfWTwAWSOjTki84CLwC3E6TAv58rr9SWJJnZBwTxwCOBxWY2N56rb8iXppn9nSA+eB7wBzN7LmbcbXAPIt7b5whCfDn124rtKajPzgfOl/T5hO2GBPw9gbsJT6T/P+AeST0bGgSjvbsIYoovADfEHzxVOxTcEvBA00LJ/dPEf8iphC+5PSSdFsvX5v7BKrVJyCz9F0KqcRr6i9DM/gG8Snjq+A/ASw2xk+fjsYRA2JbwpX6RpK3iF11Ff+/xS/dc4D3gGEk/iX6n/tKMQcZiEFgIfAE4WNK10dYaYMsKfEq2+yjwNeBkSadGWwCtN7wyle0DgZ8D3zezhZI2T/if6r2L9QYB9xDUd49oaLCJf5/Dgd+Y2Y/N7HqCGm/3im5sfXv/A4wys5sIP27+ChwnqW9je15OCTZGplHfmmcDDgG+DOwTj08E/g84HuhGeMK3fQo7uQd79yFoU+SOpxH+aQEOBAam9EuJ/a0JmhejgG8Bu8XyzRpwvx0Ictu94vExhGGq7wMdGmBrBtA7HvcDfgdc2gC/jiJ8gX8xHu8EvAFcDvQAxsX3QSntHUnQBBmQ+JwXAScAhxN6m+3K2SMMo9YnjocCtxJk1L8NPEJIRVJX4f1uB7SJr5fEe/9Csb+BEnbq4r12TJT9MPc3V4mtXD2gS+J4IjCF0Ct8DfhcpZ+tb+k279G0UCQdDowFDgJmSPqchWGv2wlfIo8Cc81sRRk7uV+0A4HxBP2Jn0naDfgMsKekMYQvt7JKe/DxL9tjJf2WkDPpWeBGoA/wGUlnAndJ2qryO2dLQhCFcI9vE4LtxbmhpZTUs37PYB7wIjBM0o/SGonzFD8HlgJXSvqumf0/QmD+LGF+ZYyZ/dvit18Ze/2AnxICyfckDQdmAl8FLgCuAu41s/+WsiepNaF3tZukL8Ye1hTCZ/AAkEtB8hfg0LT3C2Bm/wRWxdfbgXeAIyUdoLBw4dA092qhp/yEmS1LFD8HLI/3cKCkHmlsRXtmZn+N1+4E3GxmR1vo3fyIkFYl1d+wUyHNHel8y34j/Er+FXB0PP4CYZ7m2Hi8I9CjAnu9gfuBroRfqT8n/KLuSvgy/hywfwo7uZ5Qd8Jw3uXAtYQvoy5Az2j3CWBIBfY+Tcgi25bQWxsFHBLPHUqY+O2e0lY3F1dv+AAAIABJREFUQu+ijjAE9zDQLZ77TPRvv5Tv2+6EHuSX4/GBhASE3yX8um4D7Jhsv4y9PQi9rKPica4nOBzYDNgC2KkCe18CXiYMXx4Ty1oD2yT+jl4C9m7g32HuPd0xvpcPAB8Bxzfib3sgIZjuB7wOfCaj/5lBhLmb+izs+Zb3/ja3A741wYcKFwPPE4aiNotlnyesdDoxpY3cl0Q9Ybji38AeseyT/7+9M4+3sqz2+PdhRhBQEwVxRhCVCJWTM6ggCiihol4HFE3LSs0c05wySU3LMtPKikxTQ9ObgTldzdLS8BoqqU3e5nuz2y0bnNJ1//it1/2w2cO7D+979jnwrM9nfc5+h/07zzvsZz1rBq4ArsLNVC2MbZJPbHN8exsf75ci/KHxGJrgzQYeQW1nv+CC5gNI+7gKmZT2zjm2mcAPgI8jbbDDx/YL4FzgBWD3Fq51CipKeGc0ee+INK0zWp3U/N79J7Ao2jcduBFpqf1y4vTyvwOROXApMAEYFD3zPZF5b/9WxtjgPZoL/AM3T+V5tnXwpqH2wcuA/Qr6veyMAgOS6awkbvsAEhf4MLUCPdE/n4JW8lOA/r7vgFZWgMBO/rcfst3fBmzi+7YEPp0Jh5x4vZyfBR6K9o8FzvNJb0jeScgF3iPIn3I+0pLWQmaliSjIYOecWFshs8yGqDXtfwDD/Ni+jjW5CUY2qW4JjPbPOyAN8KMRXgd1OhvWwdsMGI20oInIfPnx6Lz9gPE5rzPDnAy8x+/Xu5HZbKofG+33Ycf4O6vwXgZ/vodF250VNGPQQmVaQb+ZSUjLPKCIa01c5z63ewCJC3iImrx7oxX5LcDxvv/DaKU/LRM2vj/vRH478Kx/HgQsQNFEmcO+aSBBhDXBv7+Wj/Up4OvR8a2BLVq85pFIazkMeJSKiWsnoG+L93AUcIlP2j+IsHYFBrSAMwuttr+ChOAWKBjgE8i3sk6L45qNfFiLgJvQYmKS41/VyfdlBjI7TY72zQMeAM5GZtam5sFa7xFNtLR6QiYvFlr0bNsIq0W8PsDGzfASrxq3fQCJC3iIHlGFtIH9kM/jPb7vQuCLuOkmJ95a0edbgB9l+1EU1+1Ia8gdjeQT7hK0sh/gwuZxlAPS6vWOQXb6dXwsvwC29GP7AN8HRuXEyvxCA11Y/RZY34/tjUxfI3NijUWa0IYuIJ6nYo6agvxmY1q4zrE++a+N8qCWIT9Mb+TvuZnWNMqA+ofcg/vUkInsTGB95M86jRzaAhXNaBoKsf5gg3ObCp88WPH71ujdawEv1PqcuHhu+wASr8LD04SzKcrz6PB9Q5Bmsxg41vdt2QLmdsh/MC7adxvwuH9eu8XJbSweUurC5htIc+hNRbN5Rys/dBTy+hXk/J+DTISfRtFlz5DDrxBNRjsg385glBd0M/JJzfaJfXYL4xqOHPMnA4/hGhqVMOShLT7fLZFv6GSkZW0RjTkAQzr53ixwvC8jc+XX/BmHqsm8WXj0dL9HewB/Bz5S6x3Nrh2Zz2r6kYrEKgMv8apx2weQuICHKH/MH6nkfPRFK/07cbNAC1iTkJnmdHz1jTSZPwEPt4g1FGkL10TCZnOUNb4QX+23gDc4+vxt4DL/vD0yS52H+6BqTZJVk+gA//s2lHuT+bYm+MR7MZUovYbmGeRMnoBMeYuBHwMj/Niufr1NNZkIbxcktEYgc9lSKoESe/sEulnOexbnQM1CvqjNgROBd/qxDqS59s+DmeEC1yONcBoKsR5VdU4f/zsM+c92KRurDLzEq85tH0DiFh9YZeKYBJyQ/UCAI1CuRodPvHfSJKS3Cm88igAb4XwDEjZb+iT1UZo4w2O8aHs7pMFcQWVFfhJaUW/eBGtTPLLIJ8jPAPN8eyMUoDCu2Zj8/JFUQni38Ik1u3fbAHcDE5pdT43js1EkWBZyPANpaaeiRNGngVktPN9ZKOR4im/PQ87qs5Bms7wVvGiMT6Hw4puQxpYFJkz38beiuXUgwXoB8j09SEUQHoknpvr2MBRoUDNar0isMvASF8NtH0DiTjw0mcaeQyajRSjZcQhwDMqzWIqHD+fE2xeZnL7guAcim/21qEzHr6OJtO7ES0VozUAaxpeATdAK/SLHmo8yznfKMa5dkNAcjMKCD0Sr+YtRNvc1eSdIv8ZtkZY1CUVcPeET0vHAh2gxvBXYGGXOb+rbo5G2MAFpmRdSERh5QrU3Q4Jpmwh/A3/eJyOBvVdePD9voN/3ib69m+Mci3xJV9BCxJXfw4eo5Ma8SkWT3t7fo919ewDKI5pSNlYZeImL47YPIHGLD0x+jWuBPXx7PIoUOt23hwPr+edmq/FeyHT0SDSB7QC8iEwOA5CgeHsL49sVrbpn+jjvQKvMUUhD+nfcJJUTbwgqdZP5m0YiwXAVygt6AfmNmgYmoOCBG4D5vr018uvcC7yChGyfFibxMUio/5uP51vAG3Qyv8PHcxtKsP2kT5q/opJ82rLDGpk9v0tFE+yNh777dr+82P5u3AB8ONp3id+DL/nfA6JjG1EnkrBIrDLwEhfLbR9A4hwPqaIpZNFltwKXRsff5ZNcriiw6kkF5WVsFG0fSYuhs9EYz8B9J759Olr1Zz6RgbXG0GR8U1H01fyq/e/FBW4LWAcj89ERVCLC+iFz3l6deDanIX/YbN8+FmlJuQVWFd5VPmEe7NvnAhf45zzCNM7lyXKeZiIzY5YnswcSaGvnHSPSjDZGQutmYHh07J3IRDq+2bMtGqsMvMTFc9sHkLjJA1rRHPUJtMKfhEKWs1XqBORjWL8FvL2Bs/zzrcA3o3MORcInz488w8smtQORSSv+sd9GxRzUbBKKx3e5T9yjkBntIRcQg2t9pwHWNBQocCKyy89EDv/D8JItzbDqYfvnLHppMtLmclUiqMJbKXQXaYLPkNO8E11vFnH1MDKNnQIchbLfr0bVEnKbCZHQugpFEG6EwqPPoIWQ+TKwysBLXA6noprdnMzMQgi7o+zyfzezl1Bm9H3ACSGE21HI8HWm3jB58GagH+cy33cosF4IYXEI4SNoFX2r+S85B95+wJIQwkaoNtYo4MAQwo5BnSTHoxpXNMOMxncF8qEcBpxpZo8iU8gHHbtX/J0GWDP9e88gM9mFZrYYRa3NBaZ6gcmGWHWws5Lyb3oPls8AZ5vZA3kwYIV2Dm9WfZ4U4T3Uwph2RP6nucgftgwFP/wCaar3ozJEi1soiT8IlY95NzKnnoDyb97ficKnRWKVgZeoDGq3pEvcnJEjPdM+srDM3mh1vhM5tQU/Jysns6dvD4iOHYUmpykt4GVO1snRvh2RVnMLCk44oBlOFeaFKOJsb2R22zg6NpUcgQTR+Z9Agm82inTbJDp2ME1qtVHREjanSR5M1ThrmrkivC1onsiYJaGGenhV5w9Cvp0/4jk2yKT0eeC4Trx320eftwM+4vijkCbxADn9HEVilYGXuFxu+wAS53hIKhJ5kX/u638nkSN8uQ7eIuCMqn1j8giWGli7Awv8c/9ofINx27lv54lWyybHj7qAeivDH5kOD80xngwrC9+9BlWefpBKePUBRGGuOTAPQLk2b6/+P9lnKuazpoUyUQjz3UT1yaiT9Z4Hr2ocY1HAxUIqfrHTkHbUq8lz6BN9Hohaf98X7ZuAfIE3IkFZtzRPkVhl4CXuWk6ms55BT6MugPsA64YQJuITSSsgkankq8D6IYS9fX8HSnDrTOfC3sCRIYRtzexVM3vdxznLzF42tRzGfDaoNSazt0xcF4UQ1kGakAGLzey3IYTdkKmvoWkwwpoOfCSE0B/dp7cDd5jZL0MIeyCz3J/yXJybsC4GTjWzp0IIg0MI6/v/yX4/vczsjRDCMNT3Zu0GeB1IyzrLzJ52vD4mc1nWLyf4uUOBs0MIg/KM1UTPo5ybV4DvhhAOQX6zJWb2ZoPn0Bc4PoSwRQghy0M5H/hdCOEOv7fLUJkeQ9Fqr5SNVQZeojZQuyVd4tpM7WipxWgSfpQWEuxqYG+MosHuR1E6z7MK5eCR3+Re5HTfGwnGVpzNuyHfzq6+PRBF0j2AfCnLyJmkiDSsn1EJ/w4oXPg5JGCXtTi2Q1CU2gSkGSxGvqOs6GZcxuR7VCX/sXIXy4NQ+ZyNHO9elEzZvw7erk3GV89ENxqt8B/GzZpEWkGd7+yJQsZ/GV3fUCSsv4/Mj4/h5Y66CqsMvMRdy20fQGJ/EBWTT3WpjNhEMxLlvWxefawT/28tFzhTiKrh1jl3MA3qavnxE1FBydsyoZB3fMi+fq5/znxQvZDGtjlebqURXnT/Mu3jrUnbP2+AbPcNS8GwshlvHVQT7GnkcH4H6sa4d/SdYX7tu1Vh9UXaxDZIwFzt5/7A+X3IzHMDEtAhwnuAGhnryDR2GDlKC/n/vRwJ2IaVtnE/kP/fv1EJB84i4C5DWm+uOnJFYZWBl7jrue0DSBw9DK3avkFVyG2D85s5k3PhROfXqg82FPk5jqS5M7xfJChaKZL5Pry/ChUfTwc5e6xUYb0fmcbiCtQ70VrS6QxUwudaKjXLspybHZF2lPVqGYCSUmvm86BQ6ueRlpWVwAl4uwBk1vtJNHkOQHk5k2s9HxR9+DrwdeT8Hkgl6TLThjKB1RcY5xNxzWrW0bmDon3vQoVas7ybzZD/LcTfKROrDLzE7eO2DyCxPwitVK+lUn+rbj8NGletzX5wE4HfASc1OLehKSU67xgUuTQ3nsBrnNdKQuEuKN9jF58sfoxMXCNRJNuzNOmJEmHtgMx2WzjebUhDGIFMXkvJ2QkUBVk87GP7Bsoqn4AiujqQwJgVnd+XGtFNrKiJ3oxyWDrwHCDk25pMVRdLVMa/UQZ81gZhSyTgPoWqMa9XdV4msIdRp6ZcdP9mo1yqm6kI0EOBf6IcnOXN7l+RWGXgJW4vt30Aazojk0A/ZFZZ5pN6IyEzzH90GzXA3BclJF6Pcgze3wBvHbRKXknoROd0oES455HPYiXNJjp3IFEIcYPxLUda0ps+uc9CtaduQSVxWqlh9hxK0vs7ErAnID/IEuTPylX3DZXbWQRc6dvBJ/LrkCazKStGntXUKKNJcgyqGTcChWU/SKVI6BhUa2xi9L/yVna4EzjPP8/z634c1WubHJ23ju+v67dAgmsp8hkt8Xdwfz82E0Wr7ZNzXIVhlYGXuH3c9gGsqczKvoABKLHwSqpW8qzoIH6Q+pVwA7AuKoueTWgdwH8DJ0fnxavd+2mQyY4c9U+hUvhXICG3gmbDikLwYdxZWwdvbeRQH4v6yizDhSaqerAuOXwyfnxdn4BGoyCAZ/GMcMfahBb8WUizWuD3b69o/3XIT7N2C893BtJizvdnNgjlKN2PEmL/Ro52zlWYccWAi5Gm9QyqnjANhYVnfYmGIp/Gbk2wzkGmxf2RgD8fCe6D/HhTrbdIrDLwEref2z6ANZnRqux+ZJY6Bwmbq5HAmVR1btY3o+bEUXXudajGU/aDnYfs+nOjc9ZBEU91hZb/PRn4fLT/PT6BHuFjih3Y99KglQBKrNsItZg+3SePrOfNEeRoHxxhbeOT6Sko3PUxKgmOR5DPWZ6N/R2oGOgYv45TkRlzSvz/WhjbaLQS39TH9xiVvJ6pfk+nrsJ7MxxVhngZ76Tq+7M6csOQJrPSuxJdc5yoOwTl9Wzq2w8h7fJtOe/fKmOVgZe4+3DbB7Cmsk9sS5Gz9nzgad8/HJXrv4KKttMbRQ9NroGT/Tg3olJ082MocW0t394BOY9/g/wffZCdf0oDvExL2Q2F98Ymo3uQaW5D385Wz7WipGI/yn0++V6HTHqZQ3wi8tHkDXXd2bEmICH982gimojMcu9sgpFd33QUMns1aodwEBKIp6JosD1beKZxFYFzkFbzGDDa90+hKpGQTjqvkVbzQyrBCnGS52Rg5xrfyRYe+/j1noJMgn2QPyrbvi3H/SsMqwy8xN2L2z6ANYlZ0UG8DzJ3zPQJYzPfP9In7m2qvrtOPTzH+CXyS1zo+xai6KXPIJPSVsDH8eZe2QRVB28vpHEc4pP5Z317Jqpb9i0is49PAo2aUU1CEVLHRvseRL6GTyMNKa9PZjvkNzk8uw6Ub3IN0kKW0aDkDVEkHtLqFlOJYNoHCbDpSCs4A9gu73OlolEMQM3EXqQSrbYHEtAtdTxt8D/7okCFQ6jy7VAlvIh8akjY/QQFTTzl70eWt/R1ZJqqm2dUJFYZeIm7J7d9AGsao9XmOORT+D1a8WaRSHshYbF2dP5KDmJWXL2O88l6KlrN3wBcHv2vecgfMgXZ8zdrMr6p2Q8a+D/U+2V74ExU2uQRWsxXQHkfPwEurbq2uSgIIGspnMeP8h6Ug/JJKn13hvtkdFgjLKQZXk2US4OqYM+l4rc6CnjAP9eNsKuBPR0J9rNRVNTOSJie5+PKLUxb+J8dNGlBjEyMz1HRfM72Z7wD0qizEkEbIO2hUcRbYVhl4CXuvtz2AawJzIqazKnIl9EXOXSXIy3mXeRo0+vn7odyB94G/BZY5Mf6oLDXhcAXo+9s59gN81L8+9cgDWSST45ZrbH+fnxEfE15hIOfdxBKapxJk+TBHFhH+zUegPs+WvhuP5Soer1vn44CMLKggQ4U8JDb0YyE+NPIzHgzipwb5BPpV5CT/q0cmq5875BZ9hLkfxoPzEEa4BNUgjDmEgWLlI1VBl7i7s2p1lkXkJlZCGGXEMJAZN55CJl3zkPmlKtQRNLpZvbtJuXbt0NdFwea2Z+QxjExhLC/mf0LmdAWAANCCOP9Oz9HvoanY6AQwlohhLH+eSyKCHsSlUa5BniXqdbYUagr5r/M7A/ZNcV/61FU+v52FLl1CrBvXJo/L2W1xczsq8jZfQAq81+3tph/b1B0zsb+d2wIIYvyGwhcHEL4MvKPfcPvZV7aGCWKvolMlGea2T+AP5jZfDM738y+42NveL+KJP9fy6kUnFwLaZZ/RQLwRW/jcD7KD+oSrDLwEnVzarekWxMY/YiWI5PPXOTYvzg63p8Wqs36+YuAd/v2/qil8UzfDuTQGtCk+Gk02T6AHPWHI7NFFh79dh/7tFW4/lijm486ITZt0lbn+7HZ8GRkKtygyfd38fPmIb/QSBQ8cR/wMT9ne5qY3mrgvgNFmO2LAi3+k4o5bzoqrZPb/FbSu7cdMk/dQaVG2KFocfOIP4usO2izcPLCsMrAS9x9ue0DWN0ZGOl/j0AFIqejTOe/Ald0Am8PnySno8iyrMvmTOR8zu0/QT6LBcBLrCj4FqD6WHejfjBNfQusGHFVK6EzFhZ1k02rsGr2bKkSNpvmvNZbUaTbQdG+Uaj8/+fzYNTAPAuFVmeRgtcj09xuSDi33XHtYxuDcnc+FwnS3n79K5hCuwqrDLzE3ZezH3SiEiiEsC4K5X0B+WNuQyvrR9FEPghlrf+mCU6w7BcXwmlIyOyPVtKHonL6N4UQZgMvmdmDefBCCGshn8SuaHV+l5nd4Odsh/I0epvZT+MxNMA9ANUtO9PMnqox9iz7/Y0QQm8ze6MB1ixkjjrT3OQXQuhlZm9mn/HK+PWwaty3CSiJ82gz+5Xv3wSZCc+xKtNig7Ftg57pGFSF4FJkfvsAup//AK4ys7vy3LeuIDfbnofeuW8Cj1hr5sFSsMrAS9QNqd2SbnVjKqvxrOz7BshccwEy9dyBflAb00LjMmT+GYhCZz9MJTv6MPTjPLp6DE3Gtz8yka2DAhMOQYlws5EmcUYjnBq4k1Bo8TjfHoybx6jkSMQVBD5MnUx7NFk/i+fuOFafKoy4WsK5RIUXq65zPFFHTuR8fsSveXNkKmzFbDnS79siVEJmIXBjdPxtVPKf2rYSh5XNjcjkeiUSrCuFy3cFVhl4ibs/t30AqxNHk9s0ZKq5NBII85Fv5k3gvS3iVvt4LsN9C378KJoXPewbfZ7ieFn/l/7+Pw5AIbm/Aqa3OMZO922hpJ4tSPN7AZXF+SHuy/Hn8jxyPufO0EfBEgGFVv8OOMmf62+A49v9/vkYa4YAV03oY7saqwy8xD2H2z6A1YGrVmg7o8ivg5HpZyFwvh8biExo+7aAvco+HrTKvoJKQuFReLQa6rHyJMpk3wCt0puW56ciVFepbwsl9Gzx42NQRnlWfv+LSPhl1QymAjs0ucbYFzQamXeO8u2rkCY1GfgD8vU0bKNQ0rs3hIrmuBnKdRpCDW2KJkU7i8QqAy9xz+W2D6Cns0/Mc6gUc3wXlcq6A1HU1k14c7Hoe3mcr+v6ZHkZMh99B5l6NkMROUtpUinZcXr7RLkpijTbCYVVfw8lQB6Gkg23b/HaC+nbQvE9WwYjreVpFKKdHbsWaTcrVUWogdMPCeOBKDrqGqTx3Y4WD4cCH/Jzt6VJ4mRJ794gv6b3+QS+IVqE9KJGKwkqmuBgfwd6lYFVBl7ins0pj2bVaRIyn0wNIQxGfTKOCyGMMbOXTU7x/mjF/xaZ/6qqKcs7CSH0N7M/I/PMWGRCWoJMZy+iEvtHmtmv6w0shDA8hHArMjn93Mf5WeAXyNR1oJl9Hi+Bg8x6uSiEMAllcl+Leqh8zPu5E0LIkh5PN7Ol/pU3gNPM7OHqazWzxSg0+O/An0MIg/3+vBRCmIzyLM6yirN+EHCGmX23GgtNYn9HwnkJ0BFC2Nn/z4lImI1qdn1m9hoKhvgrEi6Xm9m3zOwgH+cxwIIQwl5mttzMHs1774oiU67Od5BQPwgJ5N+b2Zs+/rcohNDHFISRaYKvmgdWFI1VBl6ink0p6qyTFEIYjkrJ3xJCmINMZd9Gjvn3orpZFwCvIX/DcWb24yaYWTTYNGR+egH4kZndHkKYj8rNnA68z8yuyzHG/ijkdl30Ywc5+SeiKKvnfOwXoBppd+a89k2Q4/bXZnaaT/JZt8frkSAcapXIs5WiwqJrHYNK3fRBmsG5aFK/24+9hDSQJ/3/hHqTkEe9HYO0mhuQ8DoBCbn7zOx7Oa8vG1s/FDgwGlUO+Et0TgcK7viamd2TB7dIyiLwQghbofygEaj1wK7IFPoiuq9rIX+e+UR+G3BRfC+KxCoDL9FqQO1WqXoqI41gIZU8loOQiexQ5FM4CfkY7kGaQyOsMn08A5DWcbd/7o98DYtQuZqtadDVsw