EDA.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import gc\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "\n",
    "p = sns.color_palette()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# All file and their sizes:\n",
      "\n",
      "clicks_test.csv               506.95MB\n",
      "clicks_train.csv              1486.73MB\n",
      "documents_categories.csv      118.02MB\n",
      "documents_entities.csv        324.1MB\n",
      "documents_meta.csv            89.38MB\n",
      "documents_topics.csv          339.47MB\n",
      "events.csv                    1208.55MB\n",
      "page_views_sample.csv         454.35MB\n",
      "promoted_content.csv          13.89MB\n",
      "sample_submission.csv         273.14MB\n",
      "sub1.csv                      273.14MB\n"
     ]
    }
   ],
   "source": [
    "print \"# All file and their sizes:\"\n",
    "print\n",
    "for f in os.listdir('../Outbrain'):\n",
    "    if \".csv\" in f:\n",
    "        print f.ljust(30) + str(round(os.path.getsize('../Outbrain/' + f) / 1e6, 2)) + 'MB'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We are missing the \"```page_views.csv```\" file, because it is over 100GB unzipped. So, we just look into the page_views_sample.csv instead on developing our analysis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "clicks_test = pd.read_csv('clicks_test.csv')\n",
    "clicks_train = pd.read_csv('clicks_train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Variable names: \n",
      "\n",
      "clicks_test                    ['display_id' 'ad_id']\n",
      "clicks_train                   ['display_id' 'ad_id' 'clicked']\n"
     ]
    }
   ],
   "source": [
    "print \"Variable names: \"\n",
    "print \n",
    "print \"clicks_test\".ljust(30), clicks_test.columns.values\n",
    "print \"clicks_train\".ljust(30), clicks_train.columns.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>display_id</th>\n",
       "      <th>ad_id</th>\n",
       "      <th>clicked</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>42337</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>139684</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>144739</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>156824</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>279295</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>296965</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2</td>\n",
       "      <td>125211</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2</td>\n",
       "      <td>156535</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2</td>\n",
       "      <td>169564</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2</td>\n",
       "      <td>308455</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>3</td>\n",
       "      <td>71547</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>3</td>\n",
       "      <td>95814</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>3</td>\n",
       "      <td>152141</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>3</td>\n",
       "      <td>183846</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>3</td>\n",
       "      <td>228657</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>3</td>\n",
       "      <td>250082</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>4</td>\n",
       "      <td>149930</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>4</td>\n",
       "      <td>153623</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>4</td>\n",
       "      <td>184709</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>4</td>\n",
       "      <td>186849</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>4</td>\n",
       "      <td>233730</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>5</td>\n",
       "      <td>54606</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>5</td>\n",
       "      <td>162868</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>5</td>\n",
       "      <td>184548</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>5</td>\n",
       "      <td>282674</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>5</td>\n",
       "      <td>326514</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>6</td>\n",
       "      <td>85397</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>6</td>\n",
       "      <td>160754</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>6</td>\n",
       "      <td>173388</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>6</td>\n",
       "      <td>175234</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141701</th>\n",
       "      <td>16874588</td>\n",
       "      <td>23690</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141702</th>\n",
       "      <td>16874588</td>\n",
       "      <td>174903</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141703</th>\n",
       "      <td>16874588</td>\n",
       "      <td>468733</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141704</th>\n",
       "      <td>16874588</td>\n",
       "      <td>526478</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141705</th>\n",
       "      <td>16874589</td>\n",
       "      <td>3314</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141706</th>\n",
       "      <td>16874589</td>\n",
       "      <td>187175</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141707</th>\n",
       "      <td>16874589</td>\n",
       "      <td>188284</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141708</th>\n",
       "      <td>16874589</td>\n",
       "      <td>226819</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141709</th>\n",
       "      <td>16874589</td>\n",
       "      <td>243297</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141710</th>\n",
       "      <td>16874589</td>\n",
       "      <td>394670</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141711</th>\n",
       "      <td>16874590</td>\n",
       "      <td>35805</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141712</th>\n",
       "      <td>16874590</td>\n",
       "      <td>80235</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141713</th>\n",
       "      <td>16874590</td>\n",
       "      <td>132287</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141714</th>\n",
       "      <td>16874590</td>\n",
       "      <td>142604</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141715</th>\n",
       "      <td>16874590</td>\n",
       "      <td>142780</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141716</th>\n",
       "      <td>16874590</td>\n",
       "      <td>142816</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141717</th>\n",
       "      <td>16874591</td>\n",
       "      <td>36108</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141718</th>\n",
       "      <td>16874591</td>\n",
       "      <td>127047</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141719</th>\n",
       "      <td>16874591</td>\n",
       "      <td>145601</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141720</th>\n",
       "      <td>16874591</td>\n",
       "      <td>285839</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141721</th>\n",
       "      <td>16874591</td>\n",
       "      <td>397506</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141722</th>\n",
       "      <td>16874591</td>\n",
       "      <td>483448</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141723</th>\n",
       "      <td>16874592</td>\n",
       "      <td>82851</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141724</th>\n",
       "      <td>16874592</td>\n",
       "      <td>153244</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141725</th>\n",
       "      <td>16874592</td>\n",
       "      <td>167294</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141726</th>\n",
       "      <td>16874592</td>\n",
       "      <td>186600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141727</th>\n",
       "      <td>16874593</td>\n",
       "      <td>151498</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141728</th>\n",
       "      <td>16874593</td>\n",
       "      <td>282350</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141729</th>\n",
       "      <td>16874593</td>\n",
       "      <td>521828</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87141730</th>\n",
       "      <td>16874593</td>\n",
       "      <td>522693</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>87141731 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          display_id   ad_id  clicked\n",
       "0                  1   42337        0\n",
       "1                  1  139684        0\n",
       "2                  1  144739        1\n",
       "3                  1  156824        0\n",
       "4                  1  279295        0\n",
       "5                  1  296965        0\n",
       "6                  2  125211        0\n",
       "7                  2  156535        0\n",
       "8                  2  169564        0\n",
       "9                  2  308455        1\n",
       "10                 3   71547        0\n",
       "11                 3   95814        0\n",
       "12                 3  152141        0\n",
       "13                 3  183846        0\n",
       "14                 3  228657        1\n",
       "15                 3  250082        0\n",
       "16                 4  149930        0\n",
       "17                 4  153623        1\n",
       "18                 4  184709        0\n",
       "19                 4  186849        0\n",
       "20                 4  233730        0\n",
       "21                 5   54606        0\n",
       "22                 5  162868        0\n",
       "23                 5  184548        0\n",
       "24                 5  282674        0\n",
       "25                 5  326514        1\n",
       "26                 6   85397        0\n",
       "27                 6  160754        0\n",
       "28                 6  173388        0\n",
       "29                 6  175234        1\n",
       "...              ...     ...      ...\n",
       "87141701    16874588   23690        0\n",
       "87141702    16874588  174903        0\n",
       "87141703    16874588  468733        1\n",
       "87141704    16874588  526478        0\n",
       "87141705    16874589    3314        0\n",
       "87141706    16874589  187175        0\n",
       "87141707    16874589  188284        1\n",
       "87141708    16874589  226819        0\n",
       "87141709    16874589  243297        0\n",
       "87141710    16874589  394670        0\n",
       "87141711    16874590   35805        0\n",
       "87141712    16874590   80235        0\n",
       "87141713    16874590  132287        0\n",
       "87141714    16874590  142604        1\n",
       "87141715    16874590  142780        0\n",
       "87141716    16874590  142816        0\n",
       "87141717    16874591   36108        0\n",
       "87141718    16874591  127047        1\n",
       "87141719    16874591  145601        0\n",
       "87141720    16874591  285839        0\n",
       "87141721    16874591  397506        0\n",
       "87141722    16874591  483448        0\n",
       "87141723    16874592   82851        0\n",
       "87141724    16874592  153244        1\n",
       "87141725    16874592  167294        0\n",
       "87141726    16874592  186600        0\n",
       "87141727    16874593  151498        1\n",
       "87141728    16874593  282350        0\n",
       "87141729    16874593  521828        0\n",
       "87141730    16874593  522693        0\n",
       "\n",
       "[87141731 rows x 3 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clicks_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The distribution of the number of ads in a display:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1155faad0>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+8AAAGJCAYAAAD7SAJLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1153f3f10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_ads = clicks_train.groupby(\"display_id\")[\"ad_id\"].count().value_counts()\n",
    "num_ads = num_ads / np.sum(num_ads)\n",
    "plt.figure(figsize=(12,4))\n",
    "sns.barplot(num_ads.index, num_ads.values, alpha=0.8, color=p[0])\n",
    "plt.xlabel(\"Number of Ads in Display\")\n",
    "plt.ylabel(\"Proportion of data\")\n",