7mZhTQt4UCe7ag/KIfoPDl64AnfP84JBQvJWpr0OxZoMmxFwo4OBZNjuOi6+9V/Z02vIPjkc9rPbSwyRrHnYsE7vG4qRKZrx6kvk+rMKwy8BL3bG77AHoyI/PVNcAxvn0wKp9/iG/3o+KzqNdnvTQfT3RutbDph8J876RBW+jq/0WBfVsouGeL3/upKLDgUSq1y9ZGoeStREfN8onvSuAC33cm8CcUSPHj6ufRhndvB6S1nRvt2w9pz/OpCkxARVxrltcvEqsMvMQ9n/uQqCWKE/DM7MYQwqvIP4OZLQwhvAkcFULog2pm/cPPrWejzHw8fUMIS6j4eG41s58CT7kJLJePpxaZ2SshhFNRef1FSBv7KAokeK3RdzOzVwhhOhIqi1GuzamohAvAsX5fHjSznzTBy+7fG8jMuC0KcDjCzP4SQpgCfN/M7q/xnVo4GfVDbRBeQiHlfwgh7I8E9/vN7JVG44pwJ6BouRnAhcC27jO6PITwvyg0/GwzW54Hr0T6OxKg/UMI65vZiyZz42AkbB9A/iUArLE5qkisMvAS9XRqt6TrSUwlaXAPpP4f5tuzUAOuzIx2MM3zWoZH35+DtJZ/Q6atU9CEviPSaJ5ohleFXTNiB2kzC1FeSjMTUqF9W6hoRYX1bEH5Su/H+8ajqKYlPt4ZKFJtRg6c2HQ5ya9nT2Qa3ML3b+t/+1d/p4vevTgBdTwKRV8Xaaln4xpx9m51FVYZeIlXP277AHoCs2JzpukoefAin3yz3iWzkGng2JyYRfp4xqIQ5TztiwfQxOxDSX1bKLBnC9KEnkdBCDei4Ab8GWR12vbLgRNXSvgUikhbCvyUSoLnDCTEWmpJUMJ7OBMJ5xuRH+ton9DvQgEd67YDqwy8xKsXt30A3Z2RM/kuKu14v4DMPNnxb6DII/zHldt2TzE+noB8Ca+jLoRZ9Fc/P9676vzYib1SwcroWKF9WyigZ0skFEYhQT3Lt9/h9+2c6NxWSqxs7xNkFhRxtj+XI5Fm07BrZxe8g72QlvYoKkXUC2kOj6F6d1sgx3vTxmBFYpWBl3j15LYPoDuzT7bXuQAY6D+gTwHzo3M2BL7aAmZ1m925yOx2jG8fiBIbD29hEt8H+D6KIrvDx7gAL1kfnZeZ/obgFQeqjg/CI8Yca2OU1HkJEmif8wn5y8ghPqeF6z4KmcZ2IUo0zSsQIiGzn09iP/Px9PexTUBC/1N+Xq6EP7/mK1FJmezat/LncBdaWOxf69l14Xs4wP8uITI9Id9e1uZgcFdjlYGXePXktg+guzPwQVRiZDmK+5+D/Ao7+fEsrHb9ZhMRBfp4amDfSSVabR5yyD4OfIgogx6F6z4OdNTAKLxvCwX2bEE+q1tdEOyJVspHRvd1Il6Is4X7FpDp8Q4kQAdEx3pRMRV2uZDB2xcjU+0gVPbmP6LjRyFz6wq14srGKgMv8erNbR9Ad2dU4fdF4LfRvnej8irXkKPvCCX4eCK8rCBhB8qxmQA8g3I/piGTVEc2DhTxs1sDvEL7tlBQzxafzD4J/JrKKnom8gccRwutl+vgb42qJnyOFqo5l/TOVWu9l6IFQz8Xhk+g8O9naVL8tEisMvASrxnc9gF0d/YJbgQyr/w8Expo9TyOSin7en6F0nw8Vf9nONI8XgbeE+3PoryGIU1mJSHDilFXpyGt5iGipmIoEfIuchTc9PO3QebGCUggb4zybz6DorkeoAWTFBXN4zYXVtl1zUbmvYbN1KqwVjKrOf7Wjn19njGV/N7t7te1JQpUOD97HqjvzRFEybJdhVUGXuLVn1MJmgYUN9ry7etQtd5aucM+AAAIH0lEQVR3mtlLOb7fD02s9yPtZTTSNJ4ys6/4ORsCl5nZ0QWMt8P/3xxTHkncKGwy8JqZ/aDqO1m5lfEo8OCHvv8S5MCfgjSanYFvWo58lBDCSFSf7AUUPXcp0jiO9ONv87G81GpjsBDCOFSosS9wqpm9nOVqNPjOWLQweMQaNJnzEjdbIyHTMB+obAoh7IYWIdeia52CtN+L2olVBl6i1Z9SUc0a5N0bsRW7OWJm70WO6CdDCL2b4ZiSIZ9DocJLUTHLh4HLQwg7+WljgDEhhPWjwpCdpSeROWr3aiEJPFwtZHyM5smY3/Jx/TCEsIGZnYtWrc8gIfnHnEJmbeTTehj5fOaiStOTQwjH+//8UyaoGwmZ7L5XUWay7AN81s/5cwOMgOqdfQ24LITwyRDCQF8EkD3HSOD9NBMyeZ5x0RRC2CaEMN/Mvo/K6geU2DoAOD+EcEI7sMrAS7QGUbtVqu7A5MhDYcXGXA2TE6u+t8o+nhavpYMWStazin1bKLBnS57n4OcFZJLMW+6mkKi8kt69taj0HJqIghyeRCVvpiBhvxXKG/oqMKkrsMrAS7zmctsH0G6mhTwUVs5JaeqAZhV9PCVf9yr1baHAni15n0N2nxo9lzr4qxyVV9Jz2BGZO09DxVRH+DuzEFUouAf4uJ87sKuwysBLvOZy2wfQHZj8K95sssu14mXlZlDXoWicIW281myizkqprIN6tywAdo7O+xI5VqgovPg1lKW/abT/syg7/xXyO60L1zwoOCqvpGdyNfAqUYtvZI6aikKEXyZn6ZYiscrAS7xmctsH0F2YAle8NQRMbF5aiHw1bcstQFrHN30yn4MqKF/ik26uUu2RwOoH/Aj1FxlWdU4HSvDMHeZa5HOowu10VF5JzyCO9NsFmRzvRGHfA6vOHddVWGXgJU7c9gG0mylgxUuJPp4SrneV+7ZEQqawni1FPIcc196BQqtHxP/TP08m0uhKfgbZ/ZvmAvRE3z4R+cfGoxyhS6vuzUr3sEisMvASJzZLgqZyIzq54qVkH08J11lI3xZK6tnS2eeQE7svMgkewspaZ1f7yGYiTfAg5FT/hO8/C9WCWw7M7WqsMvASJ277ALoT08kVLyX5eAq6pupM7sN9Enkgus79fQLOlRGPtI0foxI1X0DRZIP92HGOlbsLaFHPoQXs3FF5JT6XK1B5n9m4hhkdG0GlFlyeZNbCsMrAS5w4JWxGFELoi8xJ9wC32YrJmg0TC0MIdyIz1MUhhHmolMlPUJXiJ8zsu37eULRi/4CZPV7e1awwtmkojPlnZnZvCOFWpMEcgRIxrwBON7MlDTDeuv4QwiQU3roUNRs73Mx+GULY1syWhxD6m9mrrSZjRv+r08+hJ5A3xbsa1cdbDzjO798BKD/ojrzXWCRWGXiJEkFK2FyBzOx1VNzyt7ZisiP1flxRUuECoJ93aDwT1Xs6F5l6XvZzhyIn/Ie6UMhsiyLAtgTmhRDOMbNDgf9BeS4nAaflETIhhP1DCJ9COTGHons12yeiGShpb5iZvQqtdQGNqTPPoTtTlogbQhgbQtjMzP6F6oJNBG73+7cHEvgvNlnQFIZVBl6iRDWp3SrV6sJ0o6gmSujbQjfv2dJdOXoW+6Ak3cfxjqSo5cFzKBJxGc2LsxaGVQZe4sT1OJnOCqTO1BorYQyZ9rEf6nm/LhKAp6J8l7cjTet3ZnZqjVI1tTAHoaivw4CtzexvIYStUATSfKTh3GVmd60Opq2iKYSwI3Kkn4Ui5t6NqlDfhJ7JUOSHer4rscrAS5SoFiVBUyB1F9+CTx5noF4vo1AexJeBW8zsXyGEicAbZvZUTryAfDyXonyZ95nXPXPTYW8zez0JmZUpEtKHotbY//QFyTx0L28ys+e6GqsMvESJ6lHy0RRI1g18Cz55HI6c/L8xsywE+Rjg6BBCHzN7Mq+QAY3dV7QfBv4JfDKEMMCPvenX3WXX2MPonygy74fAZ0IIA03+uRuRufX1NmGVgZcoUU1KGs1qRpH2cQnwF+AkUyn92cDpqKvn73LgrGRSc+yxjgNwfBIu+ShU2hv0QcEgL4cQhprZX9uJVQZeokTVlATNakqhxb4toQf2bOmulENIB1SMlBz+scKwysBLlCgPJUGzGlALk0cwszdqfD+g8NWTgUXAf6OAgTfM7LUQQm8zeyMKNOid4cSf11QqUkgXLfDTAiJRd6AkaHoglTF5hBD2QS15j0ZC579QqPaVZva/0Xl9PKBgCKoG8PtVvZ6eTEUK6aIFflpAJOoulIIBehj55FF4x0gzuxfVKTvczOagBlcnA3eHED7k4dm4kBmK2lOPKu1Cewj5Pb4HdV49DzW6WwBcGEJYL5usfSLv4xP7kBDCyOqJvEisMvASJeosJUHTw6iMySN08+oG3Z2KFNJFC/y0gEjUHSgJmh5IJUxGmX/nv4CdULjr1Wb2ZTO7D3VRfDyEMAwlf15g6hu/xlORQrpogZ8WEIm6DVk3KE+QOD9Tct8WuknPlp7GFFiCqEisMvASJ26VUzBAD6UQwnBUJmQ34INm9nnfP9AUyjwMuBetUHNrH6GbVDfoiRQKLEFUJFYZeIkStUJJ0PRgKmvycNw+ZvZosSNevalIIV20wE8LiETtpCRoejClyaP7UZFCumiBnxYQidpFSdD0cEqTR6JEibo7JUGTKFGiRIlKpRTenChRokSJSqUkaBIlSpQoUamUBE2iRIkSJSqVkqBJlChRokSlUhI0iRIlSpSoVEqCJlGiRIkSlUr/DzqSNhftsQtUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Testing the correlation \n",
    "\n",
    "correlationA = Twitter_Train .iloc[:,1:12].corr(method = 'pearson')\n",
    "correlationA\n",
    "\n",
    "ax = sns.heatmap( \n",
    "    correlationA, \n",
    "    vmin=-1, vmax=1, center=0, \n",
    "    cmap=sns.diverging_palette(20, 220, n = 200),\n",
    "    square=True,\n",
    ")\n",
    "ax.set_xticklabels(  \n",
    "    ax.get_xticklabels(), \n",
    "    rotation=45, \n",
    "    horizontalalignment='right'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Text(0.5, 0, 'B_follower_count'),\n",
       " Text(1.5, 0, 'B_following_count'),\n",
       " Text(2.5, 0, 'B_listed_count'),\n",
       " Text(3.5, 0, 'B_mentions_received'),\n",
       " Text(4.5, 0, 'B_retweets_received'),\n",
       " Text(5.5, 0, 'B_mentions_sent'),\n",
       " Text(6.5, 0, 'B_retweets_sent'),\n",
       " Text(7.5, 0, 'B_posts'),\n",
       " Text(8.5, 0, 'B_network_feature_1'),\n",
       " Text(9.5, 0, 'B_network_feature_2'),\n",
       " Text(10.5, 0, 'B_network_feature_3')]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "correlationB = Twitter_Train.iloc[:,12:23].corr(method='pearson')\n",
    "correlationB\n",
    "\n",
    "ax = sns.heatmap(\n",
    "    correlationB, \n",
    "    vmin=-1, vmax=1, center=0,\n",
    "    cmap=sns.diverging_palette(20, 220, n = 200),\n",
    "    square=True,\n",
    ")\n",
    "ax.set_xticklabels(\n",
    "    ax.get_xticklabels(),\n",
    "    rotation=45,\n",
    "    horizontalalignment='right'\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Applying Machine Learning Models to identify key predictors of Social Influence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To prepare the dataset for machine learning models, the dataset is divided into Test data and Train data. By creating a train and test split of your dataset helps in quickly evaluating the performance of an algorithm on your problem. The training dataset is used to prepare a model, to train it and the test dataset is new data where the output values are withheld from the algorithm. Transforming A and B individuals into A-B variables for better interpretation of the model.For the test data, values are selected from the data. For the train data, one variable is chosen to build the model. Here, the model is built upon the variable \"choice\". \n",
    "\n",
    "x = pd.DataFrame(Twitter_Train.iloc[:,1:10].values - Twitter_Train.iloc[:,10:19].values)\n",
    "x.columns = ['A/B_follower_count','A/B_following_count','A/B_listed_count','A/B_retweets_received',\n",
    "             'A/B_mentions_sent','A/B_retweets_sent','A/B_posts','A/B_network_feature_2','A/B_network_feature_3']\n",
    "y= Twitter_Train['Choice']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The train_test_split function is for splitting a single dataset for two different purposes: training and testing. The testing subset is for building your model. The testing subset is for using the model on unknown data to evaluate the performance of the model.\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train , x_test, y_train, y_test = train_test_split (x, y, test_size = 0.30, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_Train Shape: (3850, 9)\n",
      "Y_Train Shape: (3850,)\n",
      "X_Test Shape: (1650, 9)\n",
      "Y_Test Shape: (1650,)\n"
     ]
    }
   ],
   "source": [
    "# Looking at the shape of the data to test whether training features number of columns match testing feature number of columns.The number of rows to match for the respective training and testing features and the labels\n",
    "\n",
    "print('X_Train Shape:', x_train.shape)\n",
    "print('Y_Train Shape:', y_train.shape)\n",
    "print('X_Test Shape:', x_test.shape)\n",
    "print('Y_Test Shape:', y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR: 0.709091 (0.014693)\n",
      "KNN: 0.737662 (0.016054)\n",
      "CART: 0.700000 (0.019653)\n",
      "RF: 0.757662 (0.016470)\n",
      "NB: 0.547273 (0.013196)\n",
      "SVM: 0.523117 (0.006989)\n",
      "Xgboost: 0.772987 (0.021831)\n"
     ]
    }
   ],
   "source": [
    "# Spot Check Algorithms\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.svm import SVC\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "\n",
    "\n",
    "models = []\n",
    "models.append(('LR', LogisticRegression(solver='liblinear', multi_class='ovr')))\n",
    "models.append(('KNN', KNeighborsClassifier()))\n",
    "models.append(('CART', DecisionTreeClassifier()))\n",
    "models.append((\"RF\", RandomForestClassifier()))\n",
    "models.append(('NB', GaussianNB()))\n",
    "models.append(('SVM', SVC(gamma='auto')))\n",
    "models.append(('Xgboost', XGBClassifier()))\n",
    "\n",
    "# evaluate each model in turn\n",
    "results = []\n",
    "names = []\n",
    "for name, model in models:\n",
    "    kfold = StratifiedKFold(n_splits=10, random_state=1, shuffle=True)\n",
    "    cv_results = cross_val_score(model, x_train, y_train, cv = kfold, scoring='accuracy')\n",
    "    results.append(cv_results)\n",
    "    names.append(name)\n",
    "    print('%s: %f (%f)' % (name, cv_results.mean(), cv_results.std()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ True  True  True  True  True  True  True  True  True]\n",
      "[1 1 1 1 1 1 1 1 1]\n"
     ]
    }
   ],
   "source": [
    "# Logtistic Regression\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "from sklearn.feature_selection import RFE\n",
    " \n",
    "\n",
    "model_logreg = LogisticRegression()\n",
    "rfe = RFE(model_logreg, 9)\n",
    "rfe = rfe.fit(x_train, y_train.values.ravel())\n",
    "print(rfe.support_)  #Indicates all variables have been selected by RFE with a ranking of 1 \n",
    "print(rfe.ranking_)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.625624\n",
      "         Iterations 6\n",
      "                           Results: Logit\n",
      "=====================================================================\n",
      "Model:               Logit             Pseudo R-squared:  0.097      \n",
      "Dependent Variable:  Choice            AIC:               6899.8665  \n",
      "Date:                2020-07-06 12:17  BIC:               6959.3790  \n",
      "No. Observations:    5500              Log-Likelihood:    -3440.9    \n",
      "Df Model:            8                 LL-Null:           -3811.3    \n",
      "Df Residuals:        5491              LLR p-value:       1.1797e-154\n",
      "Converged:           1.0000            Scale:             1.0000     \n",
      "No. Iterations:      6.0000                                          \n",
      "---------------------------------------------------------------------\n",
      "                       Coef.  Std.Err.    z    P>|z|   [0.025  0.975]\n",
      "---------------------------------------------------------------------\n",
      "A/B_follower_count     0.0000   0.0000 11.5580 0.0000  0.0000  0.0000\n",
      "A/B_following_count    0.0000   0.0000  1.8642 0.0623 -0.0000  0.0000\n",
      "A/B_listed_count      -0.0000   0.0000 -0.2372 0.8125 -0.0000  0.0000\n",
      "A/B_retweets_received -0.0000   0.0000 -2.1213 0.0339 -0.0000 -0.0000\n",
      "A/B_mentions_sent      0.0001   0.0000 12.3501 0.0000  0.0001  0.0001\n",
      "A/B_retweets_sent      0.0000   0.0000  0.0345 0.9725 -0.0000  0.0000\n",
      "A/B_posts              0.0000   0.0000  0.6897 0.4904 -0.0000  0.0001\n",
      "A/B_network_feature_2  0.0190   0.0018 10.4503 0.0000  0.0155  0.0226\n",
      "A/B_network_feature_3 -0.0000   0.0000 -8.6146 0.0000 -0.0000 -0.0000\n",
      "=====================================================================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import statsmodels.api as sm\n",
    "\n",
    "logit_model = sm.Logit(y,x)\n",
    "result = logit_model.fit()\n",
    "print(result.summary2())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.618079\n",
      "         Iterations 6\n",
      "                           Results: Logit\n",
      "=====================================================================\n",
      "Model:               Logit             Pseudo R-squared:  0.108      \n",
      "Dependent Variable:  Choice            AIC:               4769.2079  \n",
      "Date:                2020-07-06 12:17  BIC:               4800.4870  \n",
      "No. Observations:    3850              Log-Likelihood:    -2379.6    \n",
      "Df Model:            4                 LL-Null:           -2667.5    \n",
      "Df Residuals:        3845              LLR p-value:       2.7700e-123\n",
      "Converged:           1.0000            Scale:             1.0000     \n",
      "No. Iterations:      6.0000                                          \n",
      "---------------------------------------------------------------------\n",
      "                       Coef.  Std.Err.    z    P>|z|   [0.025  0.975]\n",
      "---------------------------------------------------------------------\n",
      "A/B_follower_count     0.0000   0.0000 10.5261 0.0000  0.0000  0.0000\n",
      "A/B_retweets_received -0.0000   0.0000 -2.2968 0.0216 -0.0000 -0.0000\n",
      "A/B_mentions_sent      0.0001   0.0000 15.0594 0.0000  0.0001  0.0001\n",
      "A/B_network_feature_2  0.0220   0.0023  9.7371 0.0000  0.0176  0.0265\n",
      "A/B_network_feature_3 -0.0000   0.0000 -8.0290 0.0000 -0.0000 -0.0000\n",
      "=====================================================================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Removing variables with p-values greater than 0.05 to improve the model \n",
    "\n",
    "columns_to_keep = ['A/B_follower_count', 'A/B_retweets_received', 'A/B_mentions_sent', \n",
    "                   'A/B_network_feature_2', 'A/B_network_feature_3']\n",
    "x = x_train[columns_to_keep]\n",
    "y = y_train \n",
    "\n",
    "logit_model = sm.Logit(y,x)\n",
    "result = logit_model.fit()\n",
    "print(result.summary2())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Logistic Regression Model Fitting \n",
    "\n",
    "model_logreg.fit(x_train, y_train)\n",
    "\n",
    "# Predicting test results \n",
    "\n",
    "y_pred = model_logreg.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy of Logistic Regression model: 0.7103896103896103\n",
      "Testing accuracy of Logistic Regression model: 0.7260606060606061\n"
     ]
    }
   ],
   "source": [
    "# Calculating the accuracy of our model \n",
    "\n",
    "print(\"Training accuracy of Logistic Regression model:\", model_logreg.score(x_train, y_train))\n",
    "print(\"Testing accuracy of Logistic Regression model:\", model_logreg.score(x_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The difference between training and testing accuracy indicates how much the model has overfitted. The Logistic Regression model has a low overfitting value of 0.02 which indicates the model has not trained data too well to negatively impact the performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[674 146]\n",
      " [306 524]]\n"
     ]
    }
   ],
   "source": [
    "# Confusion Matrix \n",
    "\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "\n",
    "confusion_matrix = confusion_matrix (y_test,y_pred)\n",
    "print(confusion_matrix)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The confusion matrix visualises the actual and predicted values of the data thereby depicting the performance of a supervised algorithm. Here, it can be inferred that there are 1,198 correct predictions and 452 incorrect predictions. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.82      0.75       820\n",
      "           1       0.78      0.63      0.70       830\n",
      "\n",
      "    accuracy                           0.73      1650\n",
      "   macro avg       0.73      0.73      0.72      1650\n",
      "weighted avg       0.74      0.73      0.72      1650\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Classification report will summarise our model by computing precision, recall, f-measure and support \n",
    "\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In social media analytics, high precision means that an algorithm returned more relevant results than irrelevant, while high recall means that an algorithm returned most of the relevant results. In the logistic regression model, there is a 74% precision that there are more relevant values and a 73% recall that most of the relevant values are included in our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Random Forest Classifier \n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "model_rf = RandomForestClassifier()\n",
    "model_rf.fit(x_train, y_train) #Train the model on training data \n",
    "\n",
    "# Predicting the test results \n",
    "  \n",
    "y_pred = model_rf.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy of Random Forest model: 0.9937662337662337\n",
      "Testing accuracy of Random Forest model: 0.7751515151515151\n"
     ]
    }
   ],
   "source": [
    "# Calculating the accuracy of our model  \n",
    "\n",
    "print(\"Training accuracy of Random Forest model:\", model_rf.score(x_train, y_train))\n",
    "print(\"Testing accuracy of Random Forest model:\", model_rf.score(x_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A/B_mentions_sent</th>\n",
       "      <td>0.160686</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_network_feature_3</th>\n",
       "      <td>0.146603</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_follower_count</th>\n",
       "      <td>0.135332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_listed_count</th>\n",
       "      <td>0.127699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_retweets_sent</th>\n",
       "      <td>0.119870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_posts</th>\n",
       "      <td>0.093620</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_retweets_received</th>\n",
       "      <td>0.078480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_network_feature_2</th>\n",
       "      <td>0.075339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A/B_following_count</th>\n",
       "      <td>0.062370</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       importance\n",
       "A/B_mentions_sent        0.160686\n",
       "A/B_network_feature_3    0.146603\n",
       "A/B_follower_count       0.135332\n",
       "A/B_listed_count         0.127699\n",
       "A/B_retweets_sent        0.119870\n",
       "A/B_posts                0.093620\n",
       "A/B_retweets_received    0.078480\n",
       "A/B_network_feature_2    0.075339\n",
       "A/B_following_count      0.062370"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# To interpret the model and report the results, feature importances is used to quantify how much a particular variable improves predctions \n",
    "\n",
    "feature_importances = pd.DataFrame(model_rf.feature_importances_,\n",
    "                          index = x_train.columns, \n",
    "                          columns = ['importance']).sort_values('importance', ascending = False)\n",
    "feature_importances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Select features with importance larger than 0.08 \n",
    "\n",
    "columns_to_keep = ['A/B_mentions_sent', 'A/B_listed_count', 'A/B_follower_count', \n",
    "                   'A/B_network_feature_3', 'A/B_retweets_sent', 'A/B_posts']\n",
    "\n",
    "x = x_train[columns_to_keep]\n",
    "y = y_train \n",
    "\n",
    "model_rf = RandomForestClassifier() #New random forest classifier for the most important features \n",
    "model_rf.fit(x_train, y_train) #Train the model on training data \n",
    "\n",
    "# Predicting the test results \n",
    "\n",
    "y_pred = model_rf.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy of Random Forest model: 0.9937662337662337\n",