    "plt.title(\"Training set\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x115b189d0>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+8AAAGJCAYAAAD7SAJLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzs3XucHHWd7/9XzyQGk2VDEpSJ3CTo+QRclhXl4hEUr7Aq\nXvJTjxyWgwZFRfRwEe+IwnpDbooielwWsz+PuiKKLIjrAuoiRJFL0FU/QCJBIURISERycTLT54+q\nkU4zIT2dnunq5PV8PPoxXd/6dvWn61GZzLu/Vd+q1et1JEmSJElSdfV1uwBJkiRJkvT4DO+SJEmS\nJFWc4V2SJEmSpIozvEuSJEmSVHGGd0mSJEmSKs7wLkmSJElSxRneJUmSJEmqOMO7JEmSJEkVZ3iX\nJEmSJKniJnW7AEmS1FkR8c/AMZvp9sPMfGGH3m8K8Cnghsz8105s83Hea1fgi8CbM/O+8XwvSZKq\nxPAuSdLW5wzgCw3LHwaeCbwaqJVtf+zg++0CvAv4WQe3uSkvBQ6bgPeRJKlSDO+SJG1lMvO3wG9H\nliPiAWB9Zt40Tm9Z23yXnnwvSZIqo1av17tdgyRJGkflafTPz8w5m1g/D/gA8AzgIeDrwAczc225\n/onAecArgB2BJcCXMvP8iNgTuBOoUwTruzLzv23ifQ4HPgL8DfBn4IfAezPzzlZqiYhjgf9TvhfA\nP2Xmce3sE0mSeo0T1kmStA2LiKOBS4HbgVdRnHL/RuBbDd0+B7wY+N8Up61fAZwTEUcB9wCvowju\nHwZeu4n3eRpwGXAD8HLgWIqA/m9jqOVy4BPl8yOAj7f5sSVJ6jmeNi9J0jYqImrAJ4HvZub8hvbF\nwPcj4iWZ+QPgecDVmTkSon8cEX8CHszMwYi4rWxfkpm3b+LtDgSmAB/PzAfL97kPeHk5sr+ulVoi\nYkm56jYnrJMkbUsM75Ikbbv2BmYDV0REf0P7D4FHgJcAPwCuA94eEU8FrgKuzMwzx/heN1KcKn9L\nRFwKfI9ixvufAkTEM1qsRZKkbZKnzUuStO2aVf78EjDY8FgPTAWeUq4/ATgNmANcAPw2Iv4zIv6m\n1TfKzCXA84GfA28BrgaWR8RHWqxl9tg/niRJWw9H3iVJ2natKn+eCPxklPUrATLzzxTXl3+8vM/6\nEcDpwL9Q3IKuJZn5M2BeREwCDgHeBpwWEbfy6Oz4j1uLJEnbKsO7JEnbrv8CVgBzMvOCkcaI2Bm4\nBPhceZu524ALMvOzmfk74MKImAv8Q/mSoc29UUScDLwDiMzcAFwXEYsoJrvbjWLiusetBbi7lfeS\nJGlrZHiXJGkblZlDEfEhipBeowjQMylOkR8AbsnMR8qR8Y9ExAbgl8BewP8CvlFuamQE/8URcecm\n7id/LcXo/WURcSHF7d6OB9ZSXEO/2Voa3qsGvDYirs7MOzq2QyRJqjCveZckadtQH60xM78IHAUc\nDHyX4pr2BJ5XjrIDzKc4Rf5U4PvA+4ELgXeW21gFfIbiNnFXRsRj/r7IzNuAVwIzgK8B3wS2B15c\nXg/fai3/AVwDfKp8SJK0TajV66P+Xz6hImIKxR8B84A1wDmZee4m+h5FcR/ZXSm+hT+p8Rv+iFhF\n8cdArWyqA9tn5prx+wSSJEmSJI2fqoy8nw3sBxxKcQrd6RExr7lTRBwMfBn4CMXtbW4EvhcRU8v1\nT6EI7nMoTrEbAGYb3CVJkiRJvazr17yXwftY4LDMXAQsioizKG5Lc1lT9wHgjMz8WvnaM4BTKIL8\nzymuwVuWmUsnqn5JkiRJksZb18M7sC9FHTc2tF0PfKC5Y2ZeOvI8IrYDTgaWA78qm/cGnLhGkiRJ\nkrRVqUJ4nw08WN42ZsRyYLuImJWZK5pfEBEvBP69XDyq4bT4vYBpEXEdEMCtwImZeef4lS9JkiRJ\n0viqwjXvU4H1TW0jy1M28ZpfUFwj/2HgKxFxQNk+l2IW2zMoZrRdC1wTEdM6WrEkSZIkSROoCiPv\n63hsSB9ZHnWiucx8AHgAuD0ingO8DfgZcBgweWQkvpyZ/nfAEcDXWymmXq/Xa7Xa5jtKkiRJktQZ\nmw2hVQjv9wI7RkRfZg6XbQPA2vK+sX8REc8GhjLz1obmX1GcLk9mDgKDIysyc31E/BbYudViVq58\nhL4+w7skSZIkaWLMmLH5k8WrEN5vowjcBwE3lG2HADeN0vdYYA/g8Ia2Z1HMNE9E3EUxG/2Ccnka\n8HTgN60WMzxcZ3i4PsaPIEmSJEnS+Ol6eM/MtRGxALgoIuYDu1Dc/u0YgIjYCVidmeuALwELI+Kd\nwPeAo4H9gX8oN3cl8NGIWAo8CJwJ3ANcNYEfSZIkSZKkjqrChHVQ3PLtZuBa4ALgtMy8vFy3DHg9\nQHm6/GuANwOLKEbgX5qZ95d9TwUuBb4KLKT4fC/PTIfSJUmSJEk9q1avm2sbPfDAw+4QSZIkSdKE\nedKTtt/sxGtVGXmXJEmSJEmbYHiXJEmSJKniDO+SJEmSJFWc4V2SJEmSpIozvEuSJEmSVHGGd0mS\nJEmSKs7wLkmSJElSxRneJUmSJEmqOMO7JEmSJEkVZ3iXJEmSJKniDO+SJEmSJFWc4V2SJEmSpIoz\nvEuSJEmSVHGGd0mSJEmSKs7wLkmSJElSxRneJUmSJEmqOMO7JEmSJEkVZ3iXJEmSJKniDO+SJEmS\nJFWc4V2SJEmSpIozvEuSJEmSVHGGd0mSJEmSKs7wLkmSJElSxRneJUmSJEmquEndLkBS5wwODrJk\nyeJul1Epc+bsyeTJk7tdhiRJkrRFDO/SVmTJksWcd/EV7DBroNulVMKqFfdz0vwjiJjb7VIkSZKk\nLWJ4l7YyO8waYMeB3bpdhiRJkqQO8pp3SZIkSZIqzvAuSZIkSVLFGd4lSZIkSao4w7skSZIkSRVn\neJckSZIkqeIqMdt8REwBLgTmAWuAczLz3E30PQr4MLArcAtwUmbe1LD+SOBMYDbwfeAtmblifD+B\nJEmSJEnjpyoj72cD+wGHAscDp0fEvOZOEXEw8GXgI8DewI3A9yJiarn+gHL96cCBwAzgknGvXpIk\nSZKkcdT18F4G72OBd2Xmosy8HDgLOGGU7gPAGZn5tcy8GzgDmEkR5AHeAXwjM7+amb8EjgZeFhG7\nj/fnkCRJkiRpvHQ9vAP7Upy+f2ND2/UUI+cbycxLM/MTABGxHXAysBz4VdnlIODHDf1/D9xTtkuS\nJEmS1JOqEN5nAw9m5oaGtuXAdhExa7QXRMQLgT8BpwEnZuaahm3d19R9ObBLZ0uWJEmSJGniVCG8\nTwXWN7WNLE/ZxGt+QXGN/IeBr5TXuj/etja1HUmSJEmSKq8Ks82v47HhemR5DaPIzAeAB4DbI+I5\nwNuAnz3Otkbdzmj6+mr09dVa7S5VSn9/jVqteAhqtRr9/TUmTarC95SSJElS+6oQ3u8FdoyIvswc\nLtsGgLWZuaqxY0Q8GxjKzFsbmn8F7NWwrYGm7Q8Ay1otZubMaQYf9azp06fS199Hf79hFaCvv4/p\n06cyY8a0bpciSZIkbZEqhPfbgEGKSeVuKNsOAW4ape+xwB7A4Q1tzwJ+Xj5fCBwMLACIiF0prndf\n2GoxK1c+4si7etbq1WsYHhpmaGh48523AcNDw6xevYaHHnqk26WoNDg4yOLFd3W7jErZc8+nMXny\n5G6XIUmSuqiVwaauh/fMXBsRC4CLImI+Rdg+BTgGICJ2AlZn5jrgS8DCiHgn8D2KW8HtX/4E+AJw\nXUQspAj05wNXZObSVusZHq4zPFzvzIeTJtjQUJ16vXgI6vU6Q0N1Nmzwy4yquOOOOznv4ivYYVbz\nSVLbplUr7uek+UcQMbfbpUiSpIrrengvnQxcCFwLrAZOK+/3DsUp728EFmTmrRHxGuATwCeBXwIv\nzcxlAJm5MCLeCpwJzAC+Dxw3kR9EkvT4dpg1wI4Du3W7DEmSpJ5SifCemWuBN5WP5nV9TctXAVc9\nzrYWUJ42L0mSJEnS1sBZrSRJkiRJqjjDuyRJkiRJFWd4lyRJkiSp4gzvkiRJkiRVnOFdkiRJkqSK\nM7xLkiRJklRxhndJkiRJkirO8C5JkiRJUsUZ3iVJkiRJqjjDuyRJkiRJFWd4lyRJkiSp4gzvkiRJ\nkiRVnOFdkiRJkqSKM7xLkiRJklRxhndJkiRJkirO8C5JkiRJUsUZ3iVJkiRJqjjDuyRJkiRJFWd4\nlyRJkiSp4gzvkiRJkiRVnOFdkiRJkqSKM7xLkiRJklRxhndJkiRJkirO8C5JkiRJUsUZ3iVJkiRJ\nqjjDuyRJkiRJFWd4lyRJkiSp4gzvkiRJkiRVnOFdkiRJkqSKM7xLkiRJklRxhndJkiRJkirO8C5J\nkiRJUsUZ3iVJkiRJqrhJ3S4AICKmABcC84A1wDmZee4m+r4c+EfgacBi4LTMvKJh/Spge6BWNtWB\n7TNzzfh9AkmSJEmSxk9VRt7PBvYDDgWOB06PiHnNnSLib4FvAV8G9gW+BFwaEfuU659CEdznAAPl\nY7bBXZIkSZLUy7o+8h4RU4FjgcMycxGwKCLOAk4ALmvqfiRwTWZ+vly+MCJeCbwe+AWwF7AsM5dO\nTPWSJEmSJI2/rod3ihH0ScCNDW3XAx8Ype8lwBNGaZ9e/twbuKOTxUmSJEmS1G1VCO+zgQczc0ND\n23Jgu4iYlZkrRhozMxtfGBHPAF5Ecb08FCPv0yLiOiCAW4ETM/PO8fwAkiRJkiSNpyqE96nA+qa2\nkeUpm3pRROxIcf37f2bmd8vmucAM4H3Aw+XPayJir8x8pJVi+vpq9PXVNt9RqqD+/hq1WvEQ1Go1\n+vtrTJpUlek95DG6MY9RSZLUqiqE93U8NqSPLI860VxE7AT8gGIm+dc1rDoMmDwyQV1EHAX8DjgC\n+HorxcycOc0/KtWzpk+fSl9/H/39BgGAvv4+pk+fyowZ07pdikoeoxvzGJUkSa2qQni/F9gxIvoy\nc7hsGwDWZuaq5s4RsTNwLTAEHNp0Wv0gMNiwvD4ifgvs3GoxK1c+4si7etbq1WsYHhpmaGh48523\nAcNDw6xevYaHHmrpxBtNAI/RjXmMSpIkoKUv8qsQ3m+jCNwHATeUbYcANzV3LGemv7rs/4LMfKBp\n/V3AGZm5oFyeBjwd+E2rxQwP1xkerrfxMaTuGxqqU68XD0G9XmdoqM6GDQbFqvAY3ZjHqCRJalXX\nw3tmro2IBcBFETEf2AU4BTgG/nKK/OrMXAd8ENiD4n7wfeU6KEbp/whcCXw0IpYCDwJnAvcAV03g\nR5IkSZIkqaOqctHhycDNFKfDXwCclpmXl+uWUdzHHWAe8ETgp8B9DY/zy/XvAS4FvgospPh8L89M\nh3gkSZIkST2r6yPvUIy+A28qH83r+hqe77WZ7awHTi0fkiRJkiRtFaoy8i5JkiRJkjbB8C5JkiRJ\nUsW1ddp8REwCdgL6y6Yaxb3Z98/Mr3aoNkmSJEmSRBvhPSJeCiwAnjTK6rUUk8VJkiRJkqQOaee0\n+Y8DtwAvB9YArwFOBP4I/EPnSpMkSZIkSdBeeH8G8L7MvBq4DXgkMy+guN3buztZnCRJkiRJai+8\nDwGry+d3AX9TPr8W2LsTRUmSJEmSpEe1E95/CbyyfP5r4ODy+S4dqUiSJEmSJG2kndnmPwlcGhF/\nBr4GfDQirgT+lmL0XZIkSZIkddCYR94z8zvAAcDCzPwdcDiwAbgceEtny5MkSZIkSWMO7xHxYeA3\nmbkIIDN/lJmvAj4AnN7h+iRJkiRJ2ua1dNp8RMwFnlwung4sioiHmrrtAxxHcds4SZIkSZLUIa1e\n874ncAVQL5e/vYl+F29xRZIkSZIkaSMthffMvDIinkpxmv0SimveH2joUgf+lJkrO16hJEmSJEnb\nuJZnm8/MewAiYg/gnsysb+YlkiRJkiSpA8Z8q7jMXBoRr4yIfYD+srkGTAH2z8yXdLJASZIkSZK2\ndWMO7xHxSeA9wHKKSezuBXYqt/W1jlYnSZIkSZLGfqs44CjgxMycDdwHHAzMBn5CcT28JEmSJEnq\noHbC+07Ad8vntwMHlBPVfQB4Q6cKkyRJkiRJhXbC+0PAX5XP7wKeUT6/B9i5E0VJkiRJkqRHtRPe\nrwM+FRE7Az8FXhcROwKvZePbx0mSJEmSpA5oJ7yfCjwFeD1wKbCeYvK6TwPnda40SZIkSZIE7d0q\n7nfAMyNiu8z8c0QcAhwO/C4zb+p4hZIkSZIkbePGHN5HZOa68uca4LKOVSRJkiRJkjbSUniPiGGg\n3krfzOzfoookSZIkSdJGWh15n8+j4X134H3AF4EbgEFgf+AdwD92ukBJkiRJkrZ1LYX3zLxk5HlE\n/Ag4ITMvbujynYj4FXAicHZHK5QkSZIkaRvXzmzzBwA/HqX9Zzx6z3dJkiRJktQh7YT3O4E3jNJ+\nHPBfW1aOJEmSJElq1s5s86cDl0bES4CbKL4A+O/A3wF/38HaJEmSJEkSbYy8Z+a3gUOAe4HDgBcD\ndwAHZuZ1nS1PkiRJkiS1dZ/3zLyBYqZ5SZIkSZI0ztoK750WEVOAC4F5wBrgnMw8dxN9X05xS7qn\nAYuB0zLziob1RwJnArOB7wNvycwV4/sJJEmSJEkaP+1MWDcezgb2Aw4FjgdOj4h5zZ0i4m+BbwFf\nBvYFvkRx/f0+5foDynWnAwcCM4BLxr98SZIkSZLGT9dH3iNiKnAscFhmLgIWRcRZwAnAZU3djwSu\nyczPl8sXRsQrgdcDvwDeAXwjM79abvtoYGlE7J6ZSyfg40iSJEmS1HEtjbxHxNcjYsfy+fMiopOh\nf1+KLxFubGi7nmLkvNklwPtGaZ9e/jyIhnvQZ+bvgXvKdkmSJEmSelKrp82/GphZPr+O4nT0TpkN\nPJiZGxralgPbRcSsxo5Z+MXIckQ8A3gR8B8N27qvafvLgV06WK8kSZIkSROq1RH0XwA/jIgEasC3\nI+LPo3XMzBeOsYapwPqmtpHlKZt6UXkmwLeA/8zM725mW5vcTrO+vhp9fbVWu0uV0t9fo1YrHoJa\nrUZ/f41Jk6oyvYc8RjfmMSpJklrVanh/HfBOihH35wO/B9Z2qIZ1PDZcjyyvGe0FEbET8AOgXta2\nuW2Nup3RzJw5zT8q1bOmT59KX38f/f0GAYC+/j6mT5/KjBnTul2KSh6jG/MYlSRJrWopvGfm3cAp\nABGxB/C2zFzVoRruBXaMiL7MHC7bBoC1o71HROwMXAsMAYc23Qbu3vK1jQaAZa0Ws3LlI468q2et\nXr2G4aFhhoaGN995GzA8NMzq1Wt46KFHul2KSh6jG/MYlSRJQEtf5I954rnMfEFE1CLicGAfYBD4\nL+DazBwac5VwW7mNg4AbyrZDgJuaO5Yz019d9n9BZj7Q1GUhcDCwoOy/K8X17gtbLWZ4uM7wcH2M\nH0GqhqGhOvV68RDU63WGhups2GBQrAqP0Y15jEqSpFaNObxHxEzg+8CzgNUU18D/NXBzRLxkrCPy\nmbk2IhYAF0XEfIqwfQpwTPl+OwGrM3Md8EFgD4r7wfeV66AYpf8j8AXguohYCPwcOB+4wtvESZIk\nSZJ6WTsXHZ5NMTHc32XmjMzcAXgmsB3wiTbrOBm4meJ0+AuA0zLz8nLdMor7uAPMA54I/JRiVvmR\nx/kAmbkQeCtwOsXt5lYA89usSZIkSZKkSmjnfu1HAK/NzNtHGjJzUUS8E/g68PaxbjAz1wJvKh/N\n6/oanu/VwrYWUJ42L0mSJEnS1qCdkffJwP2jtN9Pcfq8JEmSJEnqoHbC+82MPrp+PHDrlpUjSZIk\nSZKatXPa/IcoJoV7DvCTsu1gYF/g8E4VJkmSJEmSCmMeec/MG4HnAb8FDqMI7EuAQzLzus6WJ0mS\nJEmS2hl5JzN/Bryhw7VIkiRJkqRRtHPNuyRJkiRJmkCGd0mSJEmSKs7wLkmSJElSxRneJUmSJEmq\nuLYmrIuI3YGDgCcAtcZ1mbmgA3VJkiRJkqTSmMN7RLwFuBDoH2V1HTC8S5IkSZLUQe2MvH8AuAj4\nYGb+scP1SJIkSZKkJu1c8z4bOMfgLkmSJEnSxGgnvN8GPKPThUiSJEmSpNG1c9r8WcDnI2IO8Btg\nfePKzPxxJwqTJEmSJEmFdsL7peXPz4yyrs7oE9lJkiRJkqQ2tRPe9+h4FZIkSZIkaZPGHN4zcylA\nRGwPzAUGgcWZ+XCHa5MkSZIkSbR3n/c+4GzgeGAyUAPWR8QXgZMys97ZEiVJkiRJ2ra1c9r8+4H5\nwHuAH1HMWP884HTgXuDTHatOkiRJkiS1Fd7fDByfmf+3oe3WiHgA+CiGd0mSJEmSOqqd+7zvBPx0\nlPafArtuWTmSJEmSJKlZO+H9DuDFo7S/BLh7i6qRJEmSJEmP0c5p8+cCX4yIOcBPyraDgROAd3eq\nMEmSJEmSVGjnVnELImIm8F7g1LJ5OfChzLywk8VJkiRJkqT2Rt7JzPOB8yPiSUAtM//Q2bIkSZIk\nSdKIlsJ7RPwv4BuZub583rz+L88zc0HnypMkSZIkSa2OvF8CXA38oXy+KXXA8C5JkiRJUge1FN4z\ns2+055IkSZIkafyNOYhHxLURMX2U9idHxK2dKUuSJEmSJI1o9Zr3lwHPLhcPBT4YEX9q6vZ04Kkd\nq0ySJEmSJAGtX/N+N/A5oFYuvwEYalhfB/7Eo7eOkyRJkiRJHdLqNe+/AuYARMRSYL/MXNGpIiJi\nCnAhMA9YA5yTmedu5jUHA1/JzD2b2lcB2/PoFw11YPvMXNOpeiVJkiRJmkjtTD73B2DnDtdxNrAf\nxSn5xwOnR8S8TXWOiH2Ab/JoQB9pfwpFcJ8DDJSP2QZ3SZIkSVIva/W0+UZ7AI90qoCImAocCxyW\nmYuARRFxFnACcNko/d8KfBpYDDRPnLcXsCwzl3aqPkmSJEmSuq2d8H4W8E8R8WngLmBt48rMvGeM\n29u3rOPGhrbrgQ9sov9hwNHADsDpTev2Bu4Y4/tLkiRJklRp7YT3jwH9wPMoricfUSuX+8e4vdnA\ng5m5oaFtObBdRMxqvrY+M+cBRMQxo2xrL2BaRFwHBHArcGJm3jnGmiRJkiRJqox2wvuLO1zDVGB9\nU9vI8pQxbmsuMAN4H/Bw+fOaiNgrM1s61b+vr0ZfX23zHaUK6u+vUasVD0GtVqO/v8akSe1M76Hx\n4DG6MY9RSZLUqjGH98z80cjziHgSMJiZq7aghnU8NqSPLI91ornDgMkjE9RFxFHA74AjgK+3soGZ\nM6f5R6V61vTpU+nr76O/3yAA0Nffx/TpU5kxY1q3S1HJY3RjHqOSJKlV7Yy8ExHvAt4PPLlcvh/4\ndGae38bm7gV2jIi+zBwu2waAtWP9UiAzB4HBhuX1EfFbxjA7/sqVjzjyrp61evUahoeGGRoa3nzn\nbcDw0DCrV6/hoYc6NsemtpDH6MY8RiVJEtDSF/ljDu/lbO9nUdyX/UcUt5t7HvCJiPhjZl48xk3e\nRhG4DwJuKNsOAW5qo7a7gDMyc0G5PA14OvCbVrcxPFxneLi++Y5SBQ0N1anXi4egXq8zNFRnwwaD\nYlV4jG7MY1SSJLWqnZH3k4B3Z+bnGtq+XQbnE4ExhffMXBsRC4CLImI+sAtwCnAMQETsBKzOzHUt\nbO5K4KMRsRR4EDgTuAe4aiw1SZIkSZJUJe1cdLg78L1R2q+mGOVux8nAzcC1wAXAaZl5ebluGfD6\nFrdzKnAp8FVgIcXne3lmOsQjSZIkSepZ7Yy8LwWeDSxuat+f4hZvY5aZa4E3lY/mdaN+wZCZXwG+\n0tT2Z4oAf2o7dUiSJEmSVEXthPcvAp+PiJnAT8q2g4EzgM90qjBJkiRJklRoJ7x/huLU+fPL19co\nJpz7IvCxzpUmSZIkSZKgvfu8DwMnRsRpwFyK8P7rzHy408VJkiRJkqT2JqwjIrYH3gAcDRwJvCoi\ntutkYZIkSZIkqTDm8B4Rc4E7gPOA5wIvoDhl/vaI2KWz5UmSJEmSpHZG3i8AbgV2zcxnZebfUVwD\nvxT4bCeLkyRJkiRJ7YX35wCnZuZDIw2Z+SDwbuDFnSpMkiRJkiQV2gnv9wOjnR7/18DKLStHkiRJ\nkiQ1a+dWcacCF0bEKcAPKW4Ttz9wIXB+ROw20jEz7+lEkZIkSZIkbcvaCe/fKn9eBtQb2mvAOeWj\nVq7r36LqJEmSJElSW+H9BR2vQpIkSZIkbdKYw3tm/mjkeUQ8CRjMzFUdrUqSJEmSJP1FOxPWERHv\niohlFJPXrYiIeyPixM6WJkmSJEmSoI3wHhFvBc4CvgbMA14L/CvwiYiY39nyJEmSJElSO9e8nwS8\nOzM/19D27Yi4CzgRuLgjlUmSJEmSJKC90+Z3B743SvvVwNO3rBxJkiRJktSsnfC+FHj2KO37A8u3\nrBxJkiRJktSsndPmvwh8PiJmAj8p2w4GzgA+06nCJEmSJElSoZ3w/hmKU+fPL19fAwYpQv3HOlea\nJEmSJEmC9sL7c4FTgdOAuRTh/deZ+XAnC5MkSZIkSYV2wvu3gMMz8xbgpg7XI0mSJEmSmrQzYd0D\nwPROFyJJkiRJkkbXzsj7VcCVEXEVcCewtnFlZp7RicIkSZIkSVKhnfD+Wopbwj2rfDSqU8w6L0mS\nJEmSOmTM4T0z9xiPQiRJkiRJ0uhaDu8RsQvwGmAdcFVm3jtuVUmSJEmSpL9oacK6iDgESIp7vH8R\n+HVEvHQ8C5MkSZIkSYVWZ5s/E/gPYGdgALgaOHe8ipIkSZIkSY9qNbw/E3h/Zi7LzD8AJwF7RcT2\n41eaJEmSJEmC1sP7XwErRhbK693/DMwcj6IkSZIkSdKjWg3vNYrbwDXaAPR3thxJkiRJktSs1fAu\nSZIkSZK6ZCz3eT8lIh5pWJ4MvCsiVjZ2yswzxlpEREwBLgTmAWuAczLzcSfEi4iDga9k5p5N7UdS\nTLA3G/g+8JbMXDHKJiRJkiRJ6gmthvd7gNc3tS0DXtXUVgfGHN6Bs4H9gEOBpwILIuLuzLxstM4R\nsQ/wTWBtU/sBwJeB44BFwAXAJcARbdQkSZIkSVIltBTeM/Op41VAREwFjgUOy8xFwKKIOAs4AXhM\neI+ItwIK6DvRAAAchklEQVSfBhYD05tWvwP4RmZ+tex7NLA0InbPzKXj9RkkSZIkSRpPVbjmfV+K\nLxFubGi7HjhwE/0PA44Gzh9l3UHAj0cWMvP3FGcNHNSRSiVJkiRJ6oKxXPM+XmYDD2bmhoa25cB2\nETGr+Xr1zJwHEBHHbGJb9zW1LQd26WC96qDBwUGWLFnc7TIqY86cPZk8eXK3y5AkSZJUMVUI71OB\n9U1tI8tTOrStsW5HE2TJksWcd/EV7DBroNuldN2qFfdz0vwjiJjb7VIkSZIkVUwVwvs6HhuuR5bX\ndGhbLW+nr69GX19tjG+rdvX315ix42x2HNit26V0Xa1Wo7+/xqRJ7V/N0t9fo1YrHurMPlVneYxu\nzGNUkiS1qgrh/V5gx4joy8zhsm0AWJuZq9rYVvMQ7gDFzPgtmTlzmn9UTqDp06fS199Hf79/uPb1\n9zF9+lRmzJjW9jbcnxvrxD5VZ3mMbsxjVJIktaoK4f02YJBiUrkbyrZDgJva2NZC4GBgAUBE7Epx\nvfvCVjewcuUjjrxPoNWr1zA8NMzQ0PDmO2/lhoeGWb16DQ899Ejb23B/bqwT+1Sd5TG6MY9RSZIE\ntPRFftfDe2aujYgFwEURMZ8ibJ8CHAMQETsBqzNzXQub+wJwXUQsBH5OMSP9FWO5TdzwcJ3h4fpY\nP4baNDRUp14vHtu6er3O0FCdDRvaDzXuz411Yp+qszxGN+YxKkmSWlWV8xZPBm4GrgUuAE7LzMvL\ndcuA17eykcxcCLwVOJ3idnMrgPkdr1aSJEmSpAnU9ZF3KEbfgTeVj+Z1o37BkJlfAb4ySvsCytPm\nJUmSJEnaGlQivEtSVQ0ODrJkyeJul1Epc+bsyeTJk7tdhiRJ0jbF8C5Jj2PJksWcd/EV7DCr+UYW\n26ZVK+7npPlHEDG326VIkiRtUwzvkrQZO8waYMeB3bpdhiRJkrZhVZmwTpIkSZIkbYLhXZIkSZKk\nijO8S5IkSZJUcYZ3SZIkSZIqzvAuSZIkSVLFGd4lSZIkSao4w7skSZIkSRVneJckSZIkqeIM75Ik\nSZIkVZzhXZIkSZKkijO8S5IkSZJUcYZ3SZIkSZIqzvAuSZIkSVLFGd4lSZIkSao4w7skSZIkSRVn\neJckSZIkqeIM75IkSZIkVZzhXZIkSZKkijO8S5IkSZJUcYZ3SZIkSZIqblK3C5AkSaqSwcFBlixZ\n3O0yKmPOnD2ZPHlyt8uQpG2e4V2SJKnBkiWLOe/iK9hh1kC3S+m6VSvu56T5RxAxt9ulSNI2z/Au\nSZLUZIdZA+w4sFu3y5Ak6S+85l2SJEmSpIpz5F2SJEnjxjkEHst5BCS1w/AuSZKkceMcAhtzHgFJ\n7TK8S5IkaVw5h4AkbTmveZckSZIkqeIM75IkSZIkVZzhXZIkSZKkiqvENe8RMQW4EJgHrAHOycxz\nN9H3mcAXgH2AXwJvz8xbGtavArYHamVTHdg+M9eM3yeQJEmSJGn8VGXk/WxgP+BQ4Hjg9IiY19wp\nIqYCVwI/KvvfCFwZEU8s1z+FIrjPAQbKx2yDuyRJkiSpl3V95L0M5McCh2XmImBRRJwFnABc1tT9\nDcCazHxvuXxiRLwMeB2wANgLWJaZSyemekmSJEmSxl8VRt73pfgS4caGtuuBA0fpe2C5rtFPgOeU\nz/cG7uh0gZIkSZIkdVMVwvts4MHM3NDQthzYLiJmjdL3vqa25cAu5fO9gGkRcV1E3BcRV0bE08el\nakmSJEmSJkjXT5sHpgLrm9pGlqe02Hek31xgBvA+4OHy5zURsVdmPtJKMX19Nfr6apvvqI7o769R\nqxWPbV2tVqO/v8akSe1/p+b+3Jj7tPO2dJ+6PzfWiWNUnedx+ih/j3ae/+4ltasK4X0djw3pI8vN\nE81tqu9Iv8OAySMT1EXEUcDvgCOAr7dSzMyZ0/zPZQJNnz6Vvv4++vv9D6yvv4/p06cyY8a0trfh\n/tyY+7TztnSfuj831oljVJ3ncfoof492nv/uJbWrCuH9XmDHiOjLzOGybQBYm5mrRuk70NQ2ACwD\nyMxBYHBkRWauj4jfAju3WszKlY848j6BVq9ew/DQMENDw5vvvJUbHhpm9eo1PPRQSyeJjMr9uTH3\naedt6T51f26sE8eoOs/j9FH+Hu08/91LGk0rX+hVIbzfRhG4DwJuKNsOAW4ape9C4L1Nbc8FzgSI\niLuAMzJzQbk8DXg68JtWixkerjM8XB913eDgIEuWLG51U9uEOXP2ZPLkyW2/fmioTr1ePLZ19Xqd\noaE6Gza0/8eN+3Nj7tPO29J96v7cWCeOUXWex+mj/D3aef67l9Surof3zFwbEQuAiyJiPsXkc6cA\nxwBExE7A6sxcB1wKfCIizgO+BLyN4jr4b5abuxL4aEQsBR6kCPX3AFd1otYlSxZz3sVXsMOs5sH/\nbdOqFfdz0vwjiJjb7VIkSZIkaavW9fBeOhm4ELgWWA2clpmXl+uWAW8EFmTmwxHxCuCLwHHA7cDf\nZ+basu+pwJ+BrwLTgWuAl2dmx77q3WHWADsO7NapzUmSJEmStFmVCO9l+H5T+Whe19e0/HPgWZvY\nzp8pAvyp41CmJEmSJEld4bSfkiRJkiRVnOFdkiRJkqSKM7xLkiRJklRxhndJkiRJkirO8C5JkiRJ\nUsUZ3iVJkiRJqjjDuyRJkiRJFWd4lyRJkiSp4gzvkiRJkiRVnOFdkiRJkqSKM7xLkiRJklRxhndJ\nkiRJkirO8C5JkiRJUsUZ3iVJkiRJqjjDuyRJkiRJFWd4lyRJkiSp4gzvkiRJkiRVnOFdkiRJkqSK\nM7xLkiRJklRxhndJkiRJkirO8C5JkiRJUsUZ3iVJkiRJqjjDuyRJkiRJFWd4lyRJkiSp4gzvkiRJ\nkiRVnOFdkiRJkqSKM7xLkiRJklRxhndJkiRJkipuUrcLkCRJktS6wcFBlixZ3O0yKmPOnD2ZPHly\nt8uQxp3hXZIkSeohS5Ys5ryLr2CHWQPdLqXrVq24n5PmH0HE3G6XIo07w7skSZLUY3aYNcCOA7t1\nuwxJE8hr3iVJkiRJqrhKjLxHxBTgQmAesAY4JzPP3UTfZwJfAPYBfgm8PTNvaVh/JHAmMBv4PvCW\nzFwxvp9AkqTu8NrXx/L6V0nS1qgS4R04G9gPOBR4KrAgIu7OzMsaO0XEVOBK4F+AY4C3A1dGxJzM\nXBsRBwBfBo4DFgEXAJcAR0zMx5AkaWJ57evGvP5VkrS16np4LwP5scBhmbkIWBQRZwEnAJc1dX8D\nsCYz31sunxgRLwNeBywA3gF8IzO/Wm77aGBpROyemUsn4ONIkjThvPZVkqStXxWued+X4kuEGxva\nrgcOHKXvgeW6Rj8BnlM+Pwj48ciKzPw9cE/ZLkmSJElST6pCeJ8NPJiZGxralgPbRcSsUfre19S2\nHNilxfWSJEmSJPWcrp82D0wF1je1jSxPabHvlBbXb1ZfX42+vtqo6/r7a6xeeT+12ujrtzWrV95P\nf3+NSZPa/w7Iffoo92fnuU87b0v3qftzYx6jnec+7Sz3Z+e5TzurE/sT4De/+XWHKto6zJ271xZv\nw336qE7sT4BavV7vyIbaFRGvBT6bmU9paJsL/BcwKzNXNbT/G3B7Zn6goe2TwNzMfHVE/AmYl5n/\n3rB+IfD1zDx/Aj6OJEmSJEkdV4XT5u8FdoyIxloGgLWNwb2hb/N0ugPAshbXS5IkSZLUc6oQ3m8D\nBtl4UrlDgJtG6bsQ+O9Nbc/l0cnuFgIHj6yIiF0prndf2KliJUmSJEmaaF0/bR4gIr5AEcLnU4Tt\nS4BjMvPyiNgJWJ2Z6yJie+BO4GvAl4C3Aa8Fnlbe5/0g4DqKW8b9HDi/fO1rJvozSZIkSZLUKVUY\neQc4GbgZuBa4ADgtMy8v1y0DXg+QmQ8DrwCeRxHODwD+PjPXlusXAm8FTqe4pdwKii8EJEmSJEnq\nWZUYeZckSZIkSZtWlZF3SZIkSZK0CYZ3SZIkSZIqzvAuSZIkSVLFGd4lSZIkSao4w7skSZIkSRU3\nqdsFaGwi4inAZ4EXAGuAfwXen5l/7mphPSoi9gQ+DzyX4taCn8vMs7tb1dYhIq4Elmemt2vcAhHx\nauAyoA7Uyp/fyszXd7WwHhYRTwDOA44E1gMXZ+YHu1tV74qIY4B/ZuNjtAYMZ6Z/Z7QhInYBvkBx\na9wVwGcy8zPdraq3RcSTKPbpi4AHgI9l5le6W1XviYgpFLdrfkdm/rhseyrwf4DnAHcDJ2XmD7pV\nY68ZbZ82rAvg55m5fVeK60GbOEYPAs4B/hb4PXB2Zv5T96psnyPvvedbwHYUYfMNwBHAmV2tqEdF\nRA24ElgO/B3wNuBDEfGGrha2FSj34d93u46txN7Ad4GB8jEbeHNXK+p9n6X4A/4lwP8E3hIRb+lu\nST3t6zx6bA4AuwN3Aed3s6ge903gYWA/4ETgYxHxqu6W1PO+AzwFeD7FPj23/HJULSpD0dco/l9q\n9B3gPuBZwP8PfLv8Akqb8Tj7lIjYHbgCmDLRdfWq0fZnROwEXAVcS/H3/keACyKiJ/9O9RvxHlJ+\n+3YAsFNmPli2fRj4NPDebtbWo3YCbgWOz8xHgMURcQ1wMMUfo2pDRMwAzgJ+1u1athJ7Ab/MzAe6\nXcjWoDw+5wMvzMyby7azgQMpRo40Rpm5HvjDyHJEvL98+v7RX6HHExE7UByPx2bmYor/m66m+MLp\n8q4W16Mi4lnAQcCczFwK3B4RnwLeQxE8tRkRsRfwf0dpfyEwBzgoM9cBn4yIF1H8nj1jYqvsLZva\np+W6/4/iTJF7J7SoHvY4+/PVwLLMPK1cXhwRL6D48v57E1Vfpzjy3lvuBw4fCe6lGjC9S/X0tMy8\nPzOPLIM7EfFcilMUr+tuZT3vbGAB8OtuF7KV2Bu4o9tFbEUOBlZl5vUjDZl5VmZ6NkMHlF+OvAd4\nb2YOdrueHrUWeAR4U0RMKr+4fy5wS3fL6mlzgAfK4D7iduBZEdHfpZp6zfOBayhOja81tB8I3FIG\n9xHXl/30+Da1TwFeBrwPePdEF9XDNrU/vwe8aZT+PZmfHHnvIZm5GvjLNUTlad8nAP/RtaK2EhFx\nN7Ar8G8U1xerDeU38IcA+wAXdbmcrUUAh0fEB4F+itNpP2wwatsc4O6IOBr4APAEiuu1P5aZ9a5W\ntnU4Hrg3M7/d7UJ6VWauj4gTgM9RnN7dD/xzZl7S1cJ623Jgh4jYriFk7kbxd/B0YGXXKusRmfmX\n/9OL75P+YjbFKfONlgOeNr8Zj7NPycxjy/YXTXBZPWtT+zMz7wHuaVj3ZIpLjz88kfV1iiPvve3T\nFNduONHSlptHMX/AM/E6zbaU1xldRHEZwvpu17M1iIjdgCdSjMS9DjgFOIrisgS156+A/wYcB7yR\nYp++iyIkacsdSzGngLbMXhRzXRxAcZy+NiKO7GpFve2nwDLgcxExNSKeBpxUrntC98raKkylmPiz\n0Xq8TlsVFBHbUcwfdh/wpS6X0xbDe48qr9V6F3BUZnp68hbKzFsy8yqK/8yPiwjPShm7jwA3ZaZn\ngnRI+W3xrMw8NjNvz8zLKULmceWZNxq7DcD2wJGZ+dPM/A7wMeCt3S2r90XE/sDOwDe6XUsvK0fa\njgXmZ+atmbkA+BTwoe5W1rvKL5RfS3Gnnj8CP+LRs8P+2K26thLreGxQn0JxRySpMiJiGsVE1U8D\nXtF0qUfPMLz3oIi4gCJkHlX+4ak2RMSTR5m991cU38L/dRdK6nX/A3h1RDwcEQ9TjBD/Q0T4h9EW\nyMxVTU2/prjjxMwulLM1WAasy8zfN7QlxWUz2jKHAT8uL/FS+/YD7mw6g+lWiln81abMvDkz96SY\ncX5XirlEHsxMQ+aWuZfiLhONBih+10qVEBHbA/9OMY/QCzJzSZdLapvhvcdExOkUp3v+j8z8Zrfr\n6XF7AJdFxOyGtmdTTGrj9W9j93yKa933LR/fpZgZed9uFtXLIuKlEfFgeZrXiGcCKzJzRbfq6nEL\nge3K02ZH7E1xb2JtmQOBn3S7iK3AfcDTms4A2wv4bZfq6XkRMSMi/jMiZmTmHzJzGHgF8MMul7Y1\nWAjsV146N+Lgsl3quvJMxW8DTwWel5m/6W5FW8ZTg3tIeQuEDwEfB24o71sIQGYu71phvesm4OfA\nxRFxMkWYPwv4x65W1aMy83eNy+Xoez0z/YOzfTdQnHr45Yg4A9iT4hj9VFer6mGZeUdEXAlcEhHH\nU0y29F68pVEn/A3wL90uYitwBcW/8y9HxMeAuRS33fPWe23KzIfKU2bPioiPU9x2740UE6xqy/wI\n+B3F79QzgVcC+1PsX6kK3gwcyv9r796DrazKOI5/EUXGmXJ0FDUztNRHNDRUyrzgdbJ0Eq8xiKAk\nQV7IGq0UtdLRDBUkbnknzQTxkrehKS+MNxTMG2b6KwQTw7wn3kE9/fG8W96z2Ydz4TBnw/l9Zvaw\n937fd631vufMcJ61nrVWrm21uBQ/LZH0Voe1qo088r56OYT8mZ1F9swvItOSqlf5tBYoet77k1vy\nzCIXrhgnaWKHNsysIOldMhV5Y7Kz6QrgUkljOrRhq79BwDzgAeD3wHhJkzq0RWuGHsBq94dQvZG0\nmAwuNwPmAGOAcyVd2aENW/0NIOe6ziXXDDpSkrffa5vPduYo/S21KTkgcjRwaNXUJGuedztpXw0s\ne6aHk1vH3cmy+GkRuXDdaqdLQ4N/V8zMzMzMzMzqmUfezczMzMzMzOqcg3czMzMzMzOzOufg3czM\nzMzMzKzOOXg3MzMzMzMzq3MO3s3MzMzMzMzqnIN3MzMzMzMzszrn4N3MzMzMzMyszjl4NzMzMzMz\nM6tzDt7NzMzMzMzM6tzaHd0AMzOzjhARLwCfAr0lvVd1bArQU9J+q7D+BcAUSeeuqjpa0ZZdgOuA\nLYEJkn62gnPPB84ATpE0oQVlHwdcLalNAwYr+5yKn+WxQAPQBfgEeAuYBfxC0tzSuTOBBZK+35a6\nqurtCSwA9pF0/8qWZ2Zm5pF3MzPrrBqAnsBFHd2QOjAK+BDoBVzQ1EkR0QUYDDwHjGhh2Q3Fq612\nBS5eieshA/VNi9dWQH9yAOOhiNipdN5hwCkrWVfZyty3mZlZIx55NzOzzmw+MCIibpJ0b0c3pgNt\nADwp6YVmzjsQ2JwMfm+PiD0lPbgqGybpjXYoZomk10qfX4qIQ4FHgAlAv6Ku/7VDXWVd2rk8MzPr\nxBy8m5lZZ3YdsAdwdUTsUJ0+XxERnwLHSbq21ncR8UtgT+B+4CRgPeB64Dzgd8B+wCIy1XxGqegv\nRMQMYF/gZeBiSZNLdexOjoT3BV4D7gDOkPROcXwBcBNwELAxcISkB2q0/2DgLOCrwDvAVGCUpI+K\nMr4EdImIIcBWkl5s4nkNBZ6WdGdELAR+CDQK3iPiMOAcYBtgDnBP1fG+wBigD7AUuBf4iaSFtSos\np82XnvNdwEhgI2A2MEKSmmhzTZI+johJwJURsbmk/5TT5iNiLfLZDwR6kCnw4yRdVrRrCrAu8AYw\nhMxc+AP581la4z66AecDR5AdIO8CdwMnSnojIm4B1pe0f+maAJ4FdpD0bGvuz8zM1jxOmzczs86s\nATieHHkes5Jl9QOCDC5HAsPJ4HUqsDMZhE2pumYYcB/QGxgL/DYi+gNExI5kkDqDDLoHFuX8taqM\nk4CTgW+TI8mNFMH0bcDtZMA8HBhQtAsyLf0R4AYyrbypIHoD4BBgevHVdOCIiNiwdM7uZGfC9OKe\nrgFOLx1fC7gTmAnsQHZqbAFcVavOJuxFPuPvkB0vPYBJrbi+7GlydHynGsdOIgPto8iOiAnA5OIe\nKw4HNgN2I3+PhgDjmqjrQjItfwiwdfHv/sCZxfEpwN4RsXnpmmOBOQ7czcwMPPJuZmadnKSFEXEa\ncFmRPn93G4vqAgyX9D4wLyIuAu6WdD1AREwGDo6ITSS9UlzzJ0mji/cTI2I34FQy2P4p8JfS8fkR\nMQh4PiL6lRZBmyFp5gra9XPgZkmVuezzIuJE4NaI2E7ScxGxBPigKrW82iCgGzCt+Dy1aOtQlnV8\nnAw8KOm8Ul29gR8Vnz9Pjpa/DCyU9GJEDCAD8JZaGzhG0mKAiLgUGL3iS5pUSZNfv8axLwPvAf+W\n9F8ycH8O+GfpnLeAQZI+Ap6NiLOBcRFRa8G/OcCNkh4qPi+MiLvITg7ITppXgWOA0cX6AseQo/Vm\nZmYeeTczM5N0BTnKfWVEfK6NxbxSBO4V75Fz6is+KP5dt/TdQzQ2mxxlhxwlPygi3qm8gLlktkCv\n0jX/aqZdvWvUc1/pWEsNBR6XNB9A0uNF3cOr6nq06rpZlTfFnPLR5Ej5axFxA7A3OQLeUq9UAvfC\n22SnQltUgvZac90nkZ0NL0XEoxHxa+B1Sa+XzpldBO4Vs4q2RHVhRSdO94i4ICJujohnyFH9rsXx\nT8i0+8HFJfuTHR3TqssyM7POycG7mZlZGkYGc2ObOzEiutb4erl5zuRWdCvySdXnrkAlGFwL+COw\nI5nWXXltQ86nr/iAFau1aFrl//9abV5OkcLfB+gTEUsrL+ArwNYRUZmn3cDyf1s0qkPSKHKV/1FF\n2yYAj0bEOi1pC8ueT3vYlWzzk9UHJM0j09sPJOftHww8ERGDS6dVP7+uLNuOrpEiQ2AasA6ZWTGQ\nZVMXKq4GekVEHzLT4TZJb7f+tszMbE3k4N3MzIxMnwdOI+cu71V1eCk5CluxbTtVu0vV571YNgr9\nd2B7SQskzS9GvLuRc6q3aEUdc8k54mX9yKD1Hy0s43hgCTnHvNyRsAf5bCrbxj0J7F51bd/Km4jY\ntpg+8KqkyyV9j5yrvz21552vMsX8+xHATEkv1zg+EjhS0j2STpe0ExnEDyidtnOR3l6xB5lx0Wjx\nvGJdgOHACZJOk3Rtsb98L0qdK8Wiew8XdfRn+TUSzMysE/OcdzMzs4KkqyLiKOBbQHnF9YeBH0TE\nA2TH91hydfHWqh4FHxgRc8lF3A4jA7Z9i2NjgPsjYiIwkVxUbxLQncbzrptzITA9Is4kF5ILcrT7\nDknNllOMiB9NzteeXeP4VODoiOhRtHl2Md//cuDr5MJvFa+TI87rRcRvyMyEocCb5N7xq0q3iNik\neL8Oudf7qWTmwOAmrtkYODsi3geeIgPtrwGXlM7ZkpwLP45cgO9XwHhJH+ZC8Z9ZTKb3HxoRT5C7\nEYwkFyCsXmRwCjCefCZ3tfZGzcxszeWRdzMz66wamvh+GDkHunz8BDKYehi4kQxMX2pD+Q1V7y8i\n07GfAo4DBla2eisC5QPJEenHgFvJFesPkPRxM/fwGUm3kAHzUeQo/GQyHX/Aiq4r+S6wIdmBUMtY\ncjBgmKSnyFXg9ynu6cfkdnmVtrxJjrT3JJ/lY8X7AyS920T5DbTgPpvxTXKrvkXkOgTTyC3e+kp6\npolrziFXwR9PjqRfSnae/KZ0ziNkivzfyIyISySdWTreALktHXAkuZ7BXHJxuu7AGcD2EdG9dM0N\nZCfPNZJW9r7NzGwN0qWhwf8vmJmZmbVGsc97T0n7tXO5W5GZFdtJer49yzYzs9Wb0+bNzMzMOlhE\nfBH4Bpnl8WcH7mZmVs1p82ZmZmYdbyNyvvv6NF4nwMzMDHDavJmZmZmZmVnd88i7mZmZmZmZWZ1z\n8G5mZmZmZmZW5xy8m5mZmZmZmdU5B+9mZmZmZmZmdc7Bu5mZmZmZmVmdc/BuZmZmZmZmVuccvJuZ\nmZmZmZnVOQfvZmZmZmZmZnXu/4hvu92+McG2AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115435990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_ads = clicks_test.groupby(\"display_id\")[\"ad_id\"].count().value_counts()\n",
    "num_ads = num_ads / np.sum(num_ads)\n",
    "plt.figure(figsize=(12,4))\n",
    "sns.barplot(num_ads.index, num_ads.values, alpha=0.8, color=p[0])\n",
    "plt.xlabel(\"Number of Ads in Display\")\n",
    "plt.ylabel(\"Proportion of data\")\n",
    "plt.title(\"Test set\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can say that the distributions of the number of ads in display are the same for both training set and test set. The most frequency is 4 or 6 ads, the maximum are both 12. That is the reason why the evaluation uses $MAP@12$."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Number of appearance, number of clicks and click rates for ads"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "num_appr = clicks_train.groupby(\"ad_id\")[\"ad_id\"].count()\n",
    "num_click = clicks_train.groupby(\"ad_id\")[\"clicked\"].sum()\n",
    "click_rate = (num_click + 0.0)/num_appr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean: 181.943273828\n",
      "max: 211824\n",
      "min: 1\n",
      "median: 5.0\n",
      "std 1910.68757307\n",
      "Ads that had appeared less than 2 times: 22.69%\n",
      "Ads that had appeared less than 10 times: 61.74%\n",
      "Ads that had appeared less than 50 times: 86.93%\n",
      "Ads that had appeared less than 100 times: 91.04%\n",
      "Ads that had appeared less than 1000 times: 97.45%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAMAAAGJCAYAAAD/rfo3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x126716b50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,4))\n",
    "plt.hist(num_appr, bins = 100, log = True)\n",
    "plt.xlabel(\"Number of appearance of ads\")\n",
    "plt.ylabel(\"Log(Count of ad_id)\")\n",
    "plt.title(\"Number of apperance\")\n",
    "print \"mean:\", num_appr.mean()\n",
    "print \"max:\", num_appr.max()\n",
    "print \"min:\", num_appr.min()\n",
    "print \"median:\", np.median(num_appr)\n",
    "print \"std\", np.std(num_appr)\n",
    "for i in [2, 10, 50, 100, 1000]:\n",
    "    print('Ads that had appeared less than {} times: {}%'.format(i, round((num_appr < i).mean() * 100, 2)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean: 35.2324731183\n",
      "max: 44824\n",
      "min: 0\n",
      "median: 1.0\n",
      "std 422.647158861\n",
      "Ads that had been clicked less than 1 times: 46.94%\n",
      "Ads that had been clicked less than 2 times: 65.57%\n",
      "Ads that had been clicked less than 3 times: 73.85%\n",
      "Ads that had been clicked less than 10 times: 88.7%\n",
      "Ads that had been clicked less than 100 times: 96.59%\n",
      "Ads that had been clicked less than 1000 times: 99.31%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA/8AAAGJCAYAAADc5oOjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1266f9dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,4))\n",
    "plt.hist(num_click, bins = 100, log = True)\n",
    "plt.xlabel(\"Number of clicks\")\n",
    "plt.ylabel(\"Log(Count of ad_id)\")\n",
    "plt.title(\"Number of clicks\")\n",
    "print \"mean:\", num_click.mean()\n",
    "print \"max:\", num_click.max()\n",
    "print \"min:\", num_click.min()\n",
    "print \"median:\", np.median(num_click)\n",
    "print \"std\", np.std(num_click)\n",
    "for i in [1, 2, 3, 10, 100, 1000]:\n",
    "    print('Ads that had been clicked less than {} times: {}%'.format(i, round((num_click < i).mean() * 100, 2)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean: 0.143224101154\n",
      "max: 1.0\n",
      "min: 0.0\n",
      "median: 0.051724137931\n",
      "std 0.217684217385\n",
      "Ads that had click rate less than 0.0001: 46.94%\n",
      "Ads that had click rate less than 0.01: 47.0%\n",
      "Ads that had click rate less than 0.1: 57.23%\n",
      "Ads that had click rate less than 0.2: 72.61%\n",
      "Ads that had click rate less than 0.5: 91.9%\n",
      "Ads that had click rate less than 0.99999: 96.73%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA/UAAAGJCAYAAADLxBNqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBwAAgChRDwAAAFGiHgAAAKJEPQAAAEQN1jXkwBcZANgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12de76210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,4))\n",
    "plt.hist(click_rate, bins = 100, log = True)\n",
    "plt.xlabel(\"Click rate\")\n",
    "plt.ylabel(\"Log(Count of ad_id)\")\n",
    "plt.title(\"Click rate\")\n",
    "print \"mean:\", click_rate.mean()\n",
    "print \"max:\", click_rate.max()\n",
    "print \"min:\", click_rate.min()\n",
    "print \"median:\", np.median(click_rate)\n",
    "print \"std\", np.std(click_rate)\n",
    "for i in [0.0001, 0.01, 0.1, 0.2, 0.5, 0.99999]:\n",
    "    print('Ads that had click rate less than {}: {}%'.format(i, round((click_rate < i).mean() * 100, 2)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are about 46.94% of the ads that had never been clicked, and about 3.27% has click rate 1, which probably had only appeared once or twice. Regularization is needed."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Promoted Content"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "promoted_content = pd.read_csv('promoted_content.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Is ad_id in clicks_train a subset of ad_id in promoted_content?  True\n",
      "Is ad_id in clicks_test a subset of ad_id in promoted_content?  True\n",
      "Is ad_id in clicks_test a subset of ad_id in clicks_train?  False\n"
     ]
    }
   ],
   "source": [
    "a = set(clicks_train[\"ad_id\"])\n",
    "b = set(clicks_test[\"ad_id\"])\n",
    "c = set(promoted_content[\"ad_id\"])\n",
    "print \"Is ad_id in clicks_train a subset of ad_id in promoted_content? \", a.issubset(c)\n",
    "print \"Is ad_id in clicks_test a subset of ad_id in promoted_content? \", b.issubset(c)\n",
    "print \"Is ad_id in clicks_test a subset of ad_id in clicks_train? \", b.issubset(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "promoted_content[\"num_appr\"] = num_appr\n",
    "promoted_content[\"num_click\"] = num_click\n",
    "promoted_content[\"click_rate\"] = click_rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "promoted_content = promoted_content.fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# promoted_content[\"num_appr\"] = promoted_content[\"num_appr\"] + 100\n",
    "# promoted_content[\"num_click\"] = promoted_content[\"num_click\"] + 2\n",
    "# promoted_content[\"click_rate\"] = (promoted_content[\"num_click\"] + 0.0) / promoted_content[\"num_appr\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 300,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ad_id</th>\n",
       "      <th>document_id</th>\n",
       "      <th>campaign_id</th>\n",
       "      <th>advertiser_id</th>\n",
       "      <th>num_appr</th>\n",
       "      <th>num_click</th>\n",
       "      <th>click_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6614</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>471467</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>7692</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.045455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>471471</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>161.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.068323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>471472</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>32.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.125000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>12736</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>12808</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>471477</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>3406.0</td>\n",
       "      <td>985.0</td>\n",
       "      <td>0.289196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>13379</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>28.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.142857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>13885</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>14230</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>81.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.111111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>446701</td>\n",
       "      <td>10</td>\n",
       "      <td>19</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>13</td>\n",
       "      <td>471499</td>\n",
       "      <td>10</td>\n",
       "      <td>19</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.166667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>14</td>\n",
       "      <td>471500</td>\n",
       "      <td>10</td>\n",
       "      <td>19</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.125000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>471501</td>\n",
       "      <td>10</td>\n",
       "      <td>19</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.100000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>471514</td>\n",
       "      <td>17</td>\n",
       "      <td>19</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>17</td>\n",
       "      <td>471517</td>\n",
       "      <td>10</td>\n",
       "      <td>19</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18</td>\n",
       "      <td>471518</td>\n",
       "      <td>10</td>\n",
       "      <td>19</td>\n",
       "      <td>20.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.100000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>471519</td>\n",
       "      <td>5</td>\n",
       "      <td>19</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>20</td>\n",
       "      <td>446660</td>\n",
       "      <td>21</td>\n",
       "      <td>19</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ad_id  document_id  campaign_id  advertiser_id  num_appr  num_click  \\\n",
       "ad_id                                                                        \n",
       "1          1         6614            1              7       0.0        0.0   \n",
       "2          2       471467            2              7       2.0        0.0   \n",
       "3          3         7692            3              7      22.0        1.0   \n",
       "4          4       471471            2              7     161.0       11.0   \n",
       "5          5       471472            2              7      32.0        4.0   \n",
       "6          6        12736            1              7       1.0        0.0   \n",
       "7          7        12808            1              7       1.0        0.0   \n",
       "8          8       471477            2              7    3406.0      985.0   \n",
       "9          9        13379            1              7      28.0        4.0   \n",
       "10        10        13885            1              7       3.0        0.0   \n",
       "11        11        14230            1              7      81.0        9.0   \n",
       "12        12       446701           10             19      11.0        0.0   \n",
       "13        13       471499           10             19      12.0        2.0   \n",
       "14        14       471500           10             19       8.0        1.0   \n",
       "15        15       471501           10             19      10.0        1.0   \n",
       "16        16       471514           17             19       2.0        0.0   \n",
       "17        17       471517           10             19       3.0        1.0   \n",
       "18        18       471518           10             19      20.0        2.0   \n",
       "19        19       471519            5             19       9.0        0.0   \n",
       "20        20       446660           21             19       0.0        0.0   \n",
       "\n",
       "       click_rate  \n",
       "ad_id              \n",
       "1        0.000000  \n",
       "2        0.000000  \n",
       "3        0.045455  \n",
       "4        0.068323  \n",
       "5        0.125000  \n",
       "6        0.000000  \n",
       "7        0.000000  \n",
       "8        0.289196  \n",
       "9        0.142857  \n",
       "10       0.000000  \n",
       "11       0.111111  \n",
       "12       0.000000  \n",
       "13       0.166667  \n",
       "14       0.125000  \n",
       "15       0.100000  \n",
       "16       0.000000  \n",
       "17       0.333333  \n",
       "18       0.100000  \n",
       "19       0.000000  \n",
       "20       0.000000  "
      ]
     },
     "execution_count": 300,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "promoted_content.set_index(promoted_content[\"ad_id\"], inplace = True)\n",
    "promoted_content[0:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "campaign_ctr = (promoted_content.groupby(\"campaign_id\")[\"num_click\"].sum() + 0.0)/ promoted_content.groupby(\"campaign_id\")[\"num_appr\"].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x326ff9f90>"
      ]
     },
     "execution_count": 310,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhcAAAF9CAYAAAC3XkrpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZJUlOFCkiQV9f8BUQ2F7/QDn8wAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x336e50410>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(campaign_ctr[~np.isnan(campaign_ctr)],bins = 50)\n",
    "plt.xlabel(\"Click Through Rate\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x31c6dd2d0>"
      ]
     },
     "execution_count": 312,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh8AAAF9CAYAAACkiQodAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npK9Dt0NzkwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x3270a6210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "document_ctr = (promoted_content.groupby(\"document_id\")[\"num_click\"].sum() + 0.0)/ promoted_content.groupby(\"document_id\")[\"num_appr\"].sum()\n",
    "plt.hist(document_ctr[~np.isnan(document_ctr)],bins = 50)\n",
    "plt.xlabel(\"Click Through Rate\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 314,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x337c0fe90>"
      ]
     },
     "execution_count": 314,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAg8AAAF9CAYAAACDJ4v1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x315b95650>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "advertiser_ctr = (promoted_content.groupby(\"advertiser_id\")[\"num_click\"].sum() + 0.0)/ promoted_content.groupby(\"advertiser_id\")[\"num_appr\"].sum()\n",
    "plt.hist(advertiser_ctr[~np.isnan(advertiser_ctr)],bins = 50)\n",
    "plt.xlabel(\"Click Through Rate\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Is every ad_id corresponds to only one document_id?\n",
      "The answer is:  True\n"
     ]
    }
   ],
   "source": [
    "print \"Is every ad_id corresponds to only one document_id?\"\n",
    "a = promoted_content.groupby(\"ad_id\")[\"document_id\"].count()\n",
    "print \"The answer is: \", sum(a) == len(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Is every document_id corresponds to only one ad_id?\n",
      "The answer is:  False\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x358ff2e50>"
      ]
     },
     "execution_count": 283,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAioAAAF/CAYAAABnm+JqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzt3XmcZFV58PFfd4/MooyM4zJEREDxEZCQAXnRAIKo0Tfm\njWTUgNGIgMsbQEVwF4JLjHEEREEc4oLvJELcUFQ0RkFlERBUcAEfFQRkcVgEhqWHYbr7/ePcYoqi\nh+muruq+Vf37fj79mbrn3qr7PFPV3U+fc+65A2NjY0iSJNXR4EwHIEmStCEWKpIkqbYsVCRJUm1Z\nqEiSpNqyUJEkSbVloSJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm25sx0AJ0QEUcDLwIeAbw/\nM785wyFJkqQO6PkelYh4LrBjZu4OvBjYZoZDkiRJHVKrHpWImAtcChyamec2tZ0MLAPuBY7LzOOb\nnvYCICPi65QelTdOb9SSJKlbatOjUhUkpwPbt+w6FtgZ2Bs4BDgmIpY17X8c8BfAvsAxwGe6Hqwk\nSZoWtehRiYjtgNPGaV8AHAy8MDMvBy6PiOXAYcAZ1WG3AT/PzFHgxxGx5TSFLUmSuqwuPSp7AWcD\nzwYGmtp3ohRTFza1nQ/s1rR9AfBCgIgI4OauRipJkqZNLXpUMnNF43GpNR6wOXBrZq5ralsFzIuI\nxZl5W2Z+IyL2ioiLqv2HdD9iSZI0HWpRqDyMBcB9LW2N7bmNhsx8a7snGBsbGxsYGNj4gZIkqVXX\nf4HWvVBZQ1NBUmls39uJEwwMDLB69TAjI6OdeLkZMzQ0yMKF8/siFzCfOuunXMB86qyfcoH+zafb\n6l6o3AA8NiIGq8myAEuA4cy8o1MnGRkZZd263v/QQH/lAuZTZ/2UC5hPnfVTLtB/+XRbXSbTbshl\nwP3As5ra9gQumZlwJEnSdKp1j0pmDkfESmBFRBwEbAEcCRwws5FJkqTpUMdCZaxl+wjKyrTnAHcC\nR2fmmdMelSRJmna1K1Qyc6hlexg4sPqSJEmzSN3nqEiSpFnMQkWSJNWWhYokSaotCxVJklRbFiqS\nJKm2LFQkSVJtWahIkqTaslCRJEm1ZaEiSZJqy0JFkiTVloWKJEmqLQsVSZJUWxYqkiSptixUJElS\nbVmoSJKk2rJQkSRJtWWhIkmSastCRZIk1ZaFiiRJqi0LFUmSVFsWKpIkqbYsVCRJUm1ZqEiSpNqy\nUJEkSbVloSJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2LFQkSVJtzZnpAGbabbfdxh133MO6\ndaPj7h8cHGSzzRZNc1SSJAksVHjJwR9gziYLNrh/9U1XcOZpn2LevHnTGJUkSQILFRZsvpS5Czbb\n8AEjaxkbG5u+gCRJ0gOcoyJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2+uKqn4j4FbCq2jw/\nM/95JuORJEmd0fOFSkQsBG7JzH1mOhZJktRZtSpUImIucClwaGae29R2MrAMuBc4LjOPb3raUmBx\nRHwPWAMcnpm/m97IJUlSN9RmjkpVkJwObN+y61hgZ2Bv4BDgmIhY1rR/NfBvmfl84EPAqd2PVpIk\nTYdaFCoRsR1wEbB1S/sC4GDgTZl5eWaeCSwHDms67ArgywCZeQHwZ9MStCRJ6rpaFCrAXsDZwLOB\ngab2nSjDUxc2tZ0P7Na0fQhwDEBE7ARc19VIJUnStKnFHJXMXNF4HBHNuzYHbs3MdU1tq4B5EbE4\nM28DPgn8Z0T8ELgfeH1HgxuAOXMGmTOnLjXd+IaGBh/0b68zn/rqp1zAfOqsn3KB/s2n22pRqDyM\nBcB9LW2N7bkAmbkGeFm3AhgcHGDRokcyf/78bp2ioxYu7I04J8p86qufcgHzqbN+ygX6L59uq3uh\nsoaqIGnS2L53OgIYHR3j9tvvYc2a0ek4XduGhgZZuHA+q1cPMzJS71gnwnzqq59yAfOps37KBfo3\nn26re6FyA/DYiBjMzMa7ugQYzsw7piWCMVi3bpR163rjQzUy0juxToT51Fc/5QLmU2f9lAv0Xz7d\nVveBssso806e1dS2J3DJzIQjSZKmU617VDJzOCJWAisi4iBgC+BI4ICZjUySJE2HOhYqYy3bR1BW\npj0HuBM4ulpPRZIk9bnaFSqZOdSyPQwcWH1JkqRZpO5zVCRJ0ixmoSJJkmrLQkWSJNWWhYokSaot\nCxVJklRbFiqSJKm2JnV5ckQMAMuAlwDPpixnPwrcCFwMfA34RmaOdDhOSZI0C024RyUi9gd+B/w7\n8Ajgc5RVYt8J/BewCPg0kBHxqo5HKkmSZp0J9ahExFeBRwNvAr6Tmes2cNwcYF/gzRHx95n5tx2L\nVJIkzToTHfo5NTO/vrGDqgLmy8CXI2LfKUUmSZJmvQkN/UykSBnnOV+bfDiSJEnrTXTo59UTfcHM\nXNl+OJIkSetNdOjncy3bY8AAMAzcDywERoBbAAsVSZLUERMd+hlsfAGvBC4HlmbmIzNzMyCAS4EP\ndS9USZI027Sz4NuHgX/KzMsbDZn5W+CNwLs7FZgkSVI7hcpmwJpx2oeA+VMLR5Ikab12CpXvAydF\nxFaNhojYDvgE8M0OxSVJkjS5JfQrhwDfAa6KiNspxc6jgR9Thn8kSZI6YtKFSmbeEBE7Ac8HnkG5\nAugy4PuZOdbh+CRJ0izWTo8K1U0Hv1N9SZIkdcVEF3wbATbPzJsjYpTSizKuzBzqVHCSJGl2m2iP\nykHAnU2PHeKRJEldN6FCJTP/X9Pjz23s+IhYBJyVmX/ZfmiSJGm2a+fy5InYBNitS68tSZJmiW4V\nKpIkSVNmoSJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2OlqoRMRA0+bABg+UJEmagEkXKhFx\ndUQ8Zpz2JwK3VJs3A0+aYmySJGmWm+i9fvYDXlhtbgV8IiKGWw7bimpp/eouyjd0JkRJkjRbTfRe\nPz8C3sD64ZwtgbVN+8eAu4EDOheaJEma7SZ6r58/APsARMT3gWWZeXs3A5MkSZpoj8oDMvO53QhE\nkiSp1aQLlYgI4BPA7pSbDz5IZg51IC5JkqTJFyrAKcDjgXcAd3Y2nPZFxNOASzLz0TMdiyRJ6ox2\nCpXdgN0z86edDqZdETEf+AjQeiWSJEnqYe0UKrfy4Ct+OiYi5gKXAodm5rlNbScDy4B7geMy8/iW\np34MeB/w5W7EJUmSZkY7K9OeCPxrRCzsZCBVQXI6sH3LrmOBnYG9gUOAYyJiWdPzXgdcVvXwuBqu\nJEl9pJ0elRcAewJ/iohVwH3NOzNzm8m+YERsB5w2TvsC4GDghZl5OXB5RCwHDgPOqA57FTASES8H\nlkTEWZn54snGIEmS6qedQuX86quT9gLOBo6iDO807ESJ8cKW87+7sZGZezUeR8TvLVIkSeof7ayj\n8r5OB5GZKxqPy9XPD9gcuDUz1zW1rQLmRcTizLyt5aXGOh2bJEmaOe30qBAROwFvBp4OvBx4CfCr\nzPxhB2MDWEDL0FLT9tzWg9sZdtqoAZgzZ5A5czp6o+mOGxoafNC/vc586qufcgHzqbN+ygX6N59u\na2fBt12AC4CLgF0oBcNS4ISI2Dczv9XB+Nbw0IKksX0v02BwcIBFix7J/Pnzp+N0U7ZwYW/EOVHm\nU1/9lAuYT531Uy7Qf/l0Wzs9Kh8Gjs3MoyLiLoDMfF31+L1AJwuVG4DHRsRgZo5WbUuA4cy8o4Pn\n2aDR0TFuv/0e1qwZ3fjBM2hoaJCFC+ezevUwIyP1jnUizKe++ikXMJ8666dcoH/z6bZ2CpVnUi4T\nbvUJ4PVTC+chLgPuB55FuYMzlCuOLunweTZsDNatG2Xdut74UI2M9E6sE2E+9dVPuYD51Fk/5QL9\nl0+3tVOorAXGW0PlScA9UwvnwTJzOCJWAisi4iBgC+BI4IBOnkeSJNVTOzNhvgZ8MCI2q7bHIuLp\nlNVhv9mBmFqv3DkC+AlwDmWxuaMz88wOnEeSJNVcOz0qbwW+TVlKfxD4KaWH5XLgbVMNqPXuy5k5\nDBxYfUmSpFmknXVUVgO7R8TzKFf7DAK/BP67acKrJEnSlLW1jgpAZp5NWU1WkiSpK9pZR2VXyt2M\nnwFs0rq/dehGkiSpXe30qHyashDbW4DhzoYjSZK0XjuFyrbArpn5q04HI0mS1Kydy5MvBbbqcByS\nJEkP0U6PyuuBr0fEbsDVwIOu9MnMlZ0ITJIkqZ1CZX/gqcBR4+wbAyxUJElSR7RTqLyRUqSckJnT\ncgdjSZI0O7UzR2UQON0iRZIkdVs7hcpKxr97siRJUke1M/SzGbB/RPwDZTLt/c07M3OfTgQmSZLU\nTqEyCpzW6UAkSZJatXNTQu9iLEmSpkU79/p5zsPtz8xz2w9HkiRpvXaGfn5AWS9loKltrPoaZZwb\nFUqSJLWjnUJl63Fe42nAB4B3TjkiSZKkSjtzVK4dp/mqiFgNrAB2nHJUkiRJtLeOyobcSllaX5Ik\nqSM6NZl2IXA48MspRyRJklTp1GRagGuAV00xHkmSpAd0YjItwNrMvGmqwUiSJDWb9ByVajLtVsC2\nmXlttX1kROzR6eAkSdLsNulCJSL2B74L/EVT85OBsyPiJZ0KTJIkqZ2rft4DHJGZxzYaMvPlwNuA\n93UqMEmSpHYKlacAZ43T/k3Kwm+SJEkd0U6hch2w1zjtzwacUCtJkjqmnat+PgmcGBFPAS6u2nal\nrKPygU4FJkmS1M4S+h+LiLnAmynzVQBuBN6TmSd1MjhJkjS7tdOjQmYuB5ZHxGLg/sxc3dmwJEmS\n2ixUIuJRlFVodwTuj4hfAV+wYJEkSZ3UzjoqW1Lu6XM88JfAc4GPAT+PiC06G54kSZrN2rnq5zjg\nD8DWmbk0M3eiLKt/LbC8k8FJkqTZrZ1C5QWUBd9WNRqqx28DXtipwCRJktopVNYB947TPgzMnVo4\nkiRJ67VTqFwAHB0Rj2g0VI/fU+2TJEnqiHau+nkHcCFwVURcWrXtCmzK+CvWSpIktWXSPSqZ+WvK\nnZNPpwz1zAM+D+yUmZd3NjxJkjSbtbvg27WUnpUZFxFzgJXAFsDdwKsy808zG5UkSeqECRUqEfHZ\nib5gZh7Ufjht2Q+4PjP/ISIOAN4JvH2aY5AkSV0w0R6VrZseDwJ7AquAnwL3U4aCngicOZVgqnsI\nXQocmpnnNrWdDCyjXG10XGYe33hOZn4+Ik6vNrcA7E2RJKlPTKhQycznNh5HxL8B1wMHZeZ9VdsQ\ncAow1m4gVUFyOrB9y65jgZ2BvYGtgJURcU1mntEU32hEfBN4JmWdF0mS1AfauTz5DcC/NIoUgMwc\nAT5CGYaZtIjYDriIB/fcEBELgIOBN2Xm5Zl5JmX128NaXyMz/4aypP+X2olBkiTVTzuFylpgy3Ha\nt6dMZm3HXsDZwLOBgab2nSi9Phc2tZ0P7NbYiIjXRcSh1ea9wEibMUiSpJpp56qf04DPRMRRlPkk\ng8DuwPsoc0kmLTNXNB5HRPOuzYFbM3NdU9sqYF5ELM7M24AvUoaDXk4pcl7fTgwbNABz5gwyZ047\nNd30GRoafNC/vc586qufcgHzqbN+ygX6N59ua3fBtwXACuARlOJgDXAipVjppAXAfS1tje25AJl5\nJ/CSDp/3AYODAyxa9Ejmz5/frVN01MKFvRHnRJlPffVTLmA+ddZPuUD/5dNtky5UMnMt8IaIOBII\nygTazMx7Oh0cpQBqvX9QY3u8+w113OjoGLfffg9r1oxOx+naNjQ0yMKF81m9epiRkXrHOhHmU1/9\nlAuYT531Uy7Qv/l0W1sLvgFk5t3ATzoYy3huAB4bEYOZ2XhXlwDDmXlHl89djMG6daOsW9cbH6qR\nkd6JdSLMp776KRcwnzrrp1yg//LptroPlF1GWaflWU1tewKXzEw4kiRpOrXdozIdMnM4IlYCKyLi\nIMqCbkcCB8xsZJIkaTpMqEclIl4SEfO6HUylddG4IyhDTOdQJuweXa2nIkmS+txEe1ROo0ycvT4i\nrgZ2rS4N7rjMHGrZHgYOrL4kSdIsMtFC5U7gvRFxHmUZ+1dExOrxDszMlR2KTZIkzXITLVTeQ7nn\nzkGUoZmPb+C4McBCRZIkdcREb0p4KnAqQESMAptn5qpuBiZJktTO5clbAzcDRMTjImKzzoYkSZJU\nTLpQycxrgTdGxE3AH4HbIuKGiDi849FJkqRZbdKFSkS8AVgOnA4sA15GuTHgh6q1TiRJkjqinQXf\n3gK8NTNPamr7akT8Djgc+GxHIpMkSbNeO3NUngx8e5z2/wa2nVo4kiRJ67VTqFwLPHOc9l0BrwSS\nJEkd087QzynAJyLiMcAFVdsewPuBj3UqMEmSpHYKlY9Rhn9OqJ4/QLnD8SnABzsXmiRJmu0mXahk\n5ihweEQcDTydUqhcmZl3dTo4SZI0u7XTowJAVZhc0sFYJEmSHqSdybSSJEnTwkJFkiTVloWKJEmq\nrXaW0D9nvBsRRsTjI+JnnQlLkiRpgpNpI+KvWb/I217AuyPi7pbDtgW26lxokiRptpvoVT/XACdR\nLkUG2B8Yado/BtwNvK1jkUmSpFlvQoVKZl4BbAMQEb8Hds3MW7sZmCRJUjsLvm3djUAkSZJaTbpQ\niYglwL8AuwObsH44CIDM3KYzoUmSpNmunZVpPwXsAvwXcGdnw5EkSVqvnUJlH+BFmXlep4ORJElq\n1s6Cb3cDqzodiCRJUqt2CpWVwNsjYqjTwUiSJDVrZ+jnscArgL+JiKuA+5p3ZuY+nQhMkiSpnUIF\n4PSORiFJkjSOdtZRObAbgUiSJLVqZx2V5zzc/sw8t/1wJEmS1mtn6OcHlHv7NC/0NlZ9jVIWgZMk\nSZqydgqV1iX05wBPAz4AvHPKEUmSJFXamaNy7TjNV0XEamAFsOOUo5IkSaK9dVQ25FbgqR18PUmS\nNMt1ajLtQuBw4JdTjkiSJKnSqcm0ANcAr5piPJIkSQ/oxGRagLWZedNUg5EkSWrW9mTaiNgUeDpw\nP3BVh+OSJElqa47KIHAscAjwCMoQ0H0RcQrwlswc62yIG41nLuVGiU+grOFyeGb+eDpjkCRJ3dHO\nVT/vAg4C3g7sDOxCWT/l1cBbOxfahB0EXJmZewOvAU6YgRgkSVIXtDNH5bXAIZl5WlPbzyLiFuB9\nwEfaDabqHbkUOLSxFH/VdjKwDLgXOC4zj2962krK5F4oPTwPupuzJEnqXe0UKk8ALh6n/WLgSe0G\nUhUkpwPbt+w6ltJzszewFbAyIq7JzDMAMvOe6vmPoxQtR7Ybw3hGR0e4/PKfMXfu3Ic9bocddmST\nTbx7gCRJndROofIb4Pk8dALtCyiXKE9aRGwHnDZO+wLgYOCFmXk5cHlELAcOA85oOu5pwJeAd2Xm\nD9qJYUPuvuMWjv7EWWy6eMsNHnPXbdex/AhYunSXTp5akqRZr51C5XjglIjYBrigatuDUjy0O0dl\nL+Bs4CjK8E7DTlWMFza1nQ+8u7EREU8CzgT+MTMvbfP8D2vTxVuy2ZJtu/HSkiTpYbRzefLKiHgM\n8A7gbVXzKuCozDy5nSAyc0XjcUQ079ocuDUz1zW1rQLmRcTizLyNUtw8ElgeEQPAzZm5XztxjKt1\nWbsNGBoaZM6cTt6RYHKGhgYf9G+vM5/66qdcwHzqrJ9ygf7Np9va6VEhM08ATqjmhQxk5s2dDesB\nC3jo5NjG9twqljd06dwADE7wfVi4cD6LFj2ym6FMOI5+Yj711U+5gPnUWT/lAv2XT7dNqlCJiP2A\nb2fmaoDMvCUi3hARdwJf6MIaKmuoCpImje17mQajoxM7bvXqYW6//Z7uBvMwhoYGWbhwPqtXDzMy\nMsGga8x86qufcgHzqbN+ygX6N59um1ChEhGPoExe/WvgeZT7/TTsAbwSeGlEvKJlmGaqbgAeGxGD\nmdl4V5cAw5l5RwfPs2ETLL1GRkZZt27mP3h1iaNTzKe++ikXMJ8666dcoP/y6baJDjC9GdgV2Lv1\nqprM/Edgn+rrnzoaHVxGWaL/WU1tewKXdPg8kiSphiZaqBwAHJGZ5423sypejqJcStwxmTlMWRtl\nRUQ8MyL2payT4uqzkiTNAhMtVLYGfrSRY74HPGVq4QAPHWw5AvgJcA5wInB0Zp7ZgfNIkqSam+hk\n2ruAzTZyzHw6MME1M4datoeBA6svSZI0i0y0R+VCYP+NHPNK4OdTC0eSJGm9ifaofBQ4OyKuB05u\nugKHapG1N1KGaJZ1PkRJkjRbTahQyczzIuJIyvL5R0fEj4HbgcXAbsCmwHsy8xtdi1SSJM06E17/\nNjNPBJ4JfJFSoPwvYCHwGWCnzFzelQglSdKsNamVaTPz55RhHkmSpK6bUI9KRHw0IhZO9EUj4jER\n8fH2w5IkSZr40M91wC8j4sMRscuGDoqIpVWBcgVwbScClCRJs9dEJ9N+NCK+DhwN/Cgi/gT8ErgZ\nGAIeB+xEmbPyBWDPzPxtd0KWJEmzxYTnqGTmVcBrIuLdwIspV/ssAUaB3wGfB76Rmbd0I1BJkjT7\nTGoyLUBm3gh8qvqSJEnqmkkXKhHx6g3sGgPWAtcDF2XmyFQCkyRJmnShQpmnsjVlIu6dVdujKYXK\nQLWdEfGCzLx+6iFKkqTZasILvjU5mXJVz06ZuSgzFwE7AD8DDgWeCFwNuACcJEmaknYKlSOAf8rM\nXzQaMvNK4DDg3Zl5E3AU8ILOhChJkmardgqVzVg/5NPsXuAx1ePbgfntBiVJkgTtFSrnAcsj4tGN\nhojYDPg34EdV00uBnHp4kiRpNmtnMu1hwDnA9RGRlGJnW+BW4EUR8QJK0bJfx6KUJEmz0qR7VDLz\namA74E2UHpQfUIqXyMwEfgPsmJlndDBOSZI0C7XTo0JmDkfElynL6N8PXJWZa6t93uNHkiR1RDsL\nvg0CxwKHAI+omtdGxCnAWzJzrIPx9YTRkXVk/nqjx+2ww45ssskm0xCRJEn9oZ0elXcBBwFvB35I\nGT56DnAMcAPwkY5F1yPuueMmPnPWjWx60d0bPOau265j+RGwdOkGbz4tSZJatFOovBY4JDNPa2r7\nWUTcAryPWVioAGy6eEs2W7LtTIchSVJfaefy5CcAF4/TfjHwpKmFI0mStF47hcpvgOeP0/4C4Jop\nRSNJktSknaGf44FTImIb4IKqbQ/KJcpv7VRgkiRJky5UMnNlRDwGeAfwtqp5FXBUZp7cyeAkSdLs\n1u46KicAJ0TE44CBzLy5s2FJkiS1Wag0ZOYtjccR8Rzgc5m5zZSjkiRJor3JtBsyH3hyB19PkiTN\ncp0sVCRJkjrKQkWSJNWWhYokSaqtCU2mjYh/nsBhrh8vSZI6aqJX/Rw4weOuazeQfucdliVJmrwJ\nFSqZuXW3A+l33mFZkqTJm9I6Kpoc77AsSdLkOJlWkiTVVl8VKhGxb0R8dqbjkCRJndE3hUpELAf+\ndabjkCRJnVO7OSoRMRe4FDg0M89tajsZWAbcCxyXmce3PPVi4CzggGkMV5IkdVGtelSqguR0YPuW\nXccCOwN7A4cAx0TEsuYDMvMr0xGjJEmaPrUpVCJiO+AiYOuW9gXAwcCbMvPyzDwTWA4cNv1RSpKk\n6VSbQgXYCzgbeDYw0NS+E2WI6sKmtvOB3aYvNEmSNBNqM0clM1c0HkdE867NgVszc11T2ypgXkQs\nzszbpilESZI0zWpTqDyMBcB9LW2N7bnNjZn5Q+CHHT37wMYP6aShoUHmzJl8R9fQ0OCD/u115lNf\n/ZQLmE+d9VMu0L/5dFsvFCpraClImrbv7fbJB6f587Rw4XwWLXrklJ7fT8ynvvopFzCfOuunXKD/\n8um2XihUbgAeGxGDmTlatS0BhjPzjm6ffHR048d00urVw9x++z2Tft7Q0CALF85n9ephRkamOegu\nMJ/66qdcwHzqrJ9ygf7Np9t6oVC5DLgfeBbwo6ptT+CSaTn72LSc5QEjI6OsW9f+B3iqz68b86mv\nfsoFzKfO+ikX6L98uq32hUpmDkfESmBFRBwEbAEciQu7SZLU9+paqLT2YxxBWZn2HOBO4OhqPRVJ\nktTHalmoZOZQy/YwcGD1JUmSZon+uEZKkiT1JQsVSZJUWxYqkiSptixUJElSbdVyMu1sNTqyjsxf\nb/S4HXbYkU022WQaIpIkaWZZqNTIPXfcxGfOupFNL7p7g8fcddt1LD8Cli7dZRojkyRpZlio1Mym\ni7dksyXbznQYkiTVgnNUJElSbVmoSJKk2rJQkSRJtWWhIkmSastCRZIk1ZaFiiRJqi0LFUmSVFsW\nKpIkqbYsVCRJUm1ZqEiSpNqyUJEkSbVloSJJkmrLQkWSJNWWd0/uMaMj68j89UPah4YGWbhwPqtX\nDzMyMsoOO+zIJptsMgMRSpLUORYqPeaeO27iM2fdyKYX3b3BY+667TqWHwFLl+4yjZFJktR5Fio9\naNPFW7LZkm1nOgxJkrrOOSqSJKm2LFQkSVJtWahIkqTaslCRJEm1ZaEiSZJqy0JFkiTVlpcn96EN\nLQrXykXhZt7atWv51a9+sdHjeu296te8pLrrx+89C5U+5KJwveNXv/oFbz/+DDZdvOUGj+nF96pf\n85Lqrh+/9yxU+pSLwvWOfn2v+jUvqe767XvPOSqSJKm2LFQkSVJtWahIkqTaslCRJEm1ZaEiSZJq\ny0JFkiTVVs9fnhwRA8CngQBWA6/OzFtnNipJktQJ/dCj8nfAvZm5B/A54N0zG44kSeqUWvWoRMRc\n4FLg0Mw8t6ntZGAZcC9wXGYe3/S03YH/qR7/N/DO6YtYkiR1U216VKqC5HRg+5ZdxwI7A3sDhwDH\nRMSypv0LKUM+AHcBj+pupJIkabrUolCJiO2Ai4CtW9oXAAcDb8rMyzPzTGA5cFjTYauBTavHmwJ3\ndj9iSZI0HWpRqAB7AWcDzwYGmtp3ogxPXdjUdj6wW9P2hcBfVY//GvhR98KUJEnTqRZzVDJzReNx\nRDTv2hy4NTPXNbWtAuZFxOLMvA04A/jfEXE+sBbYv6PBDWz8kF41NDTInDnt16pr167ll7/c+O3E\nn/GMyd/anZh6AAAO+ElEQVROfGho8EH/dupc3Yz54bTm09o+kedP5b3qpA3lMt4xD2d0ZB2//W1u\n9NhOvxetJpJPHW3oszw4OMCjHjWPu+9ew+joWNf//7ppQ+/NTH0fT9V0fNam82fKdH3PDIyNjU3L\niSYqIkaBvTPz3Ih4FfCBzNy6af/WwO+AJ2XmjTMVpyRJ6r66/wmxBpjb0tbYvneaY5EkSdOs7oXK\nDcBjI6I5ziXAcGbeMUMxSZKkaVL3QuUy4H7gWU1tewKXzEw4kiRpOtViMu2GZOZwRKwEVkTEQcAW\nwJHAATMbmSRJmg51LFRaZ/ceQVmZ9hzKGilHV+upSJKkPle7q34kSZIa6j5HRZIkzWIWKpIkqbYs\nVCRJUm1ZqEiSpNqq41U/0yIi5lKuJlpGWeX2uMw8fmajeqgqzkuBQzPz3KptK+BTlJs4XgO8JTO/\n2/Sc5wMfBbah3LTxdZn5+6b9hwNvpdxt+kvAYZm5pos5/BnwceC5lP/rLwLvysy1vZZLdc6nAJ8A\ndgduA07KzGOrfT2XT9O5zwJWZeZBvZpLROxLuf/XGOVOXWPAVzLz73s0n02qmF4B3Ad8NjPfU+3r\nqXwi4gDgVB783gwAo5k5p7o9yr/3UD5bAJ8EnkP5OfCxzPxYtW8reui9qc75uCqf5wG3AB/MzP9X\nh3xmc4/KscDOwN7AIcAxEbFsRiNqURUppwPbt+z6GnAjsAvwn8BXq28aIuJJwFeBzwDPBG6tjm+8\n5kuBfwZeB+xDWUxveVcTga8A8yi/2PcH/g/wgWrfmb2US0QMAGdRbo75F8D/BY6KiMbNMHsqn6Zz\n7w/875bmXvucQfle+TplBesllBubvrba14vvzccpvzheAPwD8LqIeF2P5vNfrH9PlgBPpty37YRq\nf6993r4E3EX5PXI48MGIeEm1r9feG6oY/gzYq8rn+Krwn/F8ZuXlyRGxgPKf+cLMPK9qew/wvMzc\nZ0aDq0TEdsBp1eafA8+tbtS4D+VD8PhGRRoR3wXOy8z3R8T7gT0aeUTEfOCPwP+pnv9D4HuZ+YFq\n/+7A/wCLu1GxR7kd9hXAEzLz1qptf+AjwKsp3wA9kUt1jiWUvxxem5n3VG1fAW6iFGQ9lU91nkXA\n5ZQfRFdk5kG99jlryuU/gGsz86iW9p7Lp3pfVgH7ZOb5VdvbgacBn6cHP2st+b0LOBDYgbLieM+8\nPxGxGfAn4BmZeUXV9mXK99BX6bH3JiJ2AX4MbJOZ11Ztbwf2Bd4z0/nM1h6VnSjDXhc2tZ0P7DYz\n4YxrL+BsSlfbQFP7bsBPW97g86vjGvvPbezIzGHgp8Czo9wzaVfgvKbnXgRsQvk/6YY/Ai9qFClN\nHk2prHspFzLzj5n5iqYiZXfKD9kf0IP5VI4FVgJXNrX12uesYXvgN+O092I+ewB3NIqUKq7lmfla\nevezBjxQhL0deEdm3k/vvT/DwD3AgRExp/qDbHfgZ/Tme7MNcEujSKn8nNJDsicznM9sLVQ2B27N\nzHVNbauAeRGxeIZiepDMXJGZbx2n4tycUrU3W0W5vcDG9m9GGYJ5YH9mjlDGV7egCzLzzpaxzAHg\nMEoR1lO5tIqIayjfoBdS5kX0XD5VT8OerB+Ka+i5XCoBvCgiMiJ+FxEfiohHbCReNrJ/pvLZBrgm\nIv4xIq6MiKsi4qjqe6gX82l2CHBDZn612u6pfDLzPsrPsf9LKVquBL6VmaduJFY2sn+m3ptVwGYR\nMa+pbUvKH/RPeJh4YRryma2TaRdQJqY1a2zPneZYJmtDsc+dwP4FTdsben63fQRYSqmyj9hILHXP\nZRllrP2TlOGgnnpvqjlQK4BDMvO+8kfhA3oqF4CI2BKYT/nF8XJga8ocj/kbiZeN7J+pz9qjKMM8\nrwdeQ/mFcAplQnov5tPsYODfmrZ7MZ/tKPOhjgV2BE6MiLPpzVwupgxfnxQRb6LMVXkLZcLzvI3E\n0/V8ZmuhsoaH/ic1tu+d5lgmaw3wmJa2uayPe0O53V7tYwP7u553RHwYeBPw95l5RUT0bC4AmflT\ngIg4gjJn4DPAooeJp275vBe4JDO/N86+nntvMvO6iFicmXdUTT+PiCHK5L9T6a33BmAd5SqJV2Tm\n9QAR8WRKb8T/AK29v3XPB4CI2BV4IvCFpuae+rxFxPMoxdYWVe/Kz6rJpUdReot76r2p/lB5GeWK\nzNWUHpHllD/ARinF/obi6Xo+s3Xo5wbgsdX4WcMSYLjph1xd3UCJtdkSSjW8sf23UT44D+yvfpAv\nbnp+V0TEiZQK/ZWZ2ZgR3nO5RMTjm2b2N1xBGXO96WHihfrlsx+wb0TcFRF3Aa8EXhURq4HrHyZW\nqF8uAIzz/Xsl5S/CP9J7+dwErGkUKZWkdJn33PdOkxcC52bmnU1tvZbPzsBvqyKl4WeU4ZJeywWA\nzPxJZj6F0pvyJMpcr1uAqx4mXpiGfGZroXIZcD9l0lPDnsAlMxPOpFwE7Fx12zfsUbU39u/R2FFd\n4bQUuDAzxyg57tH03L8E1lKu+uiKiDiG0n29X2Z+qWlXz+VCGU44IyI2b2p7JnAzZYLZLj2Uz16U\nLuudqq+vU2b370TpCu6p9yYi/ioibm0ZZ19KucLvPHrrvWnENC8intrUtj1lHYuL6L18GnYDLmhp\n67WfBTcCT42I5lGJ7YDf04PvTUQsiojzImJRZt6cmaPA31AuErh4pvOZlZcnA0TEJymztA+i/IXy\nOeCAzDxzJuMaT0SMAntXl3oNUt7gX1ImQP4t8C5gh8y8vuoavgJ4H/BN4BjgaZm5tHqt/SjzEl5D\n+Wb7LOXSsbd0KfbtKLPH/5WywF6zW3opl+qcg5TJs3+izLHZmjLk88Eqv58Dv+iVfFpyOxUYqy5P\n7qnPWXXOR1UxnQu8H3gKZZGqj1ZfPffeRMTXKUMih1DmqKyscvtkL+ZTnfv3lKt9vtjU1lOft4hY\nSOmt+y7le//p1TnfVf3bc+9NRPwU+AnlZ/XzgI9R/oC/jBl+b2ZrjwqUXzI/Ac4BTgSOrmORUnmg\nmqwq3ZdQutIupSwCtW+je7i6vGwZpQD7MWXW9b5Nz/8C8CHKpLzvUH7pvqOLsf8t5XN2FOVDeiOl\ny+/GKpd9eyiX5v//e4AfUVbSPCEzT6r2/W0v5bMhPfg5IzPvpgwrPI7yV9yngBWZeVwPvzevpCyK\ndh7lj6mPZ+YnejgfgMdT5i88oNc+b5m5mvLLfPMqnuOA92fmp3v4vdkPeCqlyHoT8LLM/Gkd3ptZ\n26MiSZLqbzb3qEiSpJqzUJEkSbVloSJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2LFQkSVJt\nWahIkqTaslCR+lREXBMRV0fEI8fZd2pEnNPl8/8+Iv65m+eYqIjYJSKujIjhiFjewdcdjYhXd+r1\nJD2UhYrUv8aAJwMfmelAauDdlNvNb0e574ikHmGhIvW3q4E3RMQ+Mx3IDFsEXJaZ12Tm7Rs9WlJt\nzJnpACR11X8CuwOfjYgdMvOe8Q6KiFHgNZm5cry2iDgG2AM4FzgUWACcBvwL8ElgH8qdsd+cmd9q\neuk/i4hvAc+l3DX72Mw8uekcf0np4dgVuAX4BvCuzLyr2v974MvAX1PuivzSzDxvnPhfTLlD9zOA\nu4DTgXdn5n3Va2wJDFTDNFtn5nUtzx8A3gkcAGwF3AdcAByamb+vjnkicHKVyx203AE2IuZT7sT+\nYsodZK8EPpCZXx3v/1zSxNijIvW3MeBgSo/CcVN8recAQSlY3gi8nnJb99OBnSm/mE9tec5rgR8C\nOwLHAx+LiJcARMSfA98FvkUpMF5Rvc7/tLzGocBhwIuAi1qDioi/A84Evg4sreLar4oL4JnV875A\nuVX9H8bJ7c3AkcBbgG0pt7V/GtX/WUQMUW5R/xhgT+DlwNso/78N/1Ll8SLg6cC3gf+KiC3HOZ+k\nCbJHRepzmfmHiHgrcEpEfDkzv9fmSw0Ar8/Me4HfRcRHgO9l5mkAEXEy8OKIeEJmrqqe89XM/HD1\n+KSIeBalIDiT8ov+O037r46IVwJXRcRzMvPcqv1bmfn9h4nrHcBXMrMx9+R3EXEI8LWIeHpm/joi\n1gLDmXnLBl7jt8CrM/Pb1fYfIuJLwMuq7edT5rc8JTOvqfI9EPhZ02tsQ+nNuSYz74yIo4EfAA41\nSVNgoSLNApn5qYh4GfDpiNixzZdZVRUpDfdQ5sA0DFf/zm1qu6DlNS6mDONA6f14akTc1XLMGKUo\naBQqv91IXDtShqGa/bBp36838nwy86yI+F8R8T5Kr1EAOwDXV4c8A7i9UaRUz7k8IoabXubDlF6d\nWyLiYkrP0GmNYSxJ7XHoR5o9Xgs8mjIE87CqoY5W94/TNrqRlxpp2R6izP+A8vPn88CfAzs1fW3L\ngwuP5mJgPAPjtDV+to0X80NExDuB7wOLge8BbwCObTpkjPF/Xj7w+pl5EfAkYBnwE+DVwJUR8dyJ\nxCBpfBYq0iyRmX8A3kqZs7Jny+77gYVN20/r0Gl3adneE/hF9fiXwPaZ+fvMvDozrwY2AU6g/MKf\nqJ9T5s00ew6luLhigq/xLuC9mXlYZn46M39M6VVpFEGXAY+OiO0aT4iIbWn6P4uI9wJ7ZuY3M/Pw\n6vlXAS+dRC6SWjj0I80imfmZiHg58FdA85UvFwKvi4jzKH/AHE9Zd2SyWns3XhERPwe+CfwdZZJq\no4fhOODciDgJOIky4fcTwDzgN5M453LgixHxHuCLlALhROAbmTnR1/kD8FcR8U1KL9Crq3j/WO3/\nPmXi8H9ExKHVMSfy4B6jbYBXRsTrKQXKsyhXG7UOf0maBHtUpP41toH211Iur23e/0/AnygFy5eA\nf2f9/IzJvP5Yy+OPUC7XvRx4DfCKxuXFmXkx8ELKcM9PgK9Rrhx6fmau20gOD8jMMyhXDL2c0rty\nMmVIab+NPbfJqyiXXF9Cmd+yA2X45/ERsUVmjlHm1vyacvXPNyjDU82Tcw8Bzgb+A0jgfcDbM/N0\nJLVtYGxsoz8HJEmSZoQ9KpIkqbYsVCRJUm1ZqEiSpNqyUJEkSbVloSJJkmrLQkWSJNWWhYokSaot\nCxVJklRbFiqSJKm2LFQkSVJtWahIkqTa+v/ehf3sgzYXMAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x35ece7dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print \"Is every document_id corresponds to only one ad_id?\"\n",