      "Testing accuracy of Random Forest model: 0.7751515151515151\n"
     ]
    }
   ],
   "source": [
    "# Calculating the accuracy of the model \n",
    "\n",
    "print(\"Training accuracy of Random Forest model:\", model_rf.score(x_train, y_train))\n",
    "print(\"Testing accuracy of Random Forest model:\", model_rf.score(x_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The difference between training and testing accuracy indicates how much the model has overfitted. The Random Forest model has a low overfitting value of 0.22 which indicates the model has not trained data too well to negatively impact the performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[629 191]\n",
      " [180 650]]\n"
     ]
    }
   ],
   "source": [
    "# Confusion Matrix \n",
    "\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "\n",
    "conf_matrix = confusion_matrix (y_test,y_pred)\n",
    "print(conf_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The confusion matrix visualises the actual and predicted values of the data thereby depicting the performance of a supervised algorithm. Here, it can be inferred that there are 1,274 correct predictions and 376 incorrect predictions. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.77      0.77       820\n",
      "           1       0.77      0.78      0.78       830\n",
      "\n",
      "    accuracy                           0.78      1650\n",
      "   macro avg       0.78      0.78      0.78      1650\n",
      "weighted avg       0.78      0.78      0.78      1650\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Classification report will summarise our model by computing precision, recall, f-measure and support \n",
    "\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In social media analytics, high precision means that an algorithm returned more relevant results than irrelevant, while high recall means that an algorithm returned most of the relevant results. In the Random Forest model, there is a 78% precision that there are more relevant values and a 78% recall that most of the relevant values are included in our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# K-Nearest Neighbors\n",
    "\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "model_knn = KNeighborsClassifier(n_neighbors = 5, metric = 'euclidean')\n",
    "model_knn.fit(x_train, y_train)\n",
    "\n",
    "y_pred = model_knn.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy of K-Nearest Neighbors model: 0.818961038961039\n",
      "Testing accuracy of K-Nearest Neighbors model: 0.7412121212121212\n"
     ]
    }
   ],
   "source": [
    "# Calculating training and testing of our model\n",
    "\n",
    "print(\"Training accuracy of K-Nearest Neighbors model:\", model_knn.score(x_train, y_train))\n",
    "print(\"Testing accuracy of K-Nearest Neighbors model:\", model_knn.score(x_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The difference between training and testing accuracy indicates how much the model has overfitted. The KNN model has a low overfitting value of 0.07 which indicates the model has not trained data too well to negatively impact the performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[583 237]\n",
      " [190 640]]\n"
     ]
    }
   ],
   "source": [
    "# Confusion Matrix \n",
    "\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "\n",
    "confusion_matrix = confusion_matrix(y_test, y_pred)\n",
    "print(confusion_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The confusion matrix visualises the actual and predicted values of the data thereby depicting the performance of a supervised algorithm. Here, it can be inferred that there are 1,223 correct predictions and 427 incorrect predictions. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.71      0.73       820\n",
      "           1       0.73      0.77      0.75       830\n",
      "\n",
      "    accuracy                           0.74      1650\n",
      "   macro avg       0.74      0.74      0.74      1650\n",
      "weighted avg       0.74      0.74      0.74      1650\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Classification report will summarise our model by computing precision, recall, f-measure and support \n",
    "\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In social media analytics, high precision means that an algorithm returned more relevant results than irrelevant, while high recall means that an algorithm returned most of the relevant results. In the KNN model, there is a 74% precision that there are more relevant values and a 74% recall that most of the relevant values are included in our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
      "              colsample_bynode=1, colsample_bytree=1, eta=0.3, gamma=0,\n",
      "              learning_rate=0.1, max_delta_step=0, max_depth=2,\n",
      "              min_child_weight=1, missing=None, n_estimators=100, n_jobs=1,\n",
      "              nthread=None, objective='binary:logistic', random_state=0,\n",
      "              reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
      "              silent=None, subsample=1, verbosity=1)\n"
     ]
    }
   ],
   "source": [
    "# XGBoost Classifier \n",
    "\n",
    "from xgboost import XGBClassifier\n",
    "from xgboost import plot_importance\n",
    "\n",
    "model_xgb = XGBClassifier(max_depth = 2, \n",
    "                          objective = 'binary:logistic',\n",
    "                          eta = 0.3\n",
    "                         )\n",
    "model_xgb\n",
    "\n",
    "model_xgb.fit(x_train, y_train)\n",
    "print(model_xgb)\n",
    "\n",
    "y_pred = model_xgb.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy of XGBoost model: 0.7958441558441558\n",
      "Testing accuracy of XGBoost model: 0.7824242424242425\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeUAAAEWCAYAAABYNo/VAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzde5zVVb3/8dcbNEVBOYoaVEhe0BFHRyHRRB0q9SiYklZy8IJSaidvIZonS80sTSXUrLwdIwnT4zVSD1rhRgQvyU3UxDTxh+RRyAsMjTjA+/fHd41ux71nNjCXPTOf5+MxD/de3/Vd67MG4fNd67v2/so2IYQQQmh7Xdo6gBBCCCFkIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgih3ZF0vaQftHUcITQ3xeeUQ+g8JC0CtgPW5BX3t/2PDWizGvit7U9vWHTtk6SJwGu2v9/WsYT2L2bKIXQ+R9junvez3gm5OUjaqC373xCSurZ1DKFjiaQcQgBA0r6SZkl6R9L8NAOuP3aSpL9KWiHp75JOTeWbA/8L9JFUk376SJoo6dK886slvZb3fpGk70p6BlgpaaN03t2Slkp6RdKZjcT6Qfv1bUs6T9Kbkl6XdJSkwyW9KOktSd/LO/diSXdJuiONZ46kPfOOV0jKpd/Dc5K+3KDfX0l6UNJKYAwwCjgvjf0Pqd75kl5O7T8vaUReG6MlPSbpKklvp7Eelnd8K0m/lvSPdPy+vGPDJc1Lsc2StEfJf8ChXYikHEJA0qeAB4BLga2AccDdkrZJVd4EhgNbACcBEyTtbXslcBjwj/WYeY8EhgE9gbXAH4D5wKeALwJnSzq0xLY+CWyazr0QuAk4DhgIHABcKGmHvPpHAnemsd4G3CdpY0kbpzgeBrYFzgAmS9ol79z/AH4M9ABuBSYDV6SxH5HqvJz63RL4IfBbSb3z2hgMLAR6AVcA/y1J6dgkYDNgQIphAoCkvYFbgFOBrYEbgCmSNinxdxTagUjKIXQ+96WZ1jt5s7DjgAdtP2h7re0/Ak8DhwPYfsD2y85MJ0taB2xgHNfaXmy7FvgcsI3tS2y/b/vvZIn12BLbqgN+bLsOuJ0s2V1je4Xt54DngPxZ5Wzbd6X6PyNL6Pumn+7A5SmOacD9ZBcQ9X5ve2b6Pb1XKBjbd9r+R6pzB/A3YJ+8Kq/avsn2GuA3QG9gu5S4DwNOs/227br0+wb4JnCD7Sdtr7H9G2BVijl0EO32Xk4IYb0dZftPDcq2B74q6Yi8so2BRwDS8upFQH+yi/nNgAUbGMfiBv33kfROXllXYEaJbf0zJTiA2vTfN/KO15Il24/1bXttWlrvU3/M9tq8uq+SzcALxV2QpBOAsUC/VNSd7EKh3v/l9f+vNEnuTjZzf8v22wWa3R44UdIZeWWfyIs7dACRlEMIkCWaSba/2fBAWh69GziBbJZYl2bY9cuthT7CsZIscdf7ZIE6+ectBl6xvfP6BL8ePlP/QlIX4NNA/bL7ZyR1yUvMfYEX885tON6PvJe0Pdks/4vA47bXSJrHh7+vxiwGtpLU0/Y7BY792PaPS2gntFOxfB1CAPgtcISkQyV1lbRp2kD1abLZ2CbAUmB1mjUfknfuG8DWkrbMK5sHHJ42LX0SOLuJ/p8ClqfNX91SDLtL+lyzjfCjBkr6Str5fTbZMvATwJNkFxTnpXvM1cARZEvixbwB5N+v3pwsUS+FbJMcsHspQdl+nWzj3C8l/VuK4cB0+CbgNEmDldlc0jBJPUocc2gHIimHELC9mGzz0/fIksli4Fygi+0VwJnA/wBvk210mpJ37gvA74C/p/vUfcg2K80HFpHdf76jif7XkCW/KuAVYBlwM9lGqZbwe+DrZOM5HvhKun/7PvBlsvu6y4BfAiekMRbz38Bu9ffobT8PjAceJ0vYlcDMdYjteLJ75C+QbbA7G8D202T3la9Lcb8EjF6HdkM7EF8eEkLoVCRdDOxk+7i2jiWEhmKmHEIIIZSJSMohhBBCmYjl6xBCCKFMxEw5hBBCKBPxOeWw3nr27OmddtqprcNoNStXrmTzzTdv6zBaTWcab2caK8R429rs2bOX2d6m0LFIymG9bbfddjz99NNtHUaryeVyVFdXt3UYraYzjbczjRVivG1N0qvFjsXydQghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFAmIimHEEIIZSKScgghhFBAv379qKyspKqqikGDBgFw5513MmDAALp06cLTTz/d7H1u1OwthhBCCB3EI488Qq9evT54v/vuu3PPPfdw6qmntkh/7S4pSxoB3ANU2H4hr3wqMAaYDPQGaoFNgAm2b2ykva8ClwD/Z3tokTrVwDjbwyWNBgbZPr15RlR+JH3P9k+aqldbt4Z+5z/QGiGVhXMqVzM6xtshdaaxQoy33qLLh61zWxUVFc0RUlHtcfl6JPAYcGx9gaRuwFa2l6SiUbargP2Bn0r6RCPtjQH+s1hCbiuS2vKC6Xtt2HcIIZQFSRxyyCEMHDiQG28sOrdrVu1qpiypO1miHQpMAS5Oh6qBXIFTugMrgTVF2rsQGAJ8VtIU4AfAr4BBwGpgrO1HGolne+AWYBtgKXASsAT4G7AjsCXwFlBt+1FJM1Kd14GfA5VkfwYX2/59moUPAzYFNge+UKTf84DjgbXA/9o+X1IVcD2wGfAycLLttyXlyGb5T0vqBTxtu1/q68up/o7AvbbPk3Q50E3SPOA526Ma9H0KcApAr17bcGHl6mK/ng5nu27ZFXdn0ZnG25nGCjHeerlcrtHzrrzySnr16sXbb7/NuHHjqK2tZc899wTgnXfeYfbs2dTU1DRrrO0qKQNHAVNtvyjpLUl7254DHAbcl1dvsqRVwM7A2bYLJmXbl0j6Ah8mrXNSeaWkXYGHJfVvJJ7rgFtt/0bSycC1to+S9CKwG/BZYDZwgKQngU/bfknST4Bptk+W1BN4StKfUpv7AXvYfqtQh5IOS7+Hwbb/JWmrdOhW4Azb0yVdAlwEnN1I7ABVwF7AKmChpJ+nBH96Wmko9Du7EbgRoO8OO3n8gvb2v9D6O6dyNTHejqkzjRVivPUWjaouuY358+dTV1dHdXV2Ts+ePRk4cOAHG8CaS3v7UxkJXJ1e357ezyGbPY/LqzcqJdltgFmSptp+tYT2h5DNYLH9gqRXgcaS8n7AV9LrScAV6fUM4ECypHwZ8E1gOvCXdPwQ4MuS6mPeFOibXv+xWEJOvgT82va/UpxvSdoS6Gl7eqrzG+DOJsYK8Gfb7wJIeh7YHlhcwnkAdNu4KwvX455Me5XL5dbpL3F715nG25nGCjHeUqxcuZK1a9fSo0cPVq5cycMPP8yFF17YMgHmaTf3lCVtTbace7OkRcC5wNcl7Qgstv1+w3NsLyVL2oNL7WYDw3T67wzgAGAf4EGgJ9kS+6N5/Rxtuyr99LX913RsZQkxuok6+Vbz4Z/zpg2Orcp7vYb2d5EWQggt4o033mDIkCHsueee7LPPPgwbNox///d/59577+XTn/40jz/+OMOGDePQQw9t1n7bTVIGjiFbKt7edj/bnwFeAY4DphY6QdJmZMuzL5fYx6PAqHRuf7LZ68JG6s/iww1no8g2oAE8CXweWGv7PWAecCpZsgZ4CDhDklJfe5UYH8DDwMlpbEjaKs1235Z0QKpzPNnMHGARMDC9PqbEPuokbbwOMYUQQoeyww47MH/+fObPn89zzz3HBRdcAMCIESN47bXXWLVqFW+88QYPPfRQs/bbnpLySODeBmV3A/vy8aQ8OW1Umg1MtD27xD5+CXSVtAC4Axhte1Uj9c8ETpL0DFkiPAsgnbMYeCLVmwH0ABak9z8CNgaekfRsel8S21PJNrk9ncZYvwR+InBliqWK7GNeAFcB35I0C+jVsL0ibkyxTS41rhBCCBtO9rqshJYXSZsAM2037532UJJddtnFCxc2tpDQseRyuQ82eXQGnWm8nWmsEONta5JmF8tb7foeYpqRRkIOIYTQIbTrpLwu0keSNmlQfLztBYXqtzVJlWQ7uvOtsl3qprUQQgjtTKdJyu0tmaWLhYKfFQ4hhNAxtaeNXiGEEEKHFkk5hBBCKBORlEMIIYQyEUk5hBBCKBORlEMIIYQyEUk5hBBCKBORlEMIIYQyEUk5hBBCKBORlEMIIbSINWvWsNdeezF8+HAAxowZw5577skee+zBMcccQ01NTRtHWH4iKYcQQmgR11xzDRUVFR+8nzBhAvPnz+eZZ56hb9++XHfddW0YXXlq0a/ZlDQCuAeosP1CXvlUYAwwGegN1JJ9L/UE2zc2Y//VwPu2ZzVXm6nd0cAg26eXUHcb4H7gE8CZtmc0cUqhvh62/Y/1CLXUPrYn+3PqSvZIyZ/bvr6p82rr1tDv/AdaKqyyc07lakbHeDukzjRWaL7xLrp8WNFjr732Gg888AAXXHABP/vZzwDYYostALBNbW0t6ZHyIU9Lz5RHAo8Bx9YXSOoGbGV7SSoaZbsK2B/4qaRPNGP/1cDnm7E9JK3rhcwXgRds77WuCTkZDfRZlxPWI8bXgc+nP4fBwPmS1qnPEELId/bZZ3PFFVfQpctH08xJJ53EJz/5SV544QXOOOOMNoqufLXYTFlSd7JEOxSYAlycDlUDuQKndAdWAmsaabMGuAYYTja7PtL2G2k2ej3QN1U9G1gCnAaskXQccBbwa2BHYEvgLaDa9qOSZgAnpbJbgB2AfwGn2H5G0sVkibEfsAx4OC+mYcD3gSNsL2sQbxVwBdBN0jxgP+AA4IdkKwMvAyfZrpF0IXAE0A2YBZwKHE32aMrJkmrT+X8lm6UvkzQIuMp2dcMYJR0PXJ5+35sAv7B9Q6Hfq+33895uQiMXa5JOAU4B6NVrGy6sXF2saoezXbdshtFZdKbxdqaxQvONN5fLFSx//PHHqaurY8WKFcybN49//vOfH9Q98cQTOe6447j22mv54Q9/yGGHHbbBcTSlpqamaKzlpiWXr48Cptp+UdJbkva2PQc4DLgvr95kSauAnYGzbRdNysDmwBO2L5B0BfBN4FKyRD3B9mOS+gIP2a6QdD1QY/sqAEkvArsBnwVmAwekRzp+2vZLkn4OzLV9lKQvALfy4ZOaBgJDbNemJeX65fmxwOG2324YrO15KdkOsn26pF5kCfxLtldK+m46/xLgOtuXpHYnAcNt3yXpdGCc7afTscZ+5/kxngK8a/tzkjYBZkp62PYrhU6U9BngAWAn4Nxiy+Xp9sKNAH132MnjF3SaB41xTuVqYrwdU2caKzTfeBeNqi5Y/tBDDzF79mxGjx7Ne++9x/Lly7n55pv57W9/+0GdjTbaiCuvvJKf/vSnGxxHU3K5HNXVhWMtNy35f+FI4Or0+vb0fg7Z7HlcXr1Rtp9Os91ZkqbafrVIm++T3Z+FLKkenF5/CdgtL2FtIalHgfNnAAeSJeXLyJL6dOAv6fgQstkptqdJ2lrSlunYFNu1eW0NJZvFHmJ7eZF4G9qX7KJgZor1E8Dj9e1JOg/YDNgKeA74Q4nt1suP8RBgD0nHpPdbkl34FEzKthen+n2A+yTdZfuNxjrrtnFXFjZyT6mjyeVyRf8R6og603g701ih5cd72WWXcdlll33Q11VXXcWkSZN46aWX2GmnnbDNH/7wB3bdddcWi6G9apGkLGlr4AvA7pJMtoHIaea6uMFyKQC2l0qaQ3ZPs1hSrrPt9HoNH8bfBdivQdIsNKucQbak3Qe4EDiXbHn30fpTCvRZ39/KBuV/J1vm7g88XSTehgT80fbIBnFuCvySbEa9OC1Fb1qkjdV8uLzcsE5+jALOsP1QibEBYPsfkp4jW2a/a13ODSGEYmxz4oknsnz5cmyz55578qtf/aqtwyo7LbXR6xjgVtvb2+5n+zNkM7TjgKmFTpC0GbAX2X3WdfUw8MFO6HQvF2AFkD9jfpJs49da2+8B88ju3dZvwHoUGJXaqAaWNTILfhX4CnCrpAElxvkEsL+knVIfm0nqz4fJdVm6F39M3jkNx7CIbJka0qy+iIeAb0naOPXVX9LmhSpK+nTagIekfyNbzVhY4phCCKGo6upq7r//frp06cLMmTNZsGABzz77LJMnT/5gN3b4UEsl5ZHAvQ3K7iZbvm2YlCenTVCzgYm2Z69Hf2cCgyQ9I+l5stkwZMu/IyTNk3SA7VXAYrLkCFky7gEsSO8vrm+HbJPUiY11anshWRK/U9KOTQVpeynZburfpT6eAHa1/Q5wU4rjPj5cTgeYCFyfxtCNbJPYNWlzWmP3328GngfmSHoWuIHiKyMVwJOS5pMt519le0GRuiGEEFqIPlwNbuGO0mYj24NapcPQ4nbZZRcvXNh5JtTtabNIc+hM4+1MY4UYb1uTNLtYLmy17YZplhoJOYQQQiiiLD8DkD6mtEmD4uPLeUlV0gXAVxsU32n7x20RTyGSKoFJDYpX2R7cFvGEEEL4qLJMyu0xSaTkWzYJuJB0UVPVZMUQQghtIh5IEUIIIZSJSMohhBBCmYikHEIIIZSJSMohhBBCmYikHEIIIZSJSMohhBBCmYikHEIIIZSJSMohhNCBvffee+yzzz7sueeeDBgwgIsuugiAP//5z+y9995UVVUxZMgQXnrppTaONECZfnlICCGE5rHJJpswbdo0unfvTl1dHUOGDGG77bbj6quv5ve//z0VFRX88pe/5NJLL2XixIltHW6n166SsqQRwD1Ahe0X8sqnAmOAyUBvoJbsazon2L6xDeI8G7jR9r/S+weB/0hPgyprkkYDD9v+R1N1a+vW0O/8B1o+qDJxTuVqRsd4O6T2PtZFlw8rekwS3bt3B6Curo66uroPypcvz55M++6779KnT5+WDzQ0qV0lZbJHQj4GHEv2mEXS4wy3sr1EEsAo209L2gp4WdJE2++3cpxnA78F/gVg+/BW7n9DjAaeBZpMyiGE9mHNmjUMHDiQl156iW9/+9vstttu3HzzzRx++OF069aNLbbYgieeeKLphkKLazf3lCV1B/YnmxEfm3eoGsgVOKU7sJJGnjksqUbSTyXNlvQnSftIykn6u6QvpzpdJV0p6S/pec2npvLqVPcuSS9ImqzMmUAf4BFJj6S6iyT1Sq/HSno2/ZydyvpJ+qukmyQ9J+nhdLGBpDMlPZ/6vr2RsRyUnrk8T9JcST1S+bl5sf+wsf4kHUP2JK/Jec9vDiG0c127dmXevHm89tprPPXUU7zyyitMmDCBBx98kNdee42TTjqJsWPHtnWYgfY1Uz4KmGr7RUlvSdrb9hzgMOC+vHqTJa0CdgbOtl00KQObAznb35V0L3ApcDCwG/AbYArZRcC7tj9X/0xoSQ+n8/cCBpDNKmcC+9u+VtJYYKjtZfmdSRoInAQMBgQ8KWk68HaKd6Ttb0r6H+Bostn2+cBnba+S1LORsYwDvm17ZrqAeU/SIandfVJ/UyQdCPy/Qv3Z/q2k04Fxtp8u1ImkU4BTAHr12oYLK1c3ElLHsl23bJmzs+hM423vY83lciXX7devHzNmzODJJ5+ktraWXC5H3759+cUvfrFO7bQnNTU17WZs7SkpjwSuTq9vT+/nkM2ex+XVq1++3gaYJWmq7VeLtPk+MDW9XkD2GMM6SQuAfqn8EGCPNIsE2JIsob0PPGX7NQBJ89I5jzUyhiHAvbZXpnPuAQ4gS/6v2J6X6s3O6/8ZsguN+/joxUdDM4GfSZoM3GP7tZSUDwHmpjrdU+z/r5H+GpXu0d8I0HeHnTx+QXv6X2jDnFO5mhhvx9Tex7poVHXRY0uXLmXjjTemZ8+e1NbW8oMf/IDDDz+cKVOm0KdPH/r3789///d/M3DgQKqri7fTnuVyuXYztnbxf6GkrYEvALtLMtAVsKTrgcWF7hnbXippDtmstFhSrrPt9HotsCqdu1ZS/e9GwBm2H2oQU3V9/WQNTf8+1cixhm3VLx0PAw4Evgz8QNIA2x+7pLd9uaQHgMOBJyR9KfV3me0bGsTer5H+StZt464sbGSDSUeTy+Ua/cevo+lM4+3IY3399dc58cQTWbNmDWvXruVrX/sa++23HzfddBNHH300Xbp04d/+7d+45ZZb2jrUQDtJysAxwK22T60vSMu+x/HhTPcjJG1Gtrx8xQb2/RDwLUnT0iy6P7CkiXNWAD2AZQ3KHwUmSrqcLGGOAI4v1oikLsBnbD8i6THgP8hmux/bxS1px/S85AWS9gN2TbH/SNJk2zWSPgXUlRh7CKED2GOPPZg7d+5HynK5HCNGjGDEiBFtFFUopr0k5ZHA5Q3K7ia7n/ytBuWTJdV/JGqi7dkb2PfNZEu7c5Rt715Kdn+7MTcC/yvpddtD6wttz5E0EXiqvm3bc9PMtZCuwG8lbUmWxCc08rGqsyUNJZv1Pg/8b7oPXQE8nnam15BdyDR2n30icH36He5nu7aJsYYQQmgm+nD1tn2p33Rle1Bbx9JZ7bLLLl64cGFbh9Fq2tN9qebQmcbbmcYKMd62Jml2sdzVXmbKH2N7FdnHd0IIIYQOod0m5XUh6Umy5ex8x6d7sO2KpJOAsxoUz7T97baIJ4QQQvPpFEnZ9uC2jqG52P418Ou2jiOEEELzazff6BVCCCF0dJGUQwghhDIRSTmEEEIoE5GUQwghhDIRSTmEEEIoE5GUQwghhDIRSTmEEEIoE5GUQwghhDIRSTmEENqR9957j3322Yc999yTAQMGcNFFFwFgmwsuuID+/ftTUVHBtdde28aRhvXRKb7RK4QQOopNNtmEadOm0b17d+rq6hgyZAiHHXYYf/3rX1m8eDEvvPACXbp04c0332zrUMN6aNGkLGkEcA9QYfuFvPKpwBhgMtAbqH/U4gTbNzZj/9XA+7ZnNVebqd3RwCDbp5dQdxvgfuATwJm2Z6xHXw/b/sd6hFpqH1XAr4AtyB7r+GPbdzR1Xm3dGvqd/0BLhVV2zqlczegYb4dUbmNddPmwosck0b17dwDq6uqoq6tDEr/61a+47bbb6NIlWwDddtttWyXW0Lxaevl6JPAYcGx9gaRuwFa2l6SiUbargP2Bn0r6RDP2Xw18vhnbQ9K6Xsh8EXjB9l7rmpCT0UCfdTlhPWL8F3CC7QHAvwNXS+q5jm2EEFrJmjVrqKqqYtttt+Xggw9m8ODBvPzyy9xxxx0MGjSIww47jL/97W9tHWZYDy02U5bUnSzRDgWmABenQ9VArsAp3YGVZDO1Ym3WANcAw8lm10fafiPNRq8H+qaqZwNLgNOANZKOI3uy0q+BHYEtgbeAatuPSpoBnJTKbgF2IEtUp9h+RtLFZImxH7AMeDgvpmHA94EjbC9rEG8VcAXQTdI8YD/gAOCHZCsDLwMn2a6RdCFwBNANmAWcChxN9njKyZJq0/l/JZulL5M0CLjKdnXDGCUdD1yeft+bAL+wfUOh36vtF/Ne/0PSm8A2wDsF/gxOAU4B6NVrGy6sXF2oyQ5pu27ZjKqz6EzjLbex5nK5JutcffXV1NTU8IMf/IBdd92Vf/3rXyxZsoSrrrqKRx99lKOPPrrofeWampqS+ugo2tN4W3L5+ihgqu0XJb0laW/bc4DDgPvy6k2WtArYGTjbdtGkDGwOPGH7AklXAN8ELiVL1BNsPyapL/CQ7QpJ1wM1tq8CkPQisBvwWWA2cEB6rOOnbb8k6efAXNtHSfoCcCtQlfoeCAyxXZuWlOuX58cCh9t+u2GwtuelZDvI9umSepEl8C/ZXinpu+n8S4DrbF+S2p0EDLd9l6TTgXG2n07HGvud58d4CvCu7c9J2gSYKelh26801oCkfciW2l8udDzdXrgRoO8OO3n8gs6zLeGcytXEeDumchvrolHVJdedPXs2//znP9l+++0577zz6NevHwcddBDjx4+nurpwO7lcruixjqg9jbcll69HAren17en95DNnh/LqzfK9h5ks9xxkrZvpM33ye7PQpZU+6XXXwKuS7PRKcAWknoUOH8GcGD6uQwYAnwO+Es6PgSYBGB7GrC1pC3TsSm2a/PaGgp8FxhWKCEXsS/ZRcHMFOuJQP14h0p6UtIC4AvAgBLbzJcf4yHACamfJ4GtyS58ipLUm2z8J9leux79hxBa2NKlS3nnnWwRq7a2lj/96U/suuuuHHXUUUybNg2A6dOn079//7YMM6yndb40lPRvwGdsP9NIna3JEsvukgx0BZxmrottv9/wHNtLJc0BBgOvFmm6zrbT6zV58XcB9muQNAvNKmeQLWn3AS4EziVb3n20/pQCfdb3t7JB+d/Jlrn7A08XibchAX+0PfIjhdKmwC/JZtSL01L0pkXaWM2HF1MN6+THKOAM2w+VFJi0BfAA8H3bT5RyTreNu7KwkQ0pHU0ul1unGUx715nG257G+vrrr3PiiSeyZs0a1q5dy9e+9jWGDx/OkCFDGDVqFBMmTKB79+7cfPPNbR1qWA8lJWVJOeDLqf48YKmk6bbHFjnlGOBW26fmtTEdOA6YWqSPzYC9yO7BrquHgdOBK1NbVbbnASvIdhTXe5JsSfrvtt9Ls8hTye5RQ5acRwE/Sju3l9leXmTJ+FVgHHCvpK/afq6EOJ8AfiFpp7RcvhnwaaD+swvL0r34Y4C7UtkKIH/Wv4hsmfp/ye45F/MQ8C1J02zXSeoPLLHd8OKCtLnuXrI/sztLGEcIoY3ssccezJ0792PlPXv25IEHymcHeVg/pS5fb2l7OfAV4Ne2B5ItGRczkuwf+Xx3ky3fNkzKk1NynA1MtD27xJjynQkMkvSMpOfJZsMAfwBGSJon6QDbq4DFZMkRsplzD2BBen9xfTtkm6RObKxT2wvJkvidknZsKkjbS8l2U/8u9fEEsKvtd4CbUhz38eFyOsBE4Po0hm5km8SuSZvTGrv/fjPwPDBH0rPADRS/CPsa2ZL+6NTPvLRJLYQQQivSh6vBjVTK7nMeAvwGuMD2XyQ9k+4Fl9ZR2kzOh9sAACAASURBVGxke9B6RxvKyi677OKFCxe2dRitpj1tFmkOnWm8nWmsEONta5JmF8uFpc6ULyFbDn05JeQdgHX6EJztVZGQQwghhOJKuqec7jPemff+7zR+P3ODpI8pbdKg+HjbCwrVLweSLgC+2qD4Tts/bot4CpFUSdpdnmeV7cFtEU8IIYSPKnWjV3+yr2HczvbukvYAvmz70pYIqj0miZR8yyYBF5IuauJecQghlKlSl69vAv4LqANIH4c6ttEzQgghhLBOSk3Km9l+qkFZ+XwnXQghhNABlJqUl6WP/BhA0jHA6y0WVQghhNAJlfqNXt8m+77jXSUtAV4h+3xuCCGEEJpJk0lZUheyr3/8kqTNgS62V7R8aCGEEELn0uTydXowwenp9cpIyCGEEELLKPWe8h8ljZP0GUlb1f+0aGQhhBBCJ1PqPeWT03+/nVdmsqckhRBCCKEZlDRTtv3ZAj+RkEMIHcbixYsZOnQoFRUVDBgwgGuuuQaA+fPns99++1FZWckRRxzB8uXL2zjS0JGV+o1eJxQqt31r84YTQghtY6ONNmL8+PHsvfferFixgoEDB3LwwQfzjW98g6uuuoqDDjqIW265hSuvvJIf/ehHbR1u6KBKXb7+XN7rTYEvAnPInk3cIiSNAO4BKmy/kFc+FRgDTAZ6A7Vk35M9wfaNjbT3VbIHa/yf7aFF6lQD42wPlzSabNf56Rswhi8Du9m+fH3baG2SjgJetP18U3Vr69bQ7/zO8/zWcypXMzrG2+4tunxYwfLevXvTu3dvAHr06EFFRQVLlixh4cKFHHjggQAcfPDBHHrooZGUQ4spdfn6jLyfbwJ7AZ9o2dAYCTxG3td5pucJb2V7SSoaZbsK2B/4qaTGYhoD/GexhNwSbE9pTwk5OQrYra2DCKEtLVq0iLlz5zJ48GB23313pkyZAsCdd97J4sWL2zi60JGVOlNu6F/Azs0ZSD5J3ckS7VBgCnBxOlQN5Aqc0h1YCawp0t6FwBDgs5KmAD8ge8DGILKvCx1r+5FG4tkeuAXYBlgKnAQsIXt85Y7AlsBbQLXtRyXNSHWGkGbbkiYCy1OfnwTOs31X+hz4dcBBZF/K0gW4xfZdRWL5HHANsDmwimzVoq7QeBrO9iXdD1xlOyepJrUznGy14cg0li8DB0n6PnC07Zcb9H8KcApAr17bcGFl5/m21e26ZbPHzqKjjjeXy32srKam5oPy2tpazjrrLL7xjW8wZ84cTjvtNC699FLOPfdc9t9/f7p06VKwjfYkf7ydQXsab6n3lP9A+opNsqSxG3mPcmwBRwFTbb8o6S1Je9ueAxwG3JdXb7KkVWQXCGfbLpiUbV8i6QtkS9NPSzonlVdK2hV4OD0Jq5jrgFtt/0bSycC1to+S9CLZ7+KzwGzggPTYyU/bfknSkAbt9CZL1LuSXWzcBXwF6AdUAtsCfyW7APiYtBJwB/D19FzrLcgS6lnrOB7IkvoTti+QdAXwTduXpouW+4tdFKRbBDcC9N1hJ49fsL7Xde3POZWrifG2f4tGVX+sLJfLUV1dTV1dHcOHD+e0005j7NixHxw/4YRsW82LL77Ic889R3X1x9toT+rH21m0p/GW+jfuqrzXq4FXbb/WAvHUGwlcnV7fnt7PIZs9j8urNyol2W2AWZKm2n61hPaHAD8HsP2CpFeBxpLYfmTJE7LnEV+RXs8ADiRLypcB3wSmA38p0s596ctYnpe0XV4sd6by/5NUdMYO7AK8bvsvKfblACn5r8t4AN4H7k+vZwMHN1H/Y7pt3JWFRe7PdUS5XK7gP+gdVWcbr23GjBlDRUXFRxLym2++ybbbbsvatWu59NJLOe2009owytDRlfrlIYfbnp5+Ztp+TdJPWyIgSVsDXwBulrQIOBf4enogxmLb7zc8x/ZSsqRd6nOYtYFh1q8azAAOAPYBHgR6ki2xP1rkvFUFYliXWJTXd8PyQlbz0T/jTfNe19mub2sN638rI4QOYebMmUyaNIlp06ZRVVVFVVUVDz74IL/73e/o378/u+66K3369OGkk05q61BDB1bqP8QHA99tUHZYgbLmcAzZUvGp9QWSpgPHAVMLnSBpM7LNZ1cUOl7Ao2QP1JiWlnn7AgvJZsSFzCLbcDYpnfdYKn+SbAf6322/J2kecCrZfdpSPQacKOk3ZPesq4HbitR9Aegj6XNp+boH2fJ1sfFsAfxnum/9KbKLh6asAHqsQ/whdAhDhgzhw+vUjzrrrLNaOZrQWTU6U5b0LUkLgF0kPZP38wrwTAvFNBK4t0HZ3cC+fDwpT06JcDYw0fbsEvv4JdA1je0OYLTtVY3UPxM4SdIzwPF8eA93FbAYeCLVm0GW0BaUGAdkY3sNeBa4gSzRv1uoYlol+Drwc0nzgT+SzX6LjWcm2eaxBWS3IOaUEM/twLmS5qbViRBCCK2kqZnybcD/kt0vPT+vfIXtt1oiINvVBYpvAE6wvaiJeiW1a/s9YHSBOjnS7m7bE4GJ6fUisiX1Qu0ekPf6NvJmuQ3aGN3gvO7pv2sljbNdk5bun6KRpJ7uJ+9b4FCh8Zgij9is7z+9vots0xm2ZxIfiQohhDbRaFK2/S7ZrG0kgKRtyWZm3SV1t/3/Wj7ED2akg1qjrzZyv6SeZJ/9/pHt/2vrgEIIIbS+Uj8SdQTwM6AP8CawPdlHdwa0XGjrJ30kaZMGxcfbXpcl5VZVaNYv6V6yXd35vmv7oVYJKoQQQqsrdaPXpWRLpn+yvZekoaTZc7mxXeoO7LJme0RbxxBCCKF1lfqRqDrb/wS6SOqSvv2qqgXjCiGEEDqdUmfK76SvvpxBtuP5TbLPwIYQQgihmZQ6Uz6S7Puuzyb7WNLLwBEtFVQIIYTQGZU0U7a9Mj2UYef0/c+bAV1bNrQQQgihcylppizpm2SfY70hFX2Kjz4YIoQQQggbqNTl62+TPQxiOYDtv5E90SiEEEIIzaTUpLwq/0EQkjai8IMRQgghhLCeSk3K0yV9D+gm6WCyZyn/oeXCCiGEEDqfUpPy+cBSsu9kPpXsMYXfb6mgQgihpSxevJihQ4dSUVHBgAEDuOaaawCYN28e++67L1VVVQwaNIinnnqqjSMNnVGju68l9bX9/2yvBW5KPyGE0G5ttNFGjB8/nr333psVK1YwcOBAvve973Hbbbdx0UUXcdhhh/Hggw9y3nnnkcvl2jrc0Mk09ZGo+4C9ASTdbfvolg+peUkaAdwDVNh+Ia98KjAGmAz0Jnsu8SbABNs3NtLeImCQ7WWSZtn+fCN1v2f7J+sY7+jU/unrcl5zkVQF9LH9YFN1a+vW0O/8B1ohqvJwTuVqRsd4241Flw8rWN67d2969+4NQI8ePaioqGDZsmVIYvny5QC8++679OnTp9ViDaFeU0lZea93aMlAWtBI4DHgWOBiAEndgK1sL5EEMMr205K2Al6WNDF/Y1sxjSXk5HvAOiXlMlBF9kSuJpNyCO3dokWLmDt3LqeccgrDhg3j0EMPZdy4caxdu5ZZs2a1dXihE2oqKbvI63YhfTXo/sBQYAopKQPVpOcmN9AdWAmsKbH9GtvdJfUG7gC2IPudfgsYRrYxbh7wnO1Rko4DziR7ROOTwH/aXiPpJOC/gNeBF4FVjfS5HXA9H14kfcv2LEljgZNT2c22r5bUD7jf9u7p3HFAd9sXS8qlGIYCPclWDZ4ELklxDwEus31