    "a = promoted_content.groupby(\"document_id\")[\"ad_id\"].count()\n",
    "print \"The answer is: \", sum(a) == len(a)\n",
    "plt.hist(a,bins = 50, log = True)\n",
    "plt.xlabel(\"Number of ads\")\n",
    "plt.ylabel(\"Log(Count of document_id)\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Is every ad_id corresponds to only one campaign_id?\n",
      "The answer is:  True\n"
     ]
    }
   ],
   "source": [
    "print \"Is every ad_id corresponds to only one campaign_id?\"\n",
    "a = promoted_content.groupby(\"ad_id\")[\"campaign_id\"].count()\n",
    "print \"The answer is: \", sum(a) == len(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Is every campaign_id corresponds to only one ad_id?\n",
      "The answer is:  False\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x349a12e90>"
      ]
     },
     "execution_count": 288,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAioAAAF/CAYAAABnm+JqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzt3XmcXHWZ6P9PpyEBNBEIDjAKAg4+AnIzARlXFvnp6B3v\nb1RGB1wuSFD5CbgArgiDXGdGDYt4RSaO6y8zwow6CCLOeAVUdgQVEIRHBJHVsAgESEOS7r5/nCpS\nFJ109elT1SdVn/fr1S/qfM+pU8+T06k8nPNdhsbHx5EkSaqjWTMdgCRJ0tpYqEiSpNqyUJEkSbVl\noSJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2LFQkSVJtWahIkqTa2mCmA6hCRNwALGtsXpKZ\nfzeT8UiSpGqs94VKRMwD7svMfWc6FkmSVK1aFSoRMQe4Gjg8My9qaTsd2A9YAZycmae0vG0hMD8i\nzgceBz6Ymb/tbeSSJKkbatNHpVGQnAns3LbrJGA3YB/gMOD4iNivZf9y4DOZ+Wrg08DXux+tJEnq\nhVoUKhGxE3AFsH1b+ybAIcD7M/PazDwHWAwc0XLYr4HvAGTmpcCf9iRoSZLUdbUoVIC9gQuAlwFD\nLe0LKB5PXd7Sdgnwkpbtw4DjASJiAXB7VyOVJEk9U4s+Kpm5pPk6Ilp3bQ3cn5mrW9qWARtFxPzM\nfAD4J+BfI+KnwCrgPT0IWZIk9UAtCpV12AR4oq2tuT0HIDMfB95c9gPGx8fHh4aGJj9QkiS16/o/\noHUvVB6nUZC0aG6vqOIDhoaGWL58hNHRsSpOV0vDw7OYN29j8+wT5tlfBiVPGJxcBy3Pbqt7oXIX\nsEVEzMrM5tXeChjJzIeq+pDR0TFWr+7fX6Ym8+wv5tlfBiVPGJxcByXPbqtLZ9q1uYai38lLW9r2\nBK6amXAkSVIv1fqOSmaORMRSYElELAKeCxwNHDSzkUmSpF6oY6Ey3rZ9FMXMtBcCDwPHNeZTkSRJ\nfa52hUpmDrdtjwAHN34kSdIAqXsfFUmSNMAsVCRJUm1ZqEiSpNqyUJEkSbVloSJJkmrLQkWSJNWW\nhYokSaotCxVJklRbFiqSJKm2LFQkSVJtWahIkqTaslCRJEm1ZaEiSZJqy0JFkiTVloWKJEmqLQsV\nSZJUWxYqkiSptixUJElSbVmoSJKk2rJQkSRJtWWhIkmSastCRZIk1dYGMx3ATDvm+M+wcuUoY+Pj\nE+6fM3sDjnrf4T2OSpIkgYUKl98+j2du/py17n/kNxf2MBpJktTKRz+SJKm2LFQkSVJtWahIkqTa\nslCRJEm1ZaEiSZJqy0JFkiTVloWKJEmqLQsVSZJUWxYqkiSptixUJElSbVmoSJKk2rJQkSRJtWWh\nIkmSastCRZIk1ZaFiiRJqi0LFUmSVFsWKpIkqbYsVCRJUm1ZqEiSpNqyUJEkSbVloSJJkmrLQkWS\nJNWWhYokSaotCxVJklRbfVOoRMQLIuLhmY5DkiRVpy8KlYjYGDgRGJnpWCRJUnU2mOkAWkXEHOBq\n4PDMvKil7XRgP2AFcHJmntL21s8DJwDf6WG4kiSpy2pzR6VRkJwJ7Ny26yRgN2Af4DDg+IjYr+V9\n7wauycxfAEO9iVaSJPVCLQqViNgJuALYvq19E+AQ4P2ZeW1mngMsBo5oOewdwJsj4sfAVhFxXo/C\nliRJXVaXRz97AxcAx1I83mlaQBHj5S1tlwDHNDcyc+/m64j4XWa+vruhSpKkXqlFoZKZS5qvI6J1\n19bA/Zm5uqVtGbBRRMzPzAfaTjVedWxDDLHBBrW48VTa8PCsp/y3X5lnfzHP/jMouQ5ant1Wi0Jl\nHTYBnmhra27PaT84M3eoOoBZs4bYbLNnVH3aGTFv3sYzHUJPmGd/Mc/+Myi5Dkqe3Vb3QuVxnl6Q\nNLdX0ANjY+M8+OBjvfiorhkensW8eRuzfPkIo6NjMx1O15hnfzHP/jMouQ5ant1W90LlLmCLiJiV\nmc2rvRUwkpkP9SKAccZZvbo/ftFGR8f6Jpd1Mc/+Yp79Z1ByHZQ8u21KhUpEDFHMZ/IG4GUURcMY\ncDdwJXA2cG5mjlYU3zXAKuClwGWNtj2Bqyo6vyRJqrGOe8JExAHAb4F/BjYEvgEcDXwM+DdgM+Ar\nQEbEO6oILjNHgKXAkoh4cUS8sfGZp1ZxfkmSVG8d3VGJiO8CzwLeD/ywbRRO63EbAG8EPhARf5uZ\nf10ipvaRO0dRzEx7IfAwcFxjPhVJktTnOn308/XM/N5kBzUKmO8A32nc/ZiyzBxu2x4BDm78SJKk\nAdLRo59OipQJ3nP21MORJElao9NHPwd2esLMXFo+HEmSpDU6ffTzjbbtcYoFAEcoRuXMA0aB+yg6\nv0qSJE1bp49+ZjV/gLcD1wILM/MZmbkpEMDVwKe7F6okSRo0ZSbq/yzw3sy8ttmQmTcD76NlsUBJ\nkqTpKlOobEoxtX27YcCFDSRJUmXKFCo/Bk6LiO2aDRGxE/BF4PsVxSVJklRqrZ/DgB8Ct0TEgxTF\nzrOAn1E8/pEkSarElAuVzLwrIhYArwZeRDEC6Brgx5nZPqusJElSaaVWT24sOvjDxo8kSVJXdDrh\n2yiwdWbeGxFjPH09nie1T4EvSZJUVqd3VBZRLAjYfO0jHkmS1HUdFSqZ+f+3vP7GZMdHxGbAeZn5\n8vKhSZKkQVdmeHInZgMv6dK5JUnSgOhWoSJJkjRtFiqSJKm2LFQkSVJtWahIkqTaslCRJEm1ZaEi\nSZJqq9JCJSKGWjaH1nqgJElSB6ZcqETErRGx+QTtzwHua2zeC2wzzdgkSdKA63Stn/2B1zY2twO+\nGBEjbYdtR2Nq/cYqyndVE6IkSRpUna71cxlwKGse52wLrGzZPw48ChxUXWiSJGnQdbrWzx3AvgAR\n8WNgv8x8sJuBSZIkdXpH5UmZ+apuBCJJktRuyoVKRATwReAVFIsPPkVmDlcQlyRJ0tQLFeBLwJ8A\nHwUerjYcSZKkNcoUKi8BXpGZv6g6GEmSpFZlJny7n6eO+JEkSeqKMoXKF4B/jIh5VQcjSZLUqsyj\nn9cAewJ/jIhlwBOtOzNzhyoCkyRJKlOoXNL4kSRJ6qoy86ic0I1AJEmS2pW5o0JELAA+ALwQeAvw\nBuCGzPxphbFJkqQBV2b15N2BK4EdgN2BOcBC4EcR8VfVhidJkgZZmVE/nwVOysx9aAxTzsx3A6cB\nn6wsMkmSNPDKFCovBpZO0P5FYOfphSNJkrRGmUJlJTDRHCrbAI9NLxxJkqQ1yhQqZwP/EBGbNrbH\nI+KFwOeB71cWmSRJGnhlCpUPAc+kmEr/GcAvgBuAUeDD1YUmSZIGXZl5VJYDr4iI/4ditM8s4Hrg\nvzJzrOL4JEnSACs1jwpAZl4AXFBhLJIkSU8x5UIlIvYATgdeBMxu35+ZwxXEJUmSVOqOyleAx4Ej\ngZFqw5EkSVqjTKGyI7BHZt5QdTCSJEmtyoz6uRrYruI4JEmSnqbMHZX3AN+LiJcAtwJPGemTmRPN\nWitJkjRlZQqVA4A/A46dYN84E0+v3zURsUHjM58LPAq8IzP/2MsYJElSd5QpVN5HUaScmpkrKo6n\njP2BOzPzbRFxEPAx4CMzHJMkSapAmUJlFnBmN4qUiJhD0Qfm8My8qKXtdGA/YAVwcmae0nxPZn4z\nIs5sbD4X8G6KJEl9okxn2qXAYVUH0ihIzuTpKzCfBOwG7NP43OMjYr/WAzJzLCK+T3G357yqY5Mk\nSTOjzB2VTYEDIuJtFJ1pV7XuzMx9p3rCiNgJOGOC9k2AQ4DXZua1wLURsRg4Ajir7XP/R0TsAPwA\neOFUY5AkSfVTplAZY4KiYpr2ppiO/1iKxztNCyhivLyl7RLgmOZGRLwbmJ2ZX2y8d7Ti2CRJ0gwp\nsyjhwVUHkZlLmq8jonXX1sD9mbm6pW0ZsFFEzM/MB4BvAUsj4i3AEMXw6coMMcQGG5R5QlYfw8Oz\nnvLffmWe/cU8+8+g5DpoeXZbmbV+9lrX/mYn2IpsAjzR1tbcntP4vIeBN1T4mU8xa9YQm232jG6d\nvqfmzdt4pkPoCfPsL+bZfwYl10HJs9vKPPr5CcV8KUMtbeONnzEmWKhwGh6nUZC0aG73ZGj02Ng4\nDz74WC8+qmuGh2cxb97GLF8+wujo2ORvWE+ZZ38xz/4zKLkOWp7dVqZQ2X6Cc7wA+BTFHCZVugvY\nIiJmZWbzam8FjGTmQxV/1oTGGWf16v74RRsdHeubXNbFPPuLefafQcl1UPLstjJ9VH4/QfMtEbEc\nWALsOu2o1riGYlTRS4HLGm17AldV+BmSJKmmytxRWZv7KabWr0xmjkTEUmBJRCyimNDtaOCgKj9H\nkiTVU1WdaecBHwSun3ZERV+XVkdRzEx7IfAwcFxmnlPB50iSpJqrqjMtwG3AO6YZD5k53LY9Ahzc\n+Om5sbFRfvnLn0963C677Mrs2VX2I5YkSVV0pgVYmZn3TDeYOnr04fv5yClnMXf+tms95pEHbmfx\nUbBw4e49jEySpP5XqjNtROwNbJiZ5wNExEkRcXZmXlJ5hDUwd/62bLrVjjMdhiRJA2fK08pFxAHA\nj4A/b2l+HnBBRHRt4jVJkjR4ysx/+wngqMw8qdmQmW8BPgycUFVgkiRJZQqV5wPnTdD+fYqJ3yRJ\nkipRplC5nWK143YvA/qyQ60kSZoZZUb9/BPwhYh4PnBlo20PinlUPlVVYJIkSWVG/Xw+IuYAH6Do\nrwJwN/CJzDytyuAkSdJgKzWFfmYuBhZHxHxgVWYurzYsSZKkkoVKRDyTYhbaXYFVEXED8O8WLJIk\nqUpl5lHZlmJNn1OAlwOvAj4PXBcRz602PEmSNMjKjPo5GbgD2D4zF2bmAopp9X8PLK4yOEmSNNjK\nFCqvoZjwbVmzofH6w8BrqwpMkiSpTKGyGlgxQfsIMGd64UiSJK1RplC5FDguIjZsNjRef6KxT5Ik\nqRJlRv18FLgcuCUirm607QHMZeIZayVJkkqZ8h2VzLyJYuXkMyke9WwEfBNYkJnXVhueJEkaZGUn\nfPs9xZ0VSZKkrumoUImIr3V6wsxcVD4cSZKkNTq9o7J9y+tZwJ7AMuAXwCqKR0HPAc6pNDpJkjTQ\nOipUMvNVzdcR8RngTmBRZj7RaBsGvgSMdyNISZI0mMoMTz4U+PtmkQKQmaPAicD+VQUmSZJUplBZ\nCWw7QfvOwKPTC0eSJGmNMqN+zgC+GhHHAldTFDuvAE4ATq8wNkmSNODKTvi2CbAE2BAYAh4HvkBR\nrEiSJFViyoVKZq4EDo2Io4Gg6ECbmflY1cFJkqTBVmrCN4DMfBT4eYWxSJIkPUWZzrSSJEk9YaEi\nSZJqq6NCJSLeEBEbdTsYSZKkVp3eUTkD2AIgIm6NiPndC0mSJKnQaWfah4FPRsTFwHbAWyNi+UQH\nZubSimKTJEkDrtNC5RPAScAiiuHI/3stx40DFiqSJKkSnS5K+HXg6wARMQZsnZnLuhmYJElSmVE/\n2wP3AkTEsyNi02pDkiRJKky5UMnM3wPvi4h7gD8AD0TEXRHxwcqjkyRJA23KhUpEHAosBs4E9gPe\nDHwL+HRELKo2PEmSNMjKTKF/JPChzDytpe27EfFb4IPA1yqJTJIkDbwyfVSeB/znBO3/Bew4vXAk\nSZLWKFOo/B548QTtewCOBJIkSZUp8+jnS8AXI2Jz4NJG2yuB/wV8vqrAJEmSyhQqn6d4/HNq4/1D\nwCqKAuYfqgtNkiQNuikXKpk5BnwwIo4DXkhRqNyYmY9UHZwkSRpsZe6oANAoTK6qMBZJkqSnKF2o\naI2x0dVk3rTOY3bZZVdmz57do4gkSeoPFioVeOyhe/jqeXcz94pHJ9z/yAO3s/goWLhw9x5HJknS\n+s1CpSJz52/Lpls5jYwkSVUqM4X+hRMtRBgRfxIRv6wmLEmSpA7vqETEX7Fmkre9gWMiov05x47A\ndtWF1pmImAMsBbYEZgMfzMyf9ToOSZJUvU4f/dwGnEYxFBngAGC0Zf848Cjw4coi69wiiuHR+0fE\nC4BvAC+fgTgkSVLFOipUMvPXwA4AEfE7YI/MvL/qYBp3R64GDs/Mi1raTqdYqXkFcHJmntLytqUU\nhRLAhsATVcclSZJmRpkJ37bvRiCNguRMYOe2XScBuwH7UDxaWhoRt2XmWY14Hmu8/9kURcvR3YhP\nkiT13pQLlYjYCvh74BUUfUKGWvdn5g4lzrkTcMYE7ZsAhwCvzcxrgWsjYjFwBHBWy3EvAL4NfDwz\nfzLVz5ckSfVUZnjyl4HdgX8DHq4ojr2BC4BjKR7vNC2giPHylrZLgGOaGxGxDXAO8D8z8+qK4pEk\nSTVQplDZF3hdZl5cVRCZuaT5OiJad20N3J+Zq1valgEbRcT8zHyAorh5BrA4IoaAezNz/6piG5r8\nkI4MD89igw2mPBq8ss9u/W+/Ms/+Yp79Z1ByHbQ8u61MofIoRbHQC5vw9M6xze05AJl5aDcDmDVU\nTakyb97GbLbZMyo513RiGATm2V/Ms/8MSq6Dkme3lSlUlgIfiYhDM3N00qOn53EaBUmL5vYKemBs\nfHzygzqwfPkIDz74WCXnmqrh4VnMm7cxy5ePMDo6NiMx9IJ59hfz7D+Dkuug5dltZQqVLYC3Av8j\nIm6h7Y5HZu5bRWANdwFbRMSszGxe7a2Akcx8qMLPWatqyhQYHR1j9eqZ/YWtQwy9YJ79xTz7z6Dk\nOih5dlvZtX7OrDSKtbsGWAW8FLis0bYncFWPPr8SnayuDK6wLElSuzLzqBzcjUDW8lkjEbEUWBIR\ni4DnUsyTclCvYqjCZKsrgyssS5I0kTLzqOy1rv3NGWWnof1py1EUM9NeSDEc+rjMPGean9Fzrq4s\nSdLUlXn08xOKYqJ1OMx442eMYhK40jJzuG17BDi48SNJkgZImUKlfQr9DYAXAJ8CPjbtiCRJkhrK\n9FH5/QTNt0TEcmAJsOu0o5IkSQKqnFbufuDPKjyfJEkacFV1pp0HfBC4ftoRSZIkNVTVmRbgNuAd\n04xHkiTpSVV0pgVYmZn3TDcYSZKkVqU700bEXOCFFDPH3lJxXJIkSaX6qMwCTgIOAzakeAT0RER8\nCTgyM6taHkeSJA24Mo9+Pg4sAj4C/JRi5NBewPEUiwieWFl0kiRpoJUpVN4FHJaZZ7S0/TIi7gNO\nwEJFkiRVpMw8KlsCV07QfiWwzfTCkSRJWqNMofIb4NUTtL+GYoiyJElSJco8+jkF+FJE7ABc2mh7\nJXAE8KGqApMkSSozPHlpRGwOfBT4cKN5GXBsZp5eZXCSJGmwlbmjQmaeCpwaEc8GhjLz3mrDkiRJ\nmmIflYjYPyLmNbcz8z7gTRFxQES0T6kvSZI0LR0VKhGxYUScC5wB7Na2+5WN9m9FRKk7NJIkSRPp\n9I7KB4A9gH0y8yetOzLzfwL7Nn7eW2l0kiRpoHVaqBwEHJWZF0+0s1G8HAscUlFckiRJHRcq2wOX\nTXLM+cDzpxeOJEnSGp0WKo8Am05yzMbAiumFI0mStEanhcrlwAGTHPN24LrphSNJkrRGp6N0Pgdc\nEBF3Aqdn5lhzR2NY8vuAo4D9qg9RkiQNqo4Klcy8OCKOppg+/7iI+BnwIDAfeAkwF/hEZp7btUgl\nSdLA6XjCt8z8AvBi4FsUBcpfAPOArwILMnNxVyKUJEkDa0oTtGXmdRSPeVSxsdHVZN406XG77LIr\ns2fP7kFEkiTNvI4KlYj4HHB8Zi7v8PjNgU9m5vunE9wgeeyhe/jqeXcz94pH13rMIw/czuKjYOHC\n3XsYmSRJM6fTOyq3A9dHxJnAtzLz5xMdFBELgYOBvwVOrCbEwTF3/rZsutWOMx2GJEm10Wln2s9F\nxPeA44DLIuKPwPXAvcAw8GxgAUWflX8H9szMm7sTsiRJGhQd91HJzFuAd0bEMcDrKUb7bAWMAb8F\nvgmc21hRWZIkadqmvNpxZt4NfLnxI0mS1DVTLlQi4sC17BoHVgJ3Aldk5uh0AlP3rFy5khtu+NWk\nxznCSJI006ZcqFD0U9meYg6Whxttz6IoVIYa2xkRr8nMO6cfoqp2ww2/4iOnnMXc+duu9RhHGEmS\n6qBMoXI6xciet2fmrwAiYifgXygmfzsb+AqwGHhbRXGqYo4wkiStDzqembbFUcB7m0UKQGbeCBwB\nHJOZ9wDHAq+pJkRJkjSoyhQqm7LmkU+rFcDmjdcPAhuXDUqSJAnKFSoXA4sj4lnNhojYFPgMcFmj\n6W+AnH54kiRpkJXpo3IEcCFwZ0QkRbGzI3A/8LqIeA1F0bJ/ZVFKkqSBNOU7Kpl5K7AT8H6KOyg/\noSheIjMT+A2wa2aeVWGckiRpAJW5o0JmjkTEdyim0V8F3JKZKxv7fl9hfGrhCsuSpEFTZsK3WcBJ\nwGHAho3mlRHxJeDIzByvMD61cIVlSdKgKXNH5ePAIuAjwE8pHh/tBRwP3IWrJneV859IkgZJmULl\nXcBhmXlGS9svI+I+4AQsVCRJUkXKDE/eErhygvYrgW2mF44kSdIaZQqV3wCvnqD9NcBt04pGkiSp\nRZlHP6cAX4qIHYBLG22vpBii/KGqAlM5E40MGh6exbx5G7N8+Qijo2MdjRySJKkOplyoZObSiNgc\n+Cjw4UbzMuDYzDy9yuA0dZ2MDFp261VsucMePYxKkqRyys6jcipwakQ8GxjKzHurDUvTMdnIoEce\nuKOH0UiSVF6ZPipPysz7mkVKROwVEbdWE1Y5EfHGiPjaTMYgSZKqM61Cpc3GwPMqPN+URMRi4B9n\n6vMlSVL1Sj366aaImANcDRyemRe1tJ0O7AesAE7OzFPa3nolcB5wUA/DlSRJXVTlHZVpaxQkZwI7\nt+06CdgN2Idi6v7jI2K/1gMy8z96EaMkSeqd2hQqEbETcAWwfVv7JsAhwPsz89rMPAdYTDEcWpIk\n9bGOHv1ExN91cNh0F6DZG7gAOJbi8U7TAoo4L29puwQ4ZpqfJ0mSaq7TPioHd3jc7WUDycwlzdcR\n0bpra+D+zFzd0rYM2Cgi5mfmA2U/sxND3Tx5jY2Nrubmm5Ph4XXfdHvRi3Zl9uzZPYpq3ZqxThbz\n+s48+8ug5AmDk+ug5dltHRUqmbn95Ed1zSbAE21tze05rY2Z+VOKFZ0rM2toMEuVxx66hy+fezdz\nL3tkrcc88sDtfPlTG7PHHvWaPG7evI1nOoSeMM/+Mih5wuDkOih5dlvtRv1M4HHaCpKW7RV02dj4\neLc/orYmmzgOYPnyER588LEeRbRu7UsF9Cvz7C+DkicMTq6Dlme3rQ+Fyl3AFhExKzObV3wrYCQz\nH+r2hw9umdKZ0dExVq+u11/EOsbUDebZXwYlTxicXAclz25bHx6gXQOsAl7a0rYncNXMhCNJknql\n9ndUMnMkIpYCSyJiEfBc4Gic2E2SpL5X10Kl/YnLURQz014IPAwc15hPRZIk9bFaFiqZOdy2PUIx\nRLrTYdKSJKkPrA99VCRJ0oCyUJEkSbVloSJJkmrLQkWSJNWWhYokSaotCxVJklRbtRyeLLVauXIl\nN9zwq0mPW7BgAfCM7gckSeoZCxXV3g03/IqPnHIWc+dvu9ZjHnngdk7+8Cy23HKvHkYmSeo2CxWt\nFzpZyVmS1H/soyJJkmrLQkWSJNWWj36kKei0Y+8uu+zK7NmzexCRJPU3CxVpCjrt2Lv4KFi4cPce\nRiZJ/clCRZoiO/ZKUu/YR0WSJNWWhYokSaotCxVJklRb9lFRaWOjq8m8aZ3HOPpFkjQdFioq7bGH\n7uGr593N3CsenXC/o18kSdNloaJpcQSMJKmb7KMiSZJqy0JFkiTVloWKJEmqLQsVSZJUWxYqkiSp\nthz1oxnVyWrEk83VAsWcLjfddCPz5m3M8uUjjI6OPe2YVatWAbDhhhuu81y9mPvFVZglqTMWKppR\nnaxGvOzWq9hyhz3WeZ7HHrqHL597N/922SPrPM8mz9qyFisfuwqzJHXGQkUzbrK5WB554I7KzjN3\n/ja1mffFOWgkaXL2UZEkSbVloSJJkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2LFQkSVJtWahI\nkqTacsI3dc3Y6OpJp7/vZHr89U1Vea/tPMPDs56yVEAV0+x3MqV/r6bzb8bSnudU46lTTpLKs1BR\n1zz20D189by7mXvFo2s9ppPp8dc3VeXdyXmqmmZ/sin9ezmdf1XLC9QpJ0nlWaioq6qaHn9906tl\nAapUpyn9q4qlTjlJKsc+KpIkqbYsVCRJUm1ZqEiSpNqyUJEkSbVloSJJkmrLQkWSJNWWhYokSaot\nCxVJklRbFiqSJKm21vuZaSNiCPgKEMBy4MDMvH9mo5IkSVXohzsqbwJWZOYrgW8Ax8xsOJIkqSq1\nuqMSEXOAq4HDM/OilrbTgf2AFcDJmXlKy9teAfyfxuv/Aj7Wu4glSVI31eaOSqMgORPYuW3XScBu\nwD7AYcDxEbFfy/55FI98AB4BntndSCVJUq/UolCJiJ2AK4Dt29o3AQ4B3p+Z12bmOcBi4IiWw5YD\ncxuv5wIPdz9iSZLUC7UoVIC9gQuAlwFDLe0LKB5PXd7Sdgnwkpbty4G/bLz+K+Cy7oUpSZJ6qRZ9\nVDJzSfN1RLTu2hq4PzNXt7QtAzaKiPmZ+QBwFvDfI+ISYCVwQJWxDU1+iPrI2Ohqbr45GR6euIa/\n+ebscUTrNlm8TS960a7Mnj17rfsne39Vn7Ny5Uquv/5X6zxHp3/Gw8Oz2GCDtcczWaydnKMXmnGO\njq7muuuunfT46f4Zr1q1CoANN9xwWp9TRjPX1mvTye9Et+LplonybFdF3jP9Z9fJ37EqDI2Pj/fk\ngzoVEWPAPpl5UUS8A/hUZm7fsn974LfANpl590zFKUmSuq8uj37W5nFgTltbc3tFj2ORJEk9VvdC\n5S5gi4hojXMrYCQzH5qhmCRJUo/UvVC5BlgFvLSlbU/gqpkJR5Ik9VItOtOuTWaORMRSYElELAKe\nCxwNHDSzkUmSpF6oY6HS3rv3KIqZaS+kmCPluMZ8KpIkqc/VbtSPJElSU937qEiSpAFmoSJJkmrL\nQkWSJNWnzAOrAAAMrUlEQVSWhYokSaqtOo766YmImEMxmmg/illuT87MU2Y2qqmLiDdSrHc0TrE0\n0TjwH5n5txGxHfBlisUebwOOzMwftbz31cDngB0oFnd8d2b+rqcJTKJxna4GDs/Mixpt2zGNvCLi\ng8CHKFbb/jZwRGY+3ot81mYteX4eeB9Pvbbvy8zTG/vXmzwj4k+B/w28iuLv27eAj2fmyn66npPk\