Hg/5PAU4B6NVrGy6s7Dzftrpdt2z22Fm09/E2tfRcW1vLWWedxTe+8Q1sc8EFFzBmzBgOOuggHnnkEb7yla8wfvz41gm2ldXU1HSqpfn2NF7ZxXOtpDVkSUpAN7Kv2iS9t+0tWjzCDZCS4FDbYyTNAk63PUfStcB9tqel5NSbLBHuDJxt+4ZG2lzEh8vX9Un5HGBT2z+W1BXYzPaK+uPpvArgCuArtusk/RJ4AvgjWTIcSPbs6keAucWWryXdATyekm5XsguJnYCJZE/yUmrvOOBtGk/Ks22fI+lwYKztL63L8nnfHXZyl69d01S1DuOcytWMX1Dq18W3f+19vMWWrwHq6uoYPnw4hx56KGPHjiWXy3HkkUfyzjvvIAnbbLnllh8sZ3c0uVyO6urqtg6j1ZTbeCXNtj2o0LFG/8bZbu9fpTkSuDq9vj29n0M2ex6XV69++XobYJakqbZfXYd+/gLcImljsmQ/r0CdL5Il3r+kJfNuZM+mHgzkbC+FD5Ju/0b6+gJwAoDtNcC7aVZ7r+2VqY17gAPIVgcac0/672ygXxN1P6bbxl1Z2Mg/fB1NLpdj0ajqtg6j1XTU8dpmzJgxVFRUMHbs2A/K+/Tpw/Tp06murmbatGnsvPPObRhl6Kza72VwEyRtTZbAdpdksu/qtqTrgcWF7hnbXippDlmiLDkp235U0oFkS9aTJF1p+9aGIQG/sf1fDeI8ig2/NaAi5av56MfeNm1wvH6ZfA0d+P+FEPLNnDmTSZMmUVlZSVVV9gTaY489lptuuomzzjqL1atXs+mmm3LjjUX3e4bQYjryP8THALfaPrW+QNJ0smXdqYVOSA/a2Itsmblk6WEdS2zfJGlzsh3rtwJ1kja2XQf8Gfi9pAm230ybynqQLTVfky4ilgNfBeY30t2fye5Z1y9fbw48CkyUdDlZgh4BHA+8AWyb2q4Bhhcbe54VKa4QOqQhQ4bQ8LZdLpdjyJAhzJ49u42iCiFT6peHtEcjgXsblN1Ndt+1YWKanDZkzQYm2l7Xv5nVwDxJc4GjgfobrTcCz0iabPt5si9ceVjSM2T3knvbfp1sA9rjwJ/IltcbcxYwVNKCFO8A23PI7ik/RZbkb7Y9N10MXJLK7gdeKNzkRzwC7CZpnqSvlzL4EEIIzaPDzpRtVxcovgE4wfaiJuo11m6/vNfd039/A/ymQN3vAt/Ne38H2S7thvV+Dfy6xP7fIHu+dcPynwE/K1B+LXBtgfLqvNfLSPeUbb8FfK6UWEIIITSvDpuUC7G9iuwzuCGEEELZ6VRJeV1IepLsG77yHW97QSv1fwHZ/eV8d9r+cWv0H0IIofVFUi7C9uA27v/HQCTgEELoRDryRq8QQgihXYmkHEIIIZSJSMohhBBCmYikHEIIIZSJSMohhBBCmYikHEIIIZSJSMohhBBCmYikHEJoFxYvXszQoUOpqKhgwIABXHNN9hXzF198MZ/61KeoqqqiqqqKBx98sI0jDWH9xZeHhBDahY022ojx48ez9957s2LFCgYOHMjBBx8MwHe+8x3GjRvXRAshlL9WScqSRgD3ABW2X8grnwqMASYDvYFasq+2nGB7gx9mKmk08LDtf2xoW0Xarwbetz2rJdpfV5K+DOxm+/JmaKum/oEbxdTWraHf+Q9saFftxjmVqxkd421xiy4fVrC8d+/e9O7dG4AePXpQUVHBkiVLWjO0EFpcay1fjwQeA46tL5DUDdjKdv3fqlG2q4D9gZ9K+kQpDadnChczGuizXhGXphr4/IY0IKnZLoxsT2mOhBxCuVu0aBFz585l8ODs23Cvu+469thjD04++WTefvvtNo4uhPXX4klZUneyRDuGvKRMltByBU7pDqwE1jTSZo2kS9JDI/aTNFDSdEmzJT0kqbekY8ieCDU5PRv4IEn3pPOPlFQr6ROSNpX091S+o6SpqZ0ZknZN5dtIulvSX9LP/pL6AacB30ntHyDpq5KelTRf0qONxD9a0p2S/gA8nMrOTW0/I+mHeXVPSGXzJU0qFk9eu9dJ2lLSIkldUvlmkhZL2riRMX5W0uOpvR8Viz2EtlZTU8PRRx/N1VdfzRZbbMG3vvUtXn75ZebNm0fv3r0555xz2jrEENZbayxfHwVMtf2ipLck7W17DnAYcF9evcmSVgE7A2fbLpqUgc2BZ21fKGljYDpwpO2lkr4O/Nj2yZJOB8bZfjrNSCem8w8AniV7bvBGwJOp/EbgNNt/kzQY+CXwBeAasiX1xyT1BR6yXSHpeqDG9lUAkhYAh9peIqlnE7+X/YA9bL8l6ZA07n0AAVMkHQj8E7gA2N/2MklbpXM/Fg9QUd+w7XclzQcOAh4Bjkgx10lqbIy/sn2rpG8XC1rSKcApAL16bcOFlaubGGbHsV23bEm3s2ir8eZyuaLHVq9ezX/9138xePBgttpqq4/Vrays5Lbbbmu0jUJqamrW+Zz2LMZbvlojKY8Erk6vb0/v55DNnvN3ZoxKyXMbYJakqbZfLdLmGuDu9HoXYHfgj5IAugKvNzzB9mpJL0mqIEt+PwMOTPVnpBn954E7Uzvw4aMbvwTslle+haQeBeKaCUyU9D9k99Ab80fbb6XXh6Sfuel9d7IkvSdwl+1laQz19UuJ5w7g62RJ+Vjgl02McX/g6PR6EvDTQkGne/03AvTdYSePX9B59gqeU7maGG/LWzSqumC5bU488UT2339/rr766g/KX3/99Q/uNU+YMIHBgwdTXV24jWJyudw6n9OexXjLV4v+jZO0NdksbHdJJkuATjPMxbbfb3hOmu3OAQYDxZLye3kzaQHP2d6vhJBmkM3Q64A/kc2cu5JdHHQB3kn3tRvqAuxnu7bB+BrGflqafQ4D5kmqsv3PIrGszG8KuMz2DQ3aPxPwesYzBbgsza4HAtPIVhiKjZEifRXVbeOuLCyyKacjyuVyRRNGR1Ru4505cyaTJk2isrKSqqrsf+Gf/OQn/O53v2PevHlIol+/ftxwww1NtBRC+Wrpy+BjgFttn1pfIGk6cBwwtdAJkjYD9gKuKLGPhcA2kvaz/Xhazu5v+zlgBZA/g3wUuDXFtDRdNHySLKlb0iuSvmr7TmUZbg/b88nu+54OXJlirLI9L7W/RV7sO9p+EnhS0hHAZ8iWoJvyEPAjSZNt10j6FNmFw5+BeyVNsP1PSVul2XKxeD6Q2nmKbFn6/nQRs7yRMc4km1H/FhhVyi8+hNY0ZMgQ7I9fNx5++OFtEE0ILaOlN3qNBO5tUHY3sC8fT8qTJc0DZgMTbc8upYM02z6GbMf2fGAeH+6InghcnzZidSO7d7wdWXIGeAZ4xh/+TR8FjEntPAccmcrPBAalDVfPk23wAvgDMKJ+oxdwpaQFkp5NfcwvcQwPA7cBj6f70ncBPdKFxY+B6SmmnzURT0N3kF0A3ZFXVmyMZwHflvQXYMtS4g4hhNC8VOjKs0U7lDYBZtoe1Kodh2a3yy67eOHChW0dRqtpT/elmkNnGm9nGivEeNuapNnFcmCr7+KwvYrso0ohhBBCyFPWW0mVfQ55kwbFx9te0BbxrCtJh/LxXcyv2B7RFvGEEEIob2WdlG0PbusYNoTth8g2cYUQQghNiqdEhRBCCGUiknIIIYRQJiIphxBCCGUiknIIIYRQJiIphxBCCGUiknIIIYRQJiIphxBCCGUiknIIIYRQJiIphxDKyuLFixk6dCgVFRUMGDCAa6655iPHr7rqKiSxbNmyNoowhJZT1t/oFULofDbaaCPGjx/P3nvvzYoVKxg4cCAHH3wwu+22G4sXL+aPf/wjffv2beswQ2gRkZRbiaQRwD1Ahe0X8sqnAmOAyUBvoJbs+74n2L6xhWOqBt63PWt9zq+tW0O/8x9o3qDK2DmVqxkd4202iy4fVrC8d+/e9O7dG4AePXpQUVHBkiVL2G233fjOd77DFVdcwZFHHlnw3BDau1i+bj0jgceAY+sL0jOet7K9JBWNsl0F7E/2fOhPtHBM1Xz47OkQys6iRYuYO3cugwcPZsqUKXzqU59izz33bOuwQmgxMVNuBZK6kyXaocAU4OJ0qBrIFTilO7ASWNNImzXADanNt4FjbS+VVAVcD2wGvAycbPttSWcCpwGrgeeB89P7NZKOA84APglclPp91/aBBfo9BTgFoFevbbiwcnWpv4Z2b7tu2eyxs2jp8eZyuUaP19bWctZZZ/GNb3yDWbNm8d3vfpcrr7ySXC7He++9x8yZM9lyyy2bJZaampom4+lIYrzlS7bbOoYOLyW9obbHSJoFnG57jqRrgftsT5OUI1u+XgXsDJxt+4ZG2jRwnO3Jki4EtrV9uqRngDNsT5d0CbCF7bMl/QP4rO1VknrafkfSxUCN7atSmwuAf7e9pL5OY+Pqu8NO7vK1axqr0qGcU7ma8Qs6z3VsS4+32PI1QF1dHcOHD+fQQw9l7NixLFiwgC9+8YtsttlmALz22mv06dOHp556ik9+8pMbHEsul6O6unqD22kvYrxtS9Js24MKHes8/8K0rZHA1en17en9HLLZ87i8eqNsPy1pG2CWpKm2Xy3S5lrgjvT6t8A9krYEetqensp/A9yZXj8DTJZ0H3BfkTZnAhMl/Q/Z/e9Gddu4Kwsb+Ye1o8nlciwaVd3WYbSathqvbcaMGUNFRQVjx44FoLKykjfffPODOv369ePpp5+mV69erR5fCC0p7im3MElbA18Abpa0CDgX+LqkHYHFtt9veI7tpWRJe12eJ93Ukscw4BfAQGC2pI9dkNk+Dfg+8BlgXoo9hFY1c+ZMJk2axLRp06iqqqKqqooHH3ywrcMKoVXETLnlHQPcavvU+gJJ04HjgKmFTpC0GbAXcEUj7XZJbd8O/AfwmO13Jb0t6QDbM4DjgemSugCfsf2IpMdS/e7ACmCLvH53tP0k8KSkI8iS8z/Xd+AhrI8hQ4bQ1G21RYsWtU4wIbSySMotbyRweYOyu4HDgG81KJ8sqf4jURNtz26k3ZXAAEmzgXeBr6fyE4HrU2L/O3AS0BX4bVreFtnHrd6R9AfgLklHkm30+o6knVOdPwPz12vEIYQQ1ksk5RZmu7pA8Q3ACbYXNVGvqbZ/APygQdk8YN8C1YcUOP9FYI+8ohnrGkMIIYTmE0m5DdheBRTceRdCCKHziqRc5iQ9Sbacne94293bIp4QQggtJ5JymbO9LjuwQwghtGPxkagQQgihTERSDiGEEMpEJOUQQgihTERSDiGEEMpEJOUQQgihTERSDiGEEMpEJOUQQgihTERSDiGEEMpEJOUQAieffDLbbrstu++++wdlL730Evvuuy9VVVUMGjSIp556qg0jDKFziKQcQmD06NFMnfrRJ4necMMNXHTRRcybN49LLrmE8847r42iC6HzaLdfsylpBHAPUGH7hbzyqcAYYDLQG6h/FOIE2zc2Q7+jgYdt/2ND2yrSfjXwvu1ZLdF+Cf0fBbxo+/mm6tbWraHf+Q+0QlTl4ZzK1Yxux+NddPmwoscOPPDAgs8oXr58OfD/27v3IK3q+47j74/rjXQNFiQmBo2XkFabWNKkXqqRlVwkkQoSU0OxSae0xrakOsPSSdpMI51mHGPMrZOMMalBOl6Dlxj/CFBhwaihCq6AGqoG2iRqkLYgEKqCn/5xfht21meXi8s+l/28Zpg953fOnuf7hcPzfX6/c57zgy1btnDMMcccqNAiomjaokw1T/GPgI8DVwJIGgGMsv0LSQAzbD8iaRTwjKR5tl/e04Eltdne1c/mPwXWAgekKAMdwDagLkUZmArcC+yxKEdrmzVrFnPmzKGzs5NXX32VBx+s1ykZMXw0ZVGW1A6cBZwL3EMpylQFravGr7QD24H+Ci2StgFfBs4DZkvaUdbbgU1UxfgsqikXbyrbLwcutz1N0hTgVmAk1WWBJ2yfKOkk4BvAGOBXwF/Y/omkMcB1wHElhCuAXwCXAbskXQJ8Gngz8PkS+xbb5/QT/+8A3wUOLa//UdtPleP8TWlfAfyV7V0l368Bk6lGE6YAJwEXABMkfa4c45k+r3MpcCnAUUeN4R/etbO/v9KWc/SIqrfcrLq6ugbc/vzzz7N9+/Zf77dgwQJmzpzJhAkTWLp0KdOmTePaa6898IHWwbZt2/b499NKkm/jku16x7DPSqE51/ZMSQ8Cs2yvkvR14G7bSyR1UQ1fvwSMA66w/a0BjmngYtu3SzoEWAZMsf2CpIuB82z/WTluZ+mBHww8ZfsESV8CJlAV14OBy2xPl3RfWX5K0unAVbYnSroZ+KbtH0k6Dlho+2RJVwLbbH+pxLUGmFR6/0fa3txP/P8M/Nj2TZIOBdqA44EvAtNsvyLpm2Wf+SXfC2z/QNIXgRdt/5OkecC9thfs6d/huBPf7oP+6Gt72q1lzH7XTq5d05SfY4GBh68BNmzYwOTJk1m7di0A7e3tbN26FUnYZuTIkb8ezm41XV1ddHR01DuMIZN860vSStvvrbWtWd9hpgNfLcu3lvVVVD3Zzl779QxfjwEelPRD2//ZzzF3AXeU5d8C3gksLsPgbcBzfX/B9k5JT0s6GTiNqmd9Ttn//tKj/wPge+U4sHtu5A8Ap/Rqf6OkI2rE9QAwT9LtVNfQ+/MQ8PeSxgJ3lg8B7wfeAzxcXmcEsLHs/zLVMDXASuCDAxy7phGHtLFuD2/0raSrq4sNMzrqHcaQGT16NMuWLaOjo4MlS5Ywbty4eocU0fKarihLGg1MBN5ZenttgCVdB/ys1jXj0ttdBZwO9FeU/6/XdWQBj9s+cy9Cuh/4MPAK8G/AvBJTJ9Uw8mbb42v83kHAmbZ39Mmvb+yXlR72+UC3pPG2/7tGjjdLWlH2Wyjpz0seN9r+bI3Xf8W7h0l20YTnQgye6dOn09XVxaZNmxg7dixz586ls7OT2bNns3PnTg4//HCuv/513ycZEXvQjG/EFwHzbX+qp0HSMuAS4Ie1fkHSG4B3Uw3l7o11wBhJZ9p+qAxnv8P248BWoHePdjkwv8T0QvnQ8Gaqom5J6yV9zPb3VFXcU20/BiwCZgHXlBjH2+4ux39jr9hPsr0CWCHpD4FjgdcUZUknAj+1/fWyfGp5je9L+ortjeWGtyMGGC2gRn4xDNxyyy2vaevq6mLlypV1iCZi+GrG7ylPB+7q03YHcAavLco3SeqmGp6dZ3uv3mFKb/si4GpJjwHdVMPQUPWEr5PUXe72XgEcTVWcvUk/HgAABppJREFUAVYDq3v1QmcAM8txHqe6oQqqm6/eK2m1pCeobvAC+AFwYTn++4BrJK2RtLa8xmP9hH0xsLbk+9tUHxKeAD4HLJK0GlhMdZ19ILcCcyQ9Wm5Si4iIIdJ0PWXbHTWavwV8wvaGPew30HHb+6x3U10f7rvfHey+9tzjsF7bL+2z/3pgUo3jbKIqpH3b/4Oql9vj/r0IH9tXAVfVaL8NuK1Ge3uv5QXAgrL8AHDK3rxmREQMrqYryrXYfonqq0oRERFNqyWK8r4oN0Md1qf5T2yvqUc8+0rSecDVfZrX276wHvFERMTgGXZF2fbp9Y7h9bC9EFhY7zgiImLwNeONXhERES0pRTkiIqJBpChHREQ0iBTliIiIBpGiHBER0SBSlCMiIhpEinJERESDSFGOiIhoECnKERERDSJFOSIiokGkKEdERDSIFOWIiIgGIdv1jiGalKStwLp6xzGEjgI21TuIITSc8h1OuULyrbe32R5Ta8OwmyUqBtU628NmHmtJjyTf1jSccoXk28gyfB0REdEgUpQjIiIaRIpyvB7X1zuAIZZ8W9dwyhWSb8PKjV4RERENIj3liIiIBpGiHBER0SBSlGO/SJokaZ2kpyV9pt7xDDZJN0jaKGltr7ZRkhZLeqr8/M16xjhYJB0raamkJyU9Luny0t6q+R4u6d8lPVbynVvaT5C0ouR7m6RD6x3rYJHUJulRSfeW9VbOdYOkNZK6JT1S2prmXE5Rjn0mqQ34BvBh4BRguqRT6hvVoJsHTOrT9hngPtvjgPvKeivYCcy2fTJwBvDX5d+zVfN9CZho+3eB8cAkSWcAVwNfKfn+LzCzjjEOtsuBJ3utt3KuAOfaHt/ru8lNcy6nKMf+OA142vZPbb8M3ApMqXNMg8r2cuB/+jRPAW4syzcCU4c0qAPE9nO2V5XlrVRv3m+ldfO17W1l9ZDyx8BEYEFpb5l8JY0Fzge+U9ZFi+Y6gKY5l1OUY3+8FfhZr/Wfl7ZWd7Tt56AqZMCb6hzPoJN0PPBuYAUtnG8Zzu0GNgKLgWeAzbZ3ll1a6Zz+KvC3wKtlfTStmytUH7AWSVop6dLS1jTnch6zGftDNdry3bomJ6kduAO4wvaLVYeqNdneBYyXdCRwF3Byrd2GNqrBJ2kysNH2SkkdPc01dm36XHs5y/azkt4ELJb0k3oHtC/SU4798XPg2F7rY4Fn6xTLUPqlpLcAlJ8b6xzPoJF0CFVBvsn2naW5ZfPtYXsz0EV1Lf1IST0dlVY5p88CLpC0geoy00SqnnMr5gqA7WfLz41UH7hOo4nO5RTl2B8PA+PKHZyHAh8H7qlzTEPhHuCTZfmTwPfrGMugKdcY/wV40vaXe21q1XzHlB4ykkYAH6C6jr4UuKjs1hL52v6s7bG2j6f6f7rE9gxaMFcASb8h6YieZeBDwFqa6FzOE71iv0j6CNUn7jbgBttfqHNIg0rSLUAH1ZRvvwQ+D9wN3A4cB/wX8DHbfW8GazqSzgbuB9aw+7rj31FdV27FfE+lutmnjapjcrvtf5R0IlVvchTwKHCJ7ZfqF+ngKsPXnbYnt2quJa+7yurBwM22vyBpNE1yLqcoR0RENIgMX0dERDSIFOWIiIgGkaIcERHRIFKUIyIiGkSKckRERIPIE70iouFI2kX1Fa0eU21vqFM4EUMmX4mKiIYjaZvt9iF8vYN7PQs6om4yfB0RTUfSWyQtL3PmrpX0vtI+SdKqMlfyfaVtlKS7Ja2W9OPy8BAkXSnpekmLgPllkoprJD1c9v1UHVOMYSrD1xHRiEaUWZwA1tu+sM/2PwYWlqc1tQFvkDQG+DZwju31kkaVfecCj9qeKmkiMJ9qHmWA9wBn295RZhTaYvv3JR0GPCBpke31BzLRiN5SlCOiEe2wPX6A7Q8DN5SJNO623V0eI7m8p4j2eozi2cBHS9sSSaMljSzb7rG9oyx/CDhVUs8zoUcC44AU5RgyKcoR0XRsL5d0DnA+8K+SrgE2U3sKwoGmKtzeZ79P2144qMFG7INcU46IpiPpbVTzBH+baoar3wMeAiZIOqHs0zN8vRyYUdo6gE22X6xx2IXAX5beN5LeUWYaihgy6SlHRDPqAOZIegXYBnzC9gvluvCdkg6imjP3g8CVwHclrQZ+xe4p/Pr6DnA8sKpMZ/kCMPVAJhHRV74SFRER0SAyfB0REdEgUpQjIiIaRIpyREREg0hRjoiIaBApyhEREQ0iRTkiIqJBpChHREQ0iP8H9hvyL4+gQv8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculating training and testing of our model\n",
    "\n",
    "print(\"Training accuracy of XGBoost model:\", model_xgb.score(x_train, y_train))\n",
    "print(\"Testing accuracy of XGBoost model:\", model_xgb.score(x_test, y_test))\n",
    "\n",
    "\n",