n2TfXsxHP84EvAq8AHgBOy8yTGvu2o3+u6bry7Ktr2ojpPGBZZi5qbG/HDF/LQb6jchKwG7APcBhw\nfETsN6MRlbMz8D2KGXu3oljI8V2NfecAdwO7A/8KfDcingsQEdsA3wW+CrwYuB84u6eRT6Lxj/eZ\nFDm2OpuSeUXE3wB/B7wb2JdiMsHFXU1kEuvIcyfgoxTXtHltv9Z4z/qW538AG1F82R8A/L/Apxr7\nSv+ermd59s31jIgh4DyKRWL/HPj/gGMjorkobF9c0w7y7Jtr2ojpAOC/tzXP+PftQBYqEbEJcAjw\n/sy8tjEvy2LgiJmNrJSdgOsz877MvLfxszwi9gW2Bw7Nwmcoqt1Fjfe9G7gqM0/NzBuBg4HtImKv\nGcmiTUTsBFxBkUNr+74UlXvZvN4PfC4z/zMzfw4cChwSERt1P6unW1ueDTsBv2y5rve2/J/Iu1hP\n8oxiSfS/AN6ZmTdl5qUUX15vi4hXMb3f0/Uiz8YhfXE9G7YEfgkclpm3ZOZ/ARcAr+yna8o68mzs\n75trGhGbUfw7+LOWtlp83w5koQIsoHjsdXlL2yXAS2YmnGnZGfjNBO0vAX7RdovtEorbd839FzV3\nZOYI8IuW/TNtb4ovhJdR3FJtKp1XY82oPYCLW957BTCb4ndiJkyYZ0TMBZ7DxNcWiv8zWV/y/APw\nusy8v639WRR59Mv1nCjPIeBZfXY9ycw/ZOZbM/MxgIh4BcXyJj+hj67pWvLcC/hxv11TiqcMS4Eb\nW9pq8X07qIXK1sD9mbm6pW0ZsFFEzJ+hmMoK4HURkRHx24j4dERsSJHj3W3HLqNYhoAO9s+ozFyS\nmR+a4FnmdPLalOK2/JP7M3OU4rnzjOS9jjx3onjefWxE3BER10TEgS3715s8M/PhtmfaQxR3Ly+g\nj67nOvI8nz66nu0i4jaKf6wup+gv1zfXtFVLnpdR5LkzfXJNG3dO9mTNY8qmWlzLQS1UNgGeaGtr\nbs/pcSylRcS2wMbACPAWinWQ3gacyNpzbOY32f66mk5em7Rsr+39dfFCYAz4NcUz468A/xwRb2js\nX5/zPBFYCHyC/r6eJ1L0aziW/r6e+1H0xflzik6V/XpNm3kuBE6l+J/E9f6aNvrILaF4vNUeTy2u\n5aCO+nmcp/9BNbdX9DiW0jLz9oiYn5kPNZqui4hhig5PXwc2a3vLHNbkt7Y/gwe7FW9FHgc2b2vr\nNK/HW7bX9v5ayMylEfG9lmt7fUS8AHgvRUfF9TLPiPgsxXPrv83MX0dEX17P9jyBX/fj9QTIzF8A\nRMRRwDcpOlaW/e6pba4teR5J8R07D+iHa/pJin4m50+wrxZ/Pwf1jspdwBaNZ2hNWwEjLb9064UJ\n4r2R4nbbHyhyarUVcE/j9V2T7K+ryeJe1/4HKP7yPLm/UdjNp4Z5r+XaPqfxer3LMyK+ABwJvD0z\nmyMD+u56riXPvrqeEfEnLXcOmn5N0f/gHvrkmk6S59w+uab7A2+MiEci4hHg7cA7ImI5cCc1uJaD\nWqhcA6yi6OzUtCdw1cyEU05E/GVE3N/Wg3ohxRCxi4HdG7f1ml5J0ZmJxn+bPdebI6EWtuyvqyuA\n3UrkdXlmjlNc41e2vPflwErg2m4GPVURcUJE/KiteSFwU+P1epVnRBwPvAfYPzO/3bKrr67n2vLs\nt+tJMarnrIjYuqXtxcC9FJ0ty3z31DHXteV5H/CBPrmmewO7UnRwXUAx3cU5jddXUoO/nwO7enJE\n/BPFXAeLKDr2fAM4KIuhyuuFiHgmRXV/EfC/gOdTTMzzucbPdcCvKDpI/TXwcWCXzLwzIp7XeO8J\nwPeB44EdM3O3XucxmYgYA/bJzIsad8GuBa6n87xekJkLG+fan+J57DspOnl9DTg/M4/sbVZP15bn\ni4FLKXI7G3gtcHJj/8/WpzwbQ7CvA/6RYpLFVvfRJ9dzkjy3oU+uZyOeWRSdZ/8IHEXxD/pXgX+g\nyH2q3z21zHWSPK+gj65pU0R8HRjPzEV1+b4d1DsqUPzS/Ry4EPgCcNz6VKQAZOajFH85nk1RuX4Z\nWJKZJ2fmGMUv1VYUM56+DXhjZt7ZeO/vKTqHLaIYN78p8KaeJ9GZJ6vpRl5vYGp5vbHl/f8OfBr4\nEvBDii+hj/Yki8m15nk18GbgQIov/COAt2bmzxr716c8/5riu+ZYii+ruylu/d7duJ5vpD+u57ry\n7Kfr2fr38DGKUTD/DJyamaeV/O6pZa6T5NlX13Qidfm+Hdg7KpIkqf4G+Y6KJEmqOQsVSZJUWxYq\nkiSptixUJElSbVmoSJKk2rJQkSRJtWWhIkmSastCRZIk1ZaFiiRJqi0LFalPRcRtEXFrRDxjgn1f\nj4gLu/z5v4uIv+vmZ3QqInaPiBsjYiQiFld43rGIOLCq80l6OgsVqX+NA88DTpzpQGrgGIol53ei\nWHtE0nrCQkXqb7cCh0bEvjMdyAzbDLgmM2/LzAdnOhhJndtgpgOQ1FX/CrwC+FpE7JKZj010UESM\nAe/MzKUTtUXE8cArgYuAw4FNgDOAvwf+CdiXYrXgD2TmD1pO/acR8QPgVRQrCZ+Umae3fMbLKe5w\n7AHcB5wLfDwzH2ns/x3wHeCvKFYJ/5vMvHiC+F9PsWrxi4BHgDOBYzLzicY5tgWGGo9pts/M29ve\nPwR8DDgI2A54ArgUODwzf9c45jnA6Y1cHqJtFdiI2JhiJfbXU6wieyPwqcz87kR/5pI64x0Vqb+N\nA4dQ3FE4eZrn2gsIioLlfcB7KJZ2PxPYjeIf5q+3veddwE+BXYFTgM9HxBsAIuK/AT8CfkBRYLy1\ncZ7/03aOw4EjgNcBV7QHFRFvAs4BvgcsbMS1fyMugBc33vfvFMvV3zFBbh8AjgaOBHakWNr+BTT+\nzCJimGKZ+s2BPYG3AB+m+PNt+vtGHq8DXgj8J/BvEbHtBJ8nqUPeUZH6XGbeEREfAr4UEd/JzPNL\nnmoIeE9mrgB+GxEnAudn5hkAEXE68PqI2DIzlzXe893M/Gzj9WkR8VKKguAcin/of9iy/9aIeDtw\nS0TslZkXNdp/kJk/XkdcHwX+IzObfU9+GxGHAWdHxAsz86aIWAmMZOZ9aznHzcCBmfmfje07IuLb\nwJsb26+m6N/y/My8rZHvwcAvW86xA8XdnNsy8+GIOA74CeCjJmkaLFSkAZCZX46INwNfiYhdS55m\nWaNIaXqMog9M00jjv3Na2i5tO8eVFI9xoLj78WcR8UjbMeMURUGzULl5krh2pXgM1eqnLftumuT9\nZOZ5EfEXEXECxV2jAHYB7mwc8iLgwWaR0njPtREx0nKaz1Lc1bkvIq6kuDN0RvMxlqRyfPQjDY53\nAc+ieASzTo1HHe1WTdA2NsmpRtu2hyn6f0Dx/fNN4L8BC1p+duSphUdrMTCRoQnamt9tE8X8NBHx\nMeDHwHzgfOBQ4KSWQ8aZ+PvyyfNn5hXANsB+wM+BA4EbI+JVncQgaWIWKtKAyMw7gA9R9FnZs233\nKmBey/YLKvrY3du29wR+1Xh9PbBzZv4uM2/NzFuB2cCpFP/gd+o6in4zrfaiKC5+3eE5Pg58MjOP\nyMyvZObPKO6qNIuga4BnRcROzTdExI60/JlFxCeBPTPz+5n5wcb7bwH+Zgq5SGrjox9pgGTmVyPi\nLcBfAq0jXy4H3h0RF1P8D8wpFPOOTFX73Y23RsR1wPeBN1F0Um3eYTgZuCgiTgNOo+jw+0VgI+A3\nU/jMxcC3IuITwLcoCoQvAOdmZqfnuQP4y4j4PsVdoAMb8f6hsf/HFB2H/yUiDm8c8wWeesdoB+Dt\nEfEeigLlpRSjjdoff0maAu+oSP1rfC3t76IYXtu6/73AHykKlm8D/8ya/hlTOf942+sTKYbrXgu8\nE3hrc3hxZl4JvJbicc/PgbMpRg69OjNXT5LDkzLzLIoRQ2+huLtyOsUjpf0ne2+Ld1AMub6Kon/L\nLhSPf/4kIp6bmeMUfWtuohj9cy7F46nWzrmHARcA/wIkcALwkcw8E0mlDY2PT/o9IEmSNCO8oyJJ\nkmrLQkWSJNWWhYokSaotCxVJklRbFiqSJKm2LFQkSVJtWahIkqTaslCRJEm1ZaEiSZJqy0JFkiTV\nloWKJEmqrf8LYr23x47lfwEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x3679114d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print \"Is every campaign_id corresponds to only one ad_id?\"\n",
    "a = promoted_content.groupby(\"campaign_id\")[\"ad_id\"].count()\n",
    "print \"The answer is: \", sum(a) == len(a)\n",
    "plt.hist(a,bins = 50, log = True)\n",
    "plt.xlabel(\"Number of ads\")\n",
    "plt.ylabel(\"Log(Count of document_id)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Events"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jieli/anaconda/lib/python2.7/site-packages/IPython/core/interactiveshell.py:2723: DtypeWarning: Columns (4) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    }
   ],
   "source": [
    "events = pd.read_csv(\"events.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'DataFrame' object has no attribute 'topics'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-11-47d9f157978b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mevents\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtopics\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/Users/jieli/anaconda/lib/python2.7/site-packages/pandas/core/generic.pyc\u001b[0m in \u001b[0;36m__getattr__\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m   2670\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_info_axis\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2671\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2672\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mobject\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__getattribute__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2673\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2674\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m__setattr__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'DataFrame' object has no attribute 'topics'"
     ]
    }
   ],
   "source": [
    "events"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'clicks_train' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-5da5e0e37a64>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mclicks_train\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'clicks_train' is not defined"
     ]
    }
   ],
   "source": [
    "clicks_train[1:100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ad_id</th>\n",
       "      <th>document_id</th>\n",
       "      <th>campaign_id</th>\n",
       "      <th>advertiser_id</th>\n",
       "      <th>num_appr</th>\n",
       "      <th>num_click</th>\n",
       "      <th>click_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ad_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>144739</th>\n",
       "      <td>144739</td>\n",
       "      <td>1337362</td>\n",
       "      <td>18488</td>\n",
       "      <td>2909</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ad_id  document_id  campaign_id  advertiser_id  num_appr  num_click  \\\n",
       "ad_id                                                                          \n",
       "144739  144739      1337362        18488           2909       1.0        0.0   \n",
       "\n",
       "        click_rate  \n",
       "ad_id               \n",
       "144739         0.0  "
      ]
     },
     "execution_count": 356,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "promoted_content[promoted_content[\"ad_id\"] == 144739]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "documents_topics = pd.read_csv(\"documents_topics.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2495423"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(documents_topics[\"document_id\"].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 357,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>document_id</th>\n",
       "      <th>topic_id</th>\n",
       "      <th>confidence_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>10960736</th>\n",
       "      <td>1337362</td>\n",
       "      <td>131</td>\n",
       "      <td>0.272269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10960737</th>\n",
       "      <td>1337362</td>\n",
       "      <td>198</td>\n",
       "      <td>0.212531</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          document_id  topic_id  confidence_level\n",
       "10960736      1337362       131          0.272269\n",
       "10960737      1337362       198          0.212531"
      ]
     },
     "execution_count": 357,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "documents_topics[documents_topics[\"document_id\"] == 1337362]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 359,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "559583"
      ]
     },
     "execution_count": 359,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(promoted_content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'events' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-8-bb684a3229af>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mevents\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'events' is not defined"
     ]
    }
   ],
   "source": [
    "events.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_prob(k):\n",
    "    return cc[k]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "p = clicks_test.groupby('display_id').ad_id.apply(list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0016011443699\n",
      "0.00320228873981\n",
      "0.00480343310971\n",
      "0.00640457747962\n",
      "0.00800572184952\n",
      "0.00960686621942\n",
      "0.0112080105893\n",
      "0.0128091549592\n",
      "0.0144102993291\n",
      "0.016011443699\n",
      "0.0176125880689\n",
      "0.0192137324388\n",
      "0.0208148768088\n",
      "0.0224160211787\n",
      "0.0240171655486\n",
      "0.0256183099185\n",
      "0.0272194542884\n",
      "0.0288205986583\n",
      "0.0304217430282\n",
      "0.0320228873981\n",
      "0.033624031768\n",
      "0.0352251761379\n",
      "0.0368263205078\n",
      "0.0384274648777\n",
      "0.0400286092476\n",
      "0.0416297536175\n",
      "0.0432308979874\n",
      "0.0448320423573\n",
      "0.0464331867272\n",
      "0.0480343310971\n",
      "0.049635475467\n",
      "0.0512366198369\n",
      "0.0528377642068\n",
      "0.0544389085767\n",
      "0.0560400529466\n",
      "0.0576411973165\n",
      "0.0592423416865\n",
      "0.0608434860564\n",
      "0.0624446304263\n",
      "0.0640457747962\n",
      "0.0656469191661\n",
      "0.067248063536\n",
      "0.0688492079059\n",
      "0.0704503522758\n",
      "0.0720514966457\n",
      "0.0736526410156\n",
      "0.0752537853855\n",
      "0.0768549297554\n",
      "0.0784560741253\n",
      "0.0800572184952\n",
      "0.0816583628651\n",
      "0.083259507235\n",
      "0.0848606516049\n",
      "0.0864617959748\n",
      "0.0880629403447\n",
      "0.0896640847146\n",
      "0.0912652290845\n",
      "0.0928663734544\n",
      "0.0944675178243\n",
      "0.0960686621942\n",
      "0.0976698065641\n",
      "0.0992709509341\n",
      "0.100872095304\n",
      "0.102473239674\n",
      "0.104074384044\n",
      "0.105675528414\n",
      "0.107276672784\n",
      "0.108877817153\n",
      "0.110478961523\n",
      "0.112080105893\n",
      "0.113681250263\n",
      "0.115282394633\n",
      "0.116883539003\n",
      "0.118484683373\n",
      "0.120085827743\n",
      "0.121686972113\n",
      "0.123288116483\n",
      "0.124889260853\n",
      "0.126490405222\n",
      "0.128091549592\n",
      "0.129692693962\n",
      "0.131293838332\n",
      "0.132894982702\n",
      "0.134496127072\n",
      "0.136097271442\n",
      "0.137698415812\n",
      "0.139299560182\n",
      "0.140900704552\n",
      "0.142501848921\n",
      "0.144102993291\n",
      "0.145704137661\n",
      "0.147305282031\n",
      "0.148906426401\n",
      "0.150507570771\n",
      "0.152108715141\n",
      "0.153709859511\n",
      "0.155311003881\n",
      "0.156912148251\n",
      "0.158513292621\n",
      "0.16011443699\n",
      "0.16171558136\n",
      "0.16331672573\n",
      "0.1649178701\n",
      "0.16651901447\n",
      "0.16812015884\n",
      "0.16972130321\n",
      "0.17132244758\n",
      "0.17292359195\n",
      "0.17452473632\n",
      "0.176125880689\n",
      "0.177727025059\n",
      "0.179328169429\n",
      "0.180929313799\n",
      "0.182530458169\n",
      "0.184131602539\n",
      "0.185732746909\n",
      "0.187333891279\n",
      "0.188935035649\n",
      "0.190536180019\n",
      "0.192137324388\n",
      "0.193738468758\n",
      "0.195339613128\n",
      "0.196940757498\n",
      "0.198541901868\n",
      "0.200143046238\n",
      "0.201744190608\n",
      "0.203345334978\n",
      "0.204946479348\n",
      "0.206547623718\n",
      "0.208148768088\n",
      "0.209749912457\n",
      "0.211351056827\n",
      "0.212952201197\n",
      "0.214553345567\n",