    "plot_importance(model_xgb, max_num_features=9)\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeUAAAEWCAYAAABYNo/VAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3de5wU1Z3+8c8DeAFRCEJcxIiK91uIYNB42cGoiZdEMSiyxFvcINlodON1k6wxrP6i0RjdNVnRbEANEYNERTdBCTqKNyKDKKAOiYJBzCp4BwwCfn9/1Blomp6ZbpyZLuB5v179mqrTp6qe7hn49jlV3a2IwMzMzKqvXbUDmJmZWcZF2czMLCdclM3MzHLCRdnMzCwnXJTNzMxywkXZzMwsJ1yUzWyDI+lmSf9e7RxmLU1+n7LZpkPSfGA7YFVB8+4R8fon2GcN8OuI2OGTpdswSRoDvBYRP6h2FtvweaRstun5SkR0Lritd0FuCZI6VPP4n4Sk9tXOYBsXF2UzA0DSQZKelPSupOfSCLjhvrMkvSjpA0mvSDontW8F/AHYXtKSdNte0hhJVxZsXyPptYL1+ZIulfQ8sFRSh7TdBEmLJM2T9J0msq7ef8O+JV0i6U1Jf5N0oqRjJc2V9Lak7xVse4WkuyXdlR7PDEmfLbh/L0m16XmYI+mrRcf9b0m/l7QUOBsYBlySHvv9qd9lkl5O+39B0qCCfZwp6XFJ10l6Jz3WYwru7yZptKTX0/33Ftx3vKSZKduTkvYv+xdsGwQXZTNDUi/gf4ErgW7ARcAEST1SlzeB44FtgLOAn0k6ICKWAscAr6/HyHsocBzQFfgYuB94DugFfBG4QNKXytzXPwBbpm0vB24Fvg70Aw4DLpe0S0H/E4Dx6bH+BrhX0maSNks5HgI+DZwHjJW0R8G2/wRcBWwN3A6MBX6SHvtXUp+X03G7AD8Cfi2pZ8E+BgD1QHfgJ8D/SFK67w6gE7BPyvAzAEkHAL8CzgG2BUYBEyVtUeZzZBsAF2WzTc+9aaT1bsEo7OvA7yPi9xHxcURMBqYDxwJExP9GxMuReZSsaB32CXP8Z0QsiIgPgQOBHhExMiI+iohXyArrqWXuawVwVUSsAMaRFbsbI+KDiJgDzAEKR5V1EXF36n89WUE/KN06A1enHA8DD5C9gGhwX0Q8kZ6nv5cKExHjI+L11Ocu4M/A5wu6vBoRt0bEKuA2oCewXSrcxwAjIuKdiFiRnm+AbwKjImJaRKyKiNuA5SmzbSQ22HM5ZrbeToyIPxa19QZOlvSVgrbNgEcA0vTqD4HdyV7MdwJmfcIcC4qOv72kdwva2gNTy9zXW6nAAXyYfr5RcP+HZMV2nWNHxMdpan37hvsi4uOCvq+SjcBL5S5J0unAd4GdUlNnshcKDf6v4PjL0iC5M9nI/e2IeKfEbnsDZ0g6r6Bt84LcthFwUTYzyArNHRHxzeI70vToBOB0slHiijTCbphuLfUWjqVkhbvBP5ToU7jdAmBeROy2PuHXw2caFiS1A3YAGqbdPyOpXUFh3hGYW7Bt8eNda11Sb7JR/heBpyJilaSZrHm+mrIA6Capa0S8W+K+qyLiqjL2YxsoT1+bGcCvga9I+pKk9pK2TBdQ7UA2GtsCWASsTKPmowu2fQPYVlKXgraZwLHpoqV/AC5o5vh/At5PF391TBn2lXRgiz3CtfWTdFK68vsCsmngp4FpZC8oLknnmGuAr5BNiTfmDaDwfPVWZIV6EWQXyQH7lhMqIv5GduHcLyR9KmU4PN19KzBC0gBltpJ0nKSty3zMtgFwUTYzImIB2cVP3yMrJguAi4F2EfEB8B3gt8A7ZBc6TSzY9iXgTuCVdJ56e7KLlZ4D5pOdf76rmeOvIit+fYF5wGLgl2QXSrWG+4AhZI/nNOCkdP72I+CrZOd1FwO/AE5Pj7Ex/wPs3XCOPiJeAH4KPEVWsPcDnqgg22lk58hfIrvA7gKAiJhOdl75ppT7L8CZFezXNgD+8BAz26RIugLYNSK+Xu0sZsU8UjYzM8sJF2UzM7Oc8PS1mZlZTnikbGZmlhN+n7Ktt65du8auu+5a7RjrWLp0KVtttVW1Y5SU12zOVRnnqoxzra2urm5xRPQodZ+Lsq237bbbjunTp1c7xjpqa2upqampdoyS8prNuSrjXJVxrrVJerWx+zx9bWZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlvzsZz9jn332Yd9992Xo0KH8/e9/Z968eQwYMIDddtuNIUOG8NFHH7Xa8RURrbZz27jtuMuu0e6UG6sdYx0X7reSn87qUO0YJeU1m3NVxrkqk9dcY768FTU1NavXFy5cyKGHHsoLL7xAx44dOeWUUzj22GP5/e9/z0knncSpp57KiBEj+OxnP8u3vvWt9T6upLqI6F/qvo1+pCxpkKSQtGdR+yRJvSTVSqqXNFPSi5KGN7O/+ZK6p+Unm+n7vfXIe6akmyrdrqVI6ivp2God38ysmlauXMmHH37IypUrWbZsGT179uThhx9m8ODBAJxxxhnce++9rXb8jb4oA0OBx4FTGxokdQS6RcTC1DQsIvoChwDXSNq8nB1HxBea6VJxUc6BvoCLspltcnr16sVFF13EjjvuSM+ePenSpQv9+vWja9eudOiQjfR32GEHFi5c2Mye1t9GXZQldSYrtGdTUJSBGqC2xCadgaXAqjL3vyT97CnpsTTani3pMElXAx1T29jU7+uS/pTaRklqn9rPkjRX0qMpb1PH3E7SPZKeS7cvpPbvpmPPlnRBattJ0uyCbS+SdEVarpV0TcozN2XeHBgJDEkZh5TzPJiZbQzeeecd7rvvPubNm8frr7/O0qVL+cMf/rBOP0mtliF/k/wt60RgUkTMlfS2pAMiYgZwDFA4/zBW0nJgN+CCiCirKBf4J+DBiLgqFdpOETFV0rlpBI6kvYAhwCERsULSL4BhkiYDPwL6Ae8BjwDPNnGs/wQejYhB6VidJfUDzgIGAAKmpQL/TjO5O0TE59N09Q8j4khJlwP9I+LcUhuk6f3hAN279+Dy/VY2/+y0se06Zuew8iiv2ZyrMs5VmbzmWrJkCbW1tavXa2tr2XLLLZkzZw4Ae+21F3fffTeLFi1iypQptG/fnjlz5rDllluutV1L2tiL8lDghrQ8Lq3PIBuNXlTQb1hETJfUA3hS0qSIeLWC4zwD/ErSZsC9ETGzRJ8vkhXeZ9KrrI7Am2SFtDYiFgFIugvYvYljHQGcDpBePLwn6VDgnohYmvbxO+AwYGIzuX+XftYBOzXTl3TMW4BbILvQK48Xb+T1ohLIbzbnqoxzVSavuYov9OrYsSPjx4/n85//PB07dmT06NEceeSRbL755ixatIhTTz2VcePGcdZZZ621XUvaaKevJW1LVsB+KWk+cDHZtGwfYEFErHNNeyqMM8gKZdki4jHgcGAhcIek00tFAm6LiL7ptkdEXNGwi0qO18i+S1nJ2r/jLYvuX55+rmLjf4FmZtakAQMGMHjwYA444AD2228/Pv74Y4YPH84111zD9ddfz6677spbb73F2Wef3XohImKjvAHnAKOK2h4FfgiMKGirJZuuBegEzAX6NbHf+UD3tLwk/exNNhUMcAFwQ1p+B9gsLe8N/Bn4dFrvlrbrCbwKbAtsBkwFbmri+OPIptgB2gPbAAcAz6f8WwGzgc+l/S1O+94CeBq4osTj7g7MT8tfI3vx0OxzvPvuu0cePfLII9WO0Ki8ZnOuyjhXZZxrbcD0aOT/1Y12pEw2VX1PUdsE4CBgUlH7WEkzyaZxx0REXYXHqgFmSnqWrKg1vHn3FuB5SWMj4gXgB8BDkp4HJgM9I+JvwBXAU8AfyUbqTTkfGChpVsq7T2TnyccAfwKmAb+MiGcjYgXZhVvTgAeAl8p4LI8Ae/tCLzOztrfRTllGRE2J5lHA6RExv5l+Te13p4LlzunnbcBtJfpeClxasH4XcFeJfqOB0WUe/w3ghBLt1wPXl2j/T7KLw4rbawqWF5POKUfE28CB5WQxM7OWtdEW5VIiYjlQ8lNUzMzMqm2TKsqVkDSN7DxsodMiYlYbHf/7wMlFzeMj4qq2OL6ZmbU9F+VGRERFV2C3wvGvAlyAzcw2IRvzhV5mZmYbFBdlMzOznHBRNjMzywkXZTMzs5xwUTYzM8sJF2UzM7OccFE2MzPLCRdlMzOznHBRNjMzywkXZTMzs5xwUTYzM8sJF2UzM7OccFE2MzPLCRdlMzOznHBRNjMzywkXZTMzs5xwUTYzM8sJF2UzM7OccFE2MzPLCRdlMzOznHBRNjMzywkXZTMzs5xwUTYzM8sJRUS1M9gGasdddo12p9xY7RjruHC/lfx0Vodqxygpr9mcqzLOVZm85Jp/9XFrrdfW1tKzZ0+GDBmyuu2VV15h5MiRDBw4kBEjRrBkyRJ22mknxo4dyzbbbNMiOSTVRUT/Uvd5pGxmZpusPfbYg5kzZzJz5kzq6uro1KkTgwYN4p//+Z+5+uqrmTVrFoMGDeLaa69tkzwbXFGWNEhSSNqzqH2SpF6SaiXVS5op6UVJw5vZ38mp3yNN9KmR9EBaPlPSTS3zaPJJ0veqncHMrK1NmTKFPn360Lt3b+rr6zn88MMBOOqoo5gwYUKbZNjgijIwFHgcOLWhQVJHoFtELExNwyKiL3AIcI2kzZvY39nAv0TEwNYKvD4kVXOux0XZzDY548aNY+jQoQDsu+++TJw4EYDx48ezYMGCNslQ/Un+CkjqTFZoBwITgSvSXTVAbYlNOgNLgVWN7O9y4FBgZ0kTgX8H/hvoD6wEvhsRTY2gewO/AnoAi4CzgIXAn4E+QBfgbaAmIh6TNDX1+RvwX8B+ZL+DKyLiPklnAscBWwJbAUc0ctxLgNOAj4E/RMRlkvoCNwOdgJeBb0TEO5JqgYsiYrqk7sD0iNgpHeurqX8f4J6IuETS1UBHSTOBORExrOjYw4HhAN279+Dy/VY29vRUzXYds3NYeZTXbM5VGeeqTF5y1dbWrrW+ZMmS1W0rVqxgwoQJHH/88dTW1jJixAiuvPJKLr74Yg455BDatWu3zvatYYMqysCJwKSImCvpbUkHRMQM4Bjg3oJ+YyUtB3YDLoiIkkU5IkZKOoI1RevC1L5fmh5/SNLuTeS5Cbg9Im6T9A3gPyPiRElzgb2BnYE64DBJ04AdIuIvkv4f8HBEfENSV+BPkv6Y9nkwsH9EvF3qgJKOSc/DgIhYJqlbuut24LyIeFTSSOCHwAVNZAfoC3wOWA7US/qvVODPTTMNpZ6zW4BbILvQKw8XbxTLy0UlpeQ1m3NVxrkqk5dc84fVrLVeW1tLTU3Wdt999zFgwABOOumk1feffvrpAMydO5c5c+as7tuaNrTp66HAuLQ8Lq1DNnp+vKDfsIjYH9gRuCiNaMtxKHAHQES8BLwKNFWUDwZ+k5bvSNsDTAUOT7cfp/YDgWfS/UcDl6XRaC3ZyHjHdN/kxgpyciQwOiKWpZxvS+oCdI2IR1Of29KxmzMlIt6LiL8DLwDlPk9mZhuVO++8c/XUNcCbb74JwMcff8yVV17JiBEj2iRH9V+6lEnStmTTuftKCqA9EJJuBhZExEfF20TEIkkzgAFkBbbZw3zCmA3vL5sKjAC2By4HLiabYn+s4Dhfi4j6tQ4uDSCbbm8uYyXvY1vJmhdfWxbdt7xgeRUV/j103Kw99UVvMciD2tradV4R50VeszlXZZyrMnnN1WDZsmVMnjyZUaNGrW678847+fnPfw7ASSedxFlnndUmWTakkfJgsqni3hGxU0R8BpgHfB2YVGoDSZ3IpmdfLvMYjwHD0ra7k41e65vo/yRrLjgbxprR+jTgC8DHaRQ6EziHrFgDPAicJ0npWJ8rMx/AQ8A30mNDUreIeA94R9Jhqc9pQMOoeT7QLy0PLvMYKyRtVkEmM7MNVqdOnXjrrbfo0qXL6rbzzz+fuXPnMnfuXK6++mrSf9etbkMqykOBe4raJgAHsW5RHpumhuuAMRFRV+YxfgG0lzQLuAs4MyKWN9H/O8BZkp4nK4TnA6RtFgBPp35Tga2BWWn9P4DNgOclzU7rZYmISWQXuU1Pj/GidNcZwLUpS19gZGq/DviWpCeB7mUe5paUbWy5uczM7JPbYKavI6KmRPMo4PSImN9Mv7L2m0a1Z5boU0u6ujsixgBj0vJ8GrlCOiIOK1j+DWvOPRMRH5KNnIu3Wb3vZjJfDVxd1DaT7AVKcd+XgP0Lmn5Q6lgRcXzB8qXApc3lMDOzlrXBFOVS0oi05EeVmZmZbWg26KJcifSWpC2Kmk+LiFml+lebpP1IV4IXWB4RA6qRx8zMWt8mU5Q3tGKWXiyUfK+wmZltnDakC73MzMw2ai7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWExUXZUmfkrR/8z3NzMysEmUVZUm1kraR1A14Dhgt6frWjWZmZrZpKXek3CUi3gdOAkZHRD/gyNaLZWZmtukptyh3kNQTOAV4oBXzmJmZbbLKLcojgQeBlyPiGUm7AH9uvVhmZmabnrK+kCIixgPjC9ZfAb7WWqHMzMw2ReVe6LW7pCmSZqf1/SX9oHWjmZmZbVrKnb6+Ffg3YAVARDwPnNpaoczMzDZF5RblThHxp6K2lS0dxszMbFNWblFeLKkPEACSBgN/a7VUZmZmm6CyLvQCvg3cAuwpaSEwDxjWaqnMzMw2Qc0WZUntgP4RcaSkrYB2EfFB60czMzPbtDQ7fR0RHwPnpuWlLshmZmato9xzypMlXSTpM5K6NdxaNZmZmdkmptxzyt9IP79d0BbALi0bx8zMbNOliKh2BttA7bjLrtHulBurHWMdF+63kp/OKvf1ZtvKazbnqoxzVaatcs2/+rh12urr6xkyZMjq9VdeeYWRI0fy1FNPUVdXR+fOnXn33Xfp2rUrM2fObPWMAJLqIqJ/qfvKepYknV6qPSJu/yTBzMzMWtMee+yxutiuWrWKXr16MWjQIC644AJqa2upqanhwgsvpEuXLlVOmin3nPKBBbfDgCuArza3kaRBkkLSnkXtkyT1St/TXC9ppqQXJQ2vMH9zx6+R9IWW3Gfa75mSbiqzbw9J0yQ9K+mw9TzW9pWnrOgYvSXVpd/DHEkjWvN4ZmbVMGXKFPr06UPv3r1Xt0UEv/3tbxk6dGgVk61R7hdSnFe4LqkLcEcZmw4FHif7SM4r0rYdgW4RsVASwLCImJ4uHHtZ0piI+Kj8h9CkGmAJ8GQL7Q9Jlc7BfBF4KSLOWM9DngnMBl4vdwNJHSKikk9c+xvwhYhYLqkzMFvSxIgo+5hmZnk3bty4dYrv1KlT2W677dhtt92qlGpt63VOWdJmwPMRsVcTfToD9cBAYGJE7JnajwEGRsQlkmqBi1JR3pGsePaOiFWN7HMJcCNwPPAhcEJEvCGpB3AzsGPqegGwEHgaWAUsAs4HRgN9gC7A20BNRDwmaSpwVmr7FdkFbMuA4RHxvKQrgO2BnYDFwENk790+V9JxwA+Ar0TE4qK8fYGJQMeU52CymYYfAVsALwNnRcQSSZcDX0l9nwTOIfsmrjFp2w/T9i+mYy+W1B+4LiJqSmQ8Dbia7IXJFsDPI2JUI7+uwszbAs8CB5Uqymk2YzhA9+49+l1+w63N7bLNbdcR3viw2ilKy2s256qMc1WmrXLt16vxKegVK1YwePBgRo8eTbdu2ZuHlixZwq233kqvXr045ZRTWj9gMnDgwE98Tvl+0kdskk15703BVzk24kRgUkTMlfS2pAMiYgZwDHBvQb+xkpYDuwEXNFaQk62ApyPi+5J+AnwTuJKsUP8sIh5Pxf3BiNhL0s3Akoi4Lj2OuSn7zkAdcJikacAOEfEXSf8FPBsRJ0o6Argd6JuO3Q84NCI+lHRm2t8g4LvAsRHxTnHYiJiZim1DAe9OVsCPjIilki5N248EboqIkWm/dwDHR8Tdks4lvXBJ9zX1nBdmHA68FxEHStoCeELSQxExr9SGkj4D/C+wK3BxY6PkiLiF7NPd2HGXXWNTvqhkfeQ1m3NVxrkq02YXeg2rafS+++67jwEDBnDSSSetbpsyZQpPP/00dXV17LDDDq2erxzlPkvXFSyvBF6NiNea2WYocENaHpfWZwCHABcV9GuYvu4BPClpUkS82sg+PwIeSMt1wFFp+Uhg74KCtY2krUtsPxU4nKwo/5isqD8KPJPuP5T0PdER8bCkbdNUPWSj/cLXegOB/sDREfF+I3mLHUT2ouCJlHVz4KmG/Um6BOgEdAPmAPeXud8GhRmPBvZPn1MO2ezAbmQfkbqOiFiQ+m8P3Cvp7oh4o8Ljm5nl0p133rnO1HVdXR177rlnbgoylF+Uj42ISwsbJF1T3FZw37bAEcC+kgJoD0QauS4odc44IhZJmgEMABoryitizXz7qoL87YCDi4pmqVHlVGAE2TTv5cDFZNO7jzVsUuKYDcdbWtT+Ctk09+7A9EbyFhMwOSLW+suQtCXwC7IR9YI0Fb1lI/tYyZoL9Ir7FGYUcF5EPFhmNgAi4nVJc8im2e+uZFszszxatmwZkydPZtSotc/gPfzww7m5wKtBuUX5KKC4AB9Toq3BYOD2iDinoUHSo8DXgUmlNpDUCfgc8JMyMxV6iOyjQK9N++obETOBD4BtCvpNI5uSfiUi/i5pJtm52+PT/Y+RfdHGf0iqARZHxPuNTBm/Sjbiv0fSyRExp4ycTwM/l7Rrmi7vBOwAvJnuX5zOxQ9mTUH8ACgc9c8nm6b+A2lU34gHgW9JejgiVkjaHVgYEcUvLpC0A/BWmvb+FNlsxvXNPZiOm7WnvsT7Aquttra2yWmsasprNueqjHNVptq5OnXqxFtvvbVO+2WXXUZNTU3bB2pCk2+JkvQtSbOAPSQ9X3CbBzzfxKZDgXuK2iaQTd8WF+WxqTjWAWMioq6yhwDAd4D+KdsLZKNhyKZ/B6W3+hwWEcuBBWTFEbKR89bArLR+RcN+yC6SavKK6YioJyvi45V9tWWTImIR2dXUd6ZjPA3sGRHvAremHPeyZjodsgu9bk6PoSPZRWI3povTmjr//kvgBWCGpNnAKBp/EbYXME3Sc2TT+ddFxKxG+pqZWStpbqT8G7IR2Y+BywraP4iItxvbKCJqSjSPAk6PiPnN9GtURHQuWL6bNJpMVz0PKdF/LrB/UdthBcu/IXuMDetvAyeU2M8VRetjyIolEfEs2XnixjKv7pvWHyZ7v3dxvx+QXQRW3D6B7AVNg6lkU+bNZfwY+F66NSkiJlP0PJmZWdtrsihHxHvAe2QjXyR9muw8ZmdJnSPir+UeKI1SS14CbmZmZuW/JeorZOcYtyc7/9mb7P2y+7RGqPQ2pS2Kmk/L85SqpO8DJxc1j4+Iq6qRpxRJ+7Huh74sj4gB1chjZmZrK/dCryvJzgf/MSI+J2kgafTcGjbEIpGKb24KcCnpRU3fZjuamVlVlPvZ1ysi4i2gnaR2EfEI/s/dzMysRZU7Un43vVVnKtnV0m+SvV/WzMzMWki5I+UTyD4L+gKytzS9TPY5zWZmZtZCyv2WqKWSegO7RcRt6UMv2rduNDMzs01LWSNlSd8ke09ww2eU9WLtL5UwMzOzT6jc6etvk3304vsAEfFn4NOtFcrMzGxTVG5RXl74JRKSOrDmixrMzMysBZRblB+V9D2go6SjyL5LudKvFTQzM7MmlFuULwMWkX1hwjnA7ynxOc1mZma2/pq8+lrSjhHx1/TlBremm5mZmbWC5kbKq6+wljShqY5mZmb2yTRXlFWwvEtrBjEzM9vUNVeUo5FlMzMza2HNfaLXZyW9TzZi7piWSesREdu0ajozM7NNSJNFOSL8UZpmZmZtpNy3RJmZmVkrc1E2MzPLCRdlMzOznHBRNjMzywkXZTMzs5xwUTYzM8sJF2UzM7OcaO7DQ8zMzNpUfX09Q4YMWb3+yiuvMHLkSBYuXMj999/P5ptvTp8+fRg9ejRdu3atYtKWpwh/eqatnx132TXanXJjtWOs48L9VvLTWfl8vZnXbM5VGeeqTFO55l99XJPbrlq1il69ejFt2jTq6+s54ogj6NChA5deeikA11xzzXrnqq2tpaamZr23X1+S6iKif6n72mT6WtIgSSFpz6L2SZJ6SaqVVC9ppqQXJQ1voeOeKWn7lthXI/uvkfSF1tp/pSR9VdJlLbSvJS2xHzOzT2LKlCn06dOH3r17c/TRR9OhQ1bcDzroIF577bUqp2t5bXVOeSjwOHBqQ4OkjkC3iFiYmoZFRF/gEOAaSZuXs2NJTX0U6JlAqxVloAb4REVZUou9rI2IiRFxdUvtz8ys2saNG8fQoUPXaf/Vr37FMcccU4VEravVp68ldQbqgYHAxIjYM7UfAwyMiEsk1QIXRcR0STsCTwK9I2JVI/tcAlwPfAm4EPgwrXcGFpMV40OAMcDCdP/5wPkRcZKkE4BxQBeyFyYvRMQukvoAPwd6AMuAb0bES5J6ADcDO6YIF6T9Pg2sAhYB5wH/APwwtb0XEYc3kv9M4DhgS2CriDhC0sXAKcAWwD0R8cPU93TgIrJv6Xo+Ik4rlScinkj77Q98H3gO2CUiPpbUKf0OdknblHqMOwO/IbvOYBLwrxHRuUT24cBwgO7de/S7/IZbSz3EqtquI7zxYbVTlJbXbM5VGeeqTFO59uvVpdHtVqxYweDBgxk9ejTdunVb3f7rX/+a+vp6Ro4ciaRGt2/OkiVL6Nx5nf/mWt3AgQMbnb5ui5MPJwKTImKupLclHRARM4BjgHsL+o2VtBzYjazIlCzIyVbA7Ii4XNJmwKPACRGxSNIQ4KqI+Iakc1lT7DuQFfoKaNIAABKJSURBVGmAw4DZwIFkz8G01H4LMCIi/ixpAPAL4AjgRuBnEfF4etHwYETsJelmYElEXAcgaRbwpYhYKKm5qw8OBvaPiLclHZ0e9+fJvoFroqTDgbfICuwhEbFYUsNf5Tp5gL0adhwR70l6DvhH4BHgKynzCklNPcb/jojbJX27sdARcUt6nthxl11jQzt/VW15zeZclXGuyjR5TnlYTaPb3XfffQwYMICTTjppddttt93GnDlzmDJlCp06dfpEuap1TrkpbfHbGwrckJbHpfUZZCPZiwr6DUvFswfwpKRJEfFqI/tcBUxIy3sA+wKT0yum9sDfijeIiJWS/iJpL7Lidz1weOo/NY3ovwCML3jltUX6eSSwd0H7NpK2LpHrCWCMpN8Cv2ske4PJEfF2Wj463Z5N653JivRngbsjYnF6DA39y8lzFzCErCifCvyimcd4CPC1tHwHsP5XT5iZtYA777xzranrSZMmcc011/Doo49+4oKcV61alCVtSzYK21dSkBXASCPMBRHxUfE2abQ7AxgANFaU/14wkhYwJyIOLiPSVLIR+grgj2Qj5/ZkLw7aAe+m89rF2gEHR8RaEzDF0yYRMSKNPo8DZkrqGxFvNZJlaeGugB9HxKii/X+HbNp6ffJMBH6cRtf9gIfJZhgae4w0ciwzsza3bNkyJk+ezKhRa/5bPPfcc1m+fDlHHXUUkF3sdfPNN1crYuuIiFa7AecAo4raHiU77zqioK0W6J+WOwFzgX5N7HdJwfLmwF/IihTAZsA+afl+svPWDX1rgL8CV6b1p4H5rDm3/iRwcloW8Nm0/Bvg4oL99E0/LwR+VNDep2D52YZ+JfKfCdxUsH402RR657TeC/g0sE96LrZN7d2ayVO83/Fko95fFLQ19hgnAl9Py98qfI4bu+2+++6RR4888ki1IzQqr9mcqzLOVRnnWhswPRr5f7W1r74eCtxT1DYBOIjsYqJCYyXNBOqAMRFRV84BIhttDya7Yvs5YCZrrogeA9yc3mrVkazwbQc8lu5/nuziqYYR4jDg7LSfOcAJqf07QH9Jz0t6ARiR2u8HBqX9HwZcK2mWpNnpGM+V+RgeIiu0T6Xz0ncDW0fEHOAq4NGU6fpm8hS7C/h6+tmgscd4PvBtSc+QXQBnZmZtrFWnryOipkTzKOD0iJjfTL+m9tu5aH0m2fnh4n4TWHPuucEWBfcPL+o/D/hyif0sJjs/W9w+F9i/oGlqGfGJiDGsueisoe1GsoutivveBtxWZp619hsRd5ONhgv7NPYY55FdfNbAb60yM2tjbX6ZXkQsJ3vbjpmZmRXI37XzBSRNo2Bkm5wWEbOqkadSkr7Eulcxz4uIQdXIY2Zm+ZbrohwRA6qd4ZOIiAfJ3kNsZmbWLH91o5mZWU64KJuZmeWEi7KZmVlOuCibmZnlhIuymZlZTrgom5mZ5YSLspmZWU64KJuZmeWEi7KZmVlOuCibmZnlhIuymZlZTrgom5mZ5YSLspmZWU64KJuZmeWEi7KZmVlOuCibmZnlhIuymZlZTrgom5mZ5YSLspmZWU64KJuZmeWEi7KZmVlOuCibmZnlhIuymZlVVX19PX379l1922abbbjhhhsYP348++yzD+3atWP69OnVjtkmFBHVzmAbqB132TXanXJjtWOs48L9VvLTWR2qHaOkvGZzrso4V2WKc82/+rhG+65atYpevXoxbdo0li1bRrt27TjnnHO47rrr6N+/f4vmqq2tpaampkX3WQ5JdRFR8sG06khZ0iBJIWnPovZJknpJqpVUL2mmpBclDW/h49dI+kJL7jPt90xJN5XZt4ekaZKelXTYeh5r+8pTVnSMvpKekjRH0vOShrTm8czMGjNlyhT69OlD79692Wuvvdhjjz2qHalNtfb09VDgceDUhgZJHYFuEbEwNQ2LiL7AIcA1kjZvwePXAC1alCVV+jL0i8BLEfG5iJi6Hoc8E6ioKK9HxmXA6RGxD/Bl4AZJXSvch5nZJzZu3DiGDh1a7RhV02pFWVJnskJ7NgVFmaxQ1pbYpDOwFFjVxD6XSLpK0nOSnpa0XWrvIWmCpGfS7RBJOwEjgH9NI/F/lPSKMl0lfSzp8LT9VEm7Suom6d40Wnxa0v7p/isk3SLpIeD2okzHpVFm9xJ5+wI/AY5NGTpKOjr1nyFpfHqekHR5yj47HUuSBgP9gbEF289vOJak/pJqS2WU1F7StWmfz0s6p7HnNSLmRsSf0/LrwJtAj8b6m5m1ho8++oiJEydy8sknVztK1bTmyYcTgUkRMVfS25IOiIgZwDHAvQX9xkpaDuwGXBARjRZlYCvg6Yj4vqSfAN8ErgRuBH4WEY9L2hF4MCL2knQzsCQirgOQNBfYG9gZqAMOkzQN2CEi/iLpv4BnI+JESUeQFeC+6dj9gEMj4kNJZ6b9DQK+CxwbEe8Uh42ImZIuB/pHxLmpmP4AODIilkq6NG0/ErgpIkam/d4BHB8Rd0s6F7goIqan+5p6zgszDgfei4gDJW0BPCHpoYiY19QOJH0e2Bx4uZH7hwPDAbp378Hl+61sandVsV3H7BxWHuU1m3NVxrkqU5yrtra2ZL/HH3+cnXfemRdffJEXX3xxdfu7775LXV0dS5YsadFcS5YsaTRLtbRmUR4K3JCWx6X1GWSj54sK+g2LiOmSegBPSpoUEa82ss+PgAfSch1wVFo+Eti7oGBtI2nrEttPBQ4nK8o/JivqjwLPpPsPBb4GEBEPS9pWUpd038SI+LBgXwPJRrFHR8T7jeQtdhDZi4InUtbNgaca9ifpEqAT0A2YA9xf5n4bFGY8Gtg/jbYBupC98Gm0KEvqCdwBnBERH5fqExG3ALdAdqHXhnBRSZ7kNZtzVca5KrPOhV7Dakr2u/nmm/mXf/mXdS6+6tq1K/369dtoLvRqSqv89iRtCxwB7CspgPZApJHrgoj4qHibiFgkaQYwAGisKK+INZeLr2JN/nbAwUVFs9SocirZlPb2wOXAxWTT6Y81bFLimA3HW1rU/gqwC7A7UO61+gImR8RaJ0wkbQn8gmxEvUDSFcCWjexjJWtOOxT3Kcwo4LyIeLCsYNI2wP8CP4iIp8vZxsyspSxbtozJkyczatSo1W333HMP5513HosWLeK4446jb9++PPhgWf+lbbBa6yXVYOD2iFh9HlPSo8DXgUmlNpDUCfgc2TnYSj0EnAtcm/bVNyJmAh8A2xT0m0Y2Jf1KRPxd0kzgHOD4dP9jwDDgPyTVAIsj4v1GpoxfJRvx3yPp5IiYU0bOp4GfS9o1TZd3AnYgO4cLsDidYx4M3J3aPgAKR/3zyaap/0Aa1TfiQeBbkh6OiBWSdgcWRkTxiwvSxXX3kP3OxpfxOADouFl76pt4a0O11NbWNvpKvNryms25KuNclSknV6dOnXjrrbfWahs0aBCDBg1qxWT501oXeg0l+0++0ASy6dviojw2Fcc6YExE1K3H8b4D9E8XNL1ANhqGbPp3ULpI6rCIWA4sICuOkI2ctwZmpfUrGvYDXA2c0dRBI6KerIiPl9SnuZARsYjsauo70zGeBvaMiHeBW1OOe1kznQ4wBri54UIv4EfAjZKm0sRFccAvgReAGZJmA6No/EXYKWTT+mem48xMF6mZmVkbapWRckTUlGgeRfa2m/nN9Gtqv50Llu8mjSYjYjGwzntrI2IusH9R22EFy78BflOw/jZwQon9XFG0PoasWBIRz5KdJ24s8+q+af1h4MAS/X5AdhFYcfsEshc0DaaSTZk3l/Fj4Hvp1qSI+DXw6+b6mZlZ62qzKwLSKLVlz9KbmZltRPJ3mR6Q3qa0RVHzaRExq1T/PJD0faD4zXXjI+KqauQpRdJ+ZFdXF1oeEQOqkcfMzNaWy6K8IRaJVHxzU4BLSS9qfK7YzCyn/C1RZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEy7KZmZmOeGibGZmlhMuymZmZjmhiKh2BttASfoAqK92jhK6A4urHaIRec3mXJVxrso419p6R0SPUnd0aOsktlGpj4j+1Q5RTNL0POaC/GZzrso4V2Wcq3yevjYzM8sJF2UzM7OccFG2T+KWagdoRF5zQX6zOVdlnKsyzlUmX+hlZmaWEx4pm5mZ5YSLspmZWU64KNt6kfRlSfWS/iLpsirm+JWkNyXNLmjrJmmypD+nn5+qQq7PSHpE0ouS5kg6Pw/ZJG0p6U+Snku5fpTad5Y0LeW6S9LmbZmrIF97Sc9KeiAvuSTNlzRL0kxJ01NbHv7Gukq6W9JL6e/s4GrnkrRHep4abu9LuqDauVK2f01/87Ml3Zn+LVT976uYi7JVTFJ74OfAMcDewFBJe1cpzhjgy0VtlwFTImI3YEpab2srgQsjYi/gIODb6TmqdrblwBER8VmgL/BlSQcB1wA/S7neAc5u41wNzgdeLFjPS66BEdG34D2t1f49AtwITIqIPYHPkj1vVc0VEfXpeeoL9AOWAfdUO5ekXsB3gP4RsS/QHjiV/Px9rRERvvlW0Q04GHiwYP3fgH+rYp6dgNkF6/VAz7Tck+xDTqr9nN0HHJWnbEAnYAYwgOxTjTqU+v22YZ4dyP7DPgJ4AFBOcs0Huhe1VfX3CGwDzCNdrJuXXEVZjgaeyEMuoBewAOhG9qFZDwBfysPfV/HNI2VbHw1/4A1eS215sV1E/A0g/fx0NcNI2gn4HDCNHGRLU8QzgTeBycDLwLsRsTJ1qdbv8wbgEuDjtL5tTnIF8JCkOknDU1u1f4+7AIuA0Wm6/5eStspBrkKnAnem5armioiFwHXAX4G/Ae8BdeTj72stLsq2PlSize+tK0FSZ2ACcEFEvF/tPAARsSqy6cUdgM8De5Xq1paZJB0PvBkRdYXNJbpW4+/skIg4gOx0zbclHV6FDMU6AAcA/x0RnwOWUp0p9JLSudmvAuOrnQUgncM+AdgZ2B7Yiuz3Wazq/4+5KNv6eA34TMH6DsDrVcpSyhuSegKkn29WI4SkzcgK8tiI+F2esgFExLtALdk5766SGj4Lvxq/z0OAr0qaD4wjm8K+IQe5iIjX0883yc6Pfp7q/x5fA16LiGlp/W6yIl3tXA2OAWZExBtpvdq5jgTmRcSiiFgB/A74Ajn4+yrmomzr4xlgt3Tl4uZk01QTq5yp0ETgjLR8Btn53DYlScD/AC9GxPV5ySaph6Suabkj2X9WLwKPAIOrlSsi/i0idoiIncj+nh6OiGHVziVpK0lbNyyTnSedTZV/jxHxf8ACSXukpi8CL1Q7V4GhrJm6hurn+itwkKRO6d9mw/NV1b+vkqp9Utu3DfMGHAvMJTsf+f0q5riT7BzRCrLRw9lk5yKnAH9OP7tVIdehZFNhzwMz0+3YamcD9geeTblmA5en9l2APwF/IZty3KKKv9Ma4IE85ErHfy7d5jT8rVf795gy9AWmp9/lvcCncpKrE/AW0KWgLQ+5fgS8lP7u7wC2qPbfV6mbP2bTzMwsJzx9bWZmlhMuymZmZjnhomxmZpYTLspmZmY54aJsZmaWEx2a72Jm1rYkrQJmFTSdGBHzqxTHrM34LVFmljuSlkRE5zY8XodY8xnIZlXj6Wsz2+BI6inpsfSdvbMlHZbavyxpRvq+6CmprZukeyU9L+lpSfun9isk3SLpIeD29EUd10p6JvU9p4oP0TZRnr42szzqmL7JCrLPLB5UdP8/kX3N3lXp+707SeoB3AocHhHzJHVLfX8EPBsRJ0o6Arid7NOwIPvO30Mj4sP0DVDvRcSBkrYAnpD0UETMa80HalbIRdnM8ujDyL7JqjHPAL9KX/pxb0TMlFQDPNZQRCPi7dT3UOBrqe1hSdtK6pLumxgRH6blo4H9JTV8FnIXYDey7y02axMuyma2wYmIx9JXKB4H3CHpWuBdSn/1XlNfAbm0qN95EfFgi4Y1q4DPKZvZBkdSb7LvX76V7Nu4DgCeAv5R0s6pT8P09WPAsNRWAyyO0t9t/SDwrTT6RtLu6ZuhzNqMR8pmtiGqAS6WtAJYApweEYvSeeHfSWpH9p29RwFXAKMlPQ8sY81XCBb7JbATMCN9vd8i4MTWfBBmxfyWKDMzs5zw9LWZmVlOuCibmZnlhIuymZlZTrgom5mZ5YSLspmZWU64KJuZmeWEi7KZmVlO/H9/ArEI1ZsHhAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Select the top 5 features and re-run the model \n",
    "\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "\n",
    "cols = ['A/B_follower_count','A/B_network_feature_3','A/B_listed_count',  \n",
    "        'A/B_network_feature_2','A/B_retweets_received' ]\n",
    "x = x_train[cols]\n",
    "y = y_train \n",
    "\n",
    "\n",
    "model_xgb = XGBClassifier() #New xgb classifier for the most important features \n",
    "model_xgb.fit(x_train, y_train) #Train the model on training data \n",
    "\n",
    "#Predicting the test results\n",
    "\n",
    "y_pred = model_xgb.predict(x_test)\n",
    "\n",
    "plot_importance(model_xgb, max_num_features=5)\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy of XGBoost model: 0.8166233766233766\n",
      "Testing accuracy of XGBoost model: 0.7866666666666666\n"
     ]
    }
   ],
   "source": [
    "# Calculating the accuracy of the model \n",
    "\n",
    "print(\"Training accuracy of XGBoost model:\", model_xgb.score(x_train, y_train))\n",
    "print(\"Testing accuracy of XGBoost model:\", model_xgb.score(x_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The difference between training and testing accuracy indicates how much the model has overfitted. The XGBoost model has a low overfitting value of 0.03 which indicates the model has not trained data too well to negatively impact the performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[636 184]\n",
      " [168 662]]\n"
     ]
    }
   ],
   "source": [
    "# Confusion Matrix \n",
    "\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "\n",
    "confusion_matrix = confusion_matrix(y_test, y_pred)\n",
    "print(confusion_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The confusion matrix visualises the actual and predicted values of the data thereby depicting the performance of a supervised algorithm. Here, it can be inferred that there are 1,298 correct predictions and 352 incorrect predictions. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.78      0.78       820\n",
      "           1       0.78      0.80      0.79       830\n",
      "\n",
      "    accuracy                           0.79      1650\n",
      "   macro avg       0.79      0.79      0.79      1650\n",
      "weighted avg       0.79      0.79      0.79      1650\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Classification report will summarise our model by computing precision, recall, f-measure and support \n",
    "\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In social media analytics, high precision means that an algorithm returned more relevant results than irrelevant, while high recall means that an algorithm returned most of the relevant results. In the XGBoost model, there is a 79% precision that there are more relevant values and a 79% recall that most of the relevant values are included in our model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I tried 4 models on the dataset: Logistic Regression, Random Forest, K-Nearest Neighbors and XGBoost. On comparing these models, XGBoost has the highest training and testing accuracy of 81% and 79% respectively. Further, on analyzing the confusion matrix, XGBoost model has the highest number of correctly predicted values (1,298). Lastly, according to the classification report, XGBoost model has the highest precision, recall and f-measure (79%) indicating that the model selected more relevant values than irrelevants values as well as the most relevant values in the dataset have been included in our model. Therefore, the key predictors of social influence according to XGBoost model are: A/B_listed_count, A/B_follower_count, A/B_network_feature_3, A/B_retweets_received and A/B_network_feature_2.This data can be used by companies in choosing most influencial people on media to partner with. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}