      "0.216154489937\n",
      "0.217755634307\n",
      "0.219356778677\n",
      "0.220957923047\n",
      "0.222559067417\n",
      "0.224160211787\n",
      "0.225761356156\n",
      "0.227362500526\n",
      "0.228963644896\n",
      "0.230564789266\n",
      "0.232165933636\n",
      "0.233767078006\n",
      "0.235368222376\n",
      "0.236969366746\n",
      "0.238570511116\n",
      "0.240171655486\n",
      "0.241772799856\n",
      "0.243373944225\n",
      "0.244975088595\n",
      "0.246576232965\n",
      "0.248177377335\n",
      "0.249778521705\n",
      "0.251379666075\n",
      "0.252980810445\n",
      "0.254581954815\n",
      "0.256183099185\n",
      "0.257784243555\n",
      "0.259385387924\n",
      "0.260986532294\n",
      "0.262587676664\n",
      "0.264188821034\n",
      "0.265789965404\n",
      "0.267391109774\n",
      "0.268992254144\n",
      "0.270593398514\n",
      "0.272194542884\n",
      "0.273795687254\n",
      "0.275396831623\n",
      "0.276997975993\n",
      "0.278599120363\n",
      "0.280200264733\n",
      "0.281801409103\n",
      "0.283402553473\n",
      "0.285003697843\n",
      "0.286604842213\n",
      "0.288205986583\n",
      "0.289807130953\n",
      "0.291408275323\n",
      "0.293009419692\n",
      "0.294610564062\n",
      "0.296211708432\n",
      "0.297812852802\n",
      "0.299413997172\n",
      "0.301015141542\n",
      "0.302616285912\n",
      "0.304217430282\n",
      "0.305818574652\n",
      "0.307419719022\n",
      "0.309020863391\n",
      "0.310622007761\n",
      "0.312223152131\n",
      "0.313824296501\n",
      "0.315425440871\n",
      "0.317026585241\n",
      "0.318627729611\n",
      "0.320228873981\n",
      "0.321830018351\n",
      "0.323431162721\n",
      "0.325032307091\n",
      "0.32663345146\n",
      "0.32823459583\n",
      "0.3298357402\n",
      "0.33143688457\n",
      "0.33303802894\n",
      "0.33463917331\n",
      "0.33624031768\n",
      "0.33784146205\n",
      "0.33944260642\n",
      "0.34104375079\n",
      "0.342644895159\n",
      "0.344246039529\n",
      "0.345847183899\n",
      "0.347448328269\n",
      "0.349049472639\n",
      "0.350650617009\n",
      "0.352251761379\n",
      "0.353852905749\n",
      "0.355454050119\n",
      "0.357055194489\n",
      "0.358656338859\n",
      "0.360257483228\n",
      "0.361858627598\n",
      "0.363459771968\n",
      "0.365060916338\n",
      "0.366662060708\n",
      "0.368263205078\n",
      "0.369864349448\n",
      "0.371465493818\n",
      "0.373066638188\n",
      "0.374667782558\n",
      "0.376268926927\n",
      "0.377870071297\n",
      "0.379471215667\n",
      "0.381072360037\n",
      "0.382673504407\n",
      "0.384274648777\n",
      "0.385875793147\n",
      "0.387476937517\n",
      "0.389078081887\n",
      "0.390679226257\n",
      "0.392280370626\n",
      "0.393881514996\n",
      "0.395482659366\n",
      "0.397083803736\n",
      "0.398684948106\n",
      "0.400286092476\n",
      "0.401887236846\n",
      "0.403488381216\n",
      "0.405089525586\n",
      "0.406690669956\n",
      "0.408291814326\n",
      "0.409892958695\n",
      "0.411494103065\n",
      "0.413095247435\n",
      "0.414696391805\n",
      "0.416297536175\n",
      "0.417898680545\n",
      "0.419499824915\n",
      "0.421100969285\n",
      "0.422702113655\n",
      "0.424303258025\n",
      "0.425904402394\n",
      "0.427505546764\n",
      "0.429106691134\n",
      "0.430707835504\n",
      "0.432308979874\n",
      "0.433910124244\n",
      "0.435511268614\n",
      "0.437112412984\n",
      "0.438713557354\n",
      "0.440314701724\n",
      "0.441915846094\n",
      "0.443516990463\n",
      "0.445118134833\n",
      "0.446719279203\n",
      "0.448320423573\n",
      "0.449921567943\n",
      "0.451522712313\n",
      "0.453123856683\n",
      "0.454725001053\n",
      "0.456326145423\n",
      "0.457927289793\n",
      "0.459528434162\n",
      "0.461129578532\n",
      "0.462730722902\n",
      "0.464331867272\n",
      "0.465933011642\n",
      "0.467534156012\n",
      "0.469135300382\n",
      "0.470736444752\n",
      "0.472337589122\n",
      "0.473938733492\n",
      "0.475539877862\n",
      "0.477141022231\n",
      "0.478742166601\n",
      "0.480343310971\n",
      "0.481944455341\n",
      "0.483545599711\n",
      "0.485146744081\n",
      "0.486747888451\n",
      "0.488349032821\n",
      "0.489950177191\n",
      "0.491551321561\n",
      "0.49315246593\n",
      "0.4947536103\n",
      "0.49635475467\n",
      "0.49795589904\n",
      "0.49955704341\n",
      "0.50115818778\n",
      "0.50275933215\n",
      "0.50436047652\n",
      "0.50596162089\n",
      "0.50756276526\n",
      "0.509163909629\n",
      "0.510765053999\n",
      "0.512366198369\n",
      "0.513967342739\n",
      "0.515568487109\n",
      "0.517169631479\n",
      "0.518770775849\n",
      "0.520371920219\n",
      "0.521973064589\n",
      "0.523574208959\n",
      "0.525175353329\n",
      "0.526776497698\n",
      "0.528377642068\n",
      "0.529978786438\n",
      "0.531579930808\n",
      "0.533181075178\n",
      "0.534782219548\n",
      "0.536383363918\n",
      "0.537984508288\n",
      "0.539585652658\n",
      "0.541186797028\n",
      "0.542787941397\n",
      "0.544389085767\n",
      "0.545990230137\n",
      "0.547591374507\n",
      "0.549192518877\n",
      "0.550793663247\n",
      "0.552394807617\n",
      "0.553995951987\n",
      "0.555597096357\n",
      "0.557198240727\n",
      "0.558799385097\n",
      "0.560400529466\n",
      "0.562001673836\n",
      "0.563602818206\n",
      "0.565203962576\n",
      "0.566805106946\n",
      "0.568406251316\n",
      "0.570007395686\n",
      "0.571608540056\n",
      "0.573209684426\n",
      "0.574810828796\n",
      "0.576411973165\n",
      "0.578013117535\n",
      "0.579614261905\n",
      "0.581215406275\n",
      "0.582816550645\n",
      "0.584417695015\n",
      "0.586018839385\n",
      "0.587619983755\n",
      "0.589221128125\n",
      "0.590822272495\n",
      "0.592423416865\n",
      "0.594024561234\n",
      "0.595625705604\n",
      "0.597226849974\n",
      "0.598827994344\n",
      "0.600429138714\n",
      "0.602030283084\n",
      "0.603631427454\n",
      "0.605232571824\n",
      "0.606833716194\n",
      "0.608434860564\n",
      "0.610036004933\n",
      "0.611637149303\n",
      "0.613238293673\n",
      "0.614839438043\n",
      "0.616440582413\n",
      "0.618041726783\n",
      "0.619642871153\n",
      "0.621244015523\n",
      "0.622845159893\n",
      "0.624446304263\n",
      "0.626047448632\n",
      "0.627648593002\n",
      "0.629249737372\n",
      "0.630850881742\n",
      "0.632452026112\n",
      "0.634053170482\n",
      "0.635654314852\n",
      "0.637255459222\n",
      "0.638856603592\n",
      "0.640457747962\n",
      "0.642058892332\n",
      "0.643660036701\n",
      "0.645261181071\n",
      "0.646862325441\n",
      "0.648463469811\n",
      "0.650064614181\n",
      "0.651665758551\n",
      "0.653266902921\n",
      "0.654868047291\n",
      "0.656469191661\n",
      "0.658070336031\n",
      "0.6596714804\n",
      "0.66127262477\n",
      "0.66287376914\n",
      "0.66447491351\n",
      "0.66607605788\n",
      "0.66767720225\n",
      "0.66927834662\n",
      "0.67087949099\n",
      "0.67248063536\n",
      "0.67408177973\n",
      "0.6756829241\n",
      "0.677284068469\n",
      "0.678885212839\n",
      "0.680486357209\n",
      "0.682087501579\n",
      "0.683688645949\n",
      "0.685289790319\n",
      "0.686890934689\n",
      "0.688492079059\n",
      "0.690093223429\n",
      "0.691694367799\n",
      "0.693295512168\n",
      "0.694896656538\n",
      "0.696497800908\n",
      "0.698098945278\n",
      "0.699700089648\n",
      "0.701301234018\n",
      "0.702902378388\n",
      "0.704503522758\n",
      "0.706104667128\n",
      "0.707705811498\n",
      "0.709306955867\n",
      "0.710908100237\n",
      "0.712509244607\n",
      "0.714110388977\n",
      "0.715711533347\n",
      "0.717312677717\n",
      "0.718913822087\n",
      "0.720514966457\n",
      "0.722116110827\n",
      "0.723717255197\n",
      "0.725318399567\n",
      "0.726919543936\n",
      "0.728520688306\n",
      "0.730121832676\n",
      "0.731722977046\n",
      "0.733324121416\n",
      "0.734925265786\n",
      "0.736526410156\n",
      "0.738127554526\n",
      "0.739728698896\n",
      "0.741329843266\n",
      "0.742930987635\n",
      "0.744532132005\n",
      "0.746133276375\n",
      "0.747734420745\n",
      "0.749335565115\n",
      "0.750936709485\n",
      "0.752537853855\n",
      "0.754138998225\n",
      "0.755740142595\n",
      "0.757341286965\n",
      "0.758942431335\n",
      "0.760543575704\n",
      "0.762144720074\n",
      "0.763745864444\n",
      "0.765347008814\n",
      "0.766948153184\n",
      "0.768549297554\n",
      "0.770150441924\n",
      "0.771751586294\n",
      "0.773352730664\n",
      "0.774953875034\n",
      "0.776555019403\n",
      "0.778156163773\n",
      "0.779757308143\n",
      "0.781358452513\n",
      "0.782959596883\n",
      "0.784560741253\n",
      "0.786161885623\n",
      "0.787763029993\n",
      "0.789364174363\n",
      "0.790965318733\n",
      "0.792566463103\n",
      "0.794167607472\n",
      "0.795768751842\n",
      "0.797369896212\n",
      "0.798971040582\n",
      "0.800572184952\n",
      "0.802173329322\n",
      "0.803774473692\n",
      "0.805375618062\n",
      "0.806976762432\n",
      "0.808577906802\n",
      "0.810179051171\n",
      "0.811780195541\n",
      "0.813381339911\n",
      "0.814982484281\n",
      "0.816583628651\n",
      "0.818184773021\n",
      "0.819785917391\n",
      "0.821387061761\n",
      "0.822988206131\n",
      "0.824589350501\n",
      "0.82619049487\n",
      "0.82779163924\n",
      "0.82939278361\n",
      "0.83099392798\n",
      "0.83259507235\n",
      "0.83419621672\n",
      "0.83579736109\n",
      "0.83739850546\n",
      "0.83899964983\n",
      "0.8406007942\n",
      "0.84220193857\n",
      "0.843803082939\n",
      "0.845404227309\n",
      "0.847005371679\n",
      "0.848606516049\n",
      "0.850207660419\n",
      "0.851808804789\n",
      "0.853409949159\n",
      "0.855011093529\n",
      "0.856612237899\n",
      "0.858213382269\n",
      "0.859814526638\n",
      "0.861415671008\n",
      "0.863016815378\n",
      "0.864617959748\n",
      "0.866219104118\n",
      "0.867820248488\n",
      "0.869421392858\n",
      "0.871022537228\n",
      "0.872623681598\n",
      "0.874224825968\n",
      "0.875825970338\n",
      "0.877427114707\n",
      "0.879028259077\n",
      "0.880629403447\n",
      "0.882230547817\n",
      "0.883831692187\n",
      "0.885432836557\n",
      "0.887033980927\n",
      "0.888635125297\n",
      "0.890236269667\n",
      "0.891837414037\n",
      "0.893438558406\n",
      "0.895039702776\n",
      "0.896640847146\n",
      "0.898241991516\n",
      "0.899843135886\n",
      "0.901444280256\n",
      "0.903045424626\n",
      "0.904646568996\n",
      "0.906247713366\n",
      "0.907848857736\n",
      "0.909450002106\n",
      "0.911051146475\n",
      "0.912652290845\n",
      "0.914253435215\n",
      "0.915854579585\n",
      "0.917455723955\n",
      "0.919056868325\n",
      "0.920658012695\n",
      "0.922259157065\n",
      "0.923860301435\n",
      "0.925461445805\n",
      "0.927062590174\n",
      "0.928663734544\n",
      "0.930264878914\n",
      "0.931866023284\n",
      "0.933467167654\n",
      "0.935068312024\n",
      "0.936669456394\n",
      "0.938270600764\n",
      "0.939871745134\n",
      "0.941472889504\n",
      "0.943074033873\n",
      "0.944675178243\n",
      "0.946276322613\n",
      "0.947877466983\n",
      "0.949478611353\n",
      "0.951079755723\n",
      "0.952680900093\n",
      "0.954282044463\n",
      "0.955883188833\n",
      "0.957484333203\n",
      "0.959085477573\n",
      "0.960686621942\n",
      "0.962287766312\n",
      "0.963888910682\n",
      "0.965490055052\n",
      "0.967091199422\n",
      "0.968692343792\n",
      "0.970293488162\n",
      "0.971894632532\n",
      "0.973495776902\n",
      "0.975096921272\n",
      "0.976698065641\n",
      "0.978299210011\n",
      "0.979900354381\n",
      "0.981501498751\n",
      "0.983102643121\n",
      "0.984703787491\n",
      "0.986304931861\n",
      "0.987906076231\n",
      "0.989507220601\n",
      "0.991108364971\n",
      "0.992709509341\n",
      "0.99431065371\n",
      "0.99591179808\n",
      "0.99751294245\n",
      "0.99911408682\n"
     ]
    }
   ],
   "source": [
    "ss = 1\n",
    "for x in p.index:\n",
    "    p[x] = sorted(p[x], key=get_prob, reverse=True)\n",
    "    p[x] = \" \".join([str(i) for i in p[x]])\n",
    "    ss += 1\n",
    "    if ss%10000 == 0:\n",
    "        print (ss + 0.0)/len(p)\n",
    "\n",
    "subm = pd.DataFrame(p.index)\n",
    "subm[\"ad_id\"] = p.values"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}