Click Through Rate Notebook.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2e4642d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import random\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, StackingClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from xgboost import XGBClassifier\n",
    "\n",
    "\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import precision_score, recall_score, f1_score\n",
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "\n",
    "num_lines = 40428968\n",
    "sample_size = 100000\n",
    "skip = sorted(random.sample(range(1, num_lines), num_lines-sample_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f1cd3dab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 99999 entries, 0 to 99998\n",
      "Data columns (total 24 columns):\n",
      " #   Column            Non-Null Count  Dtype \n",
      "---  ------            --------------  ----- \n",
      " 0   id                99999 non-null  uint64\n",
      " 1   click             99999 non-null  int64 \n",
      " 2   hour              99999 non-null  int64 \n",
      " 3   C1                99999 non-null  int64 \n",
      " 4   banner_pos        99999 non-null  int64 \n",
      " 5   site_id           99999 non-null  object\n",
      " 6   site_domain       99999 non-null  object\n",
      " 7   site_category     99999 non-null  object\n",
      " 8   app_id            99999 non-null  object\n",
      " 9   app_domain        99999 non-null  object\n",
      " 10  app_category      99999 non-null  object\n",
      " 11  device_id         99999 non-null  object\n",
      " 12  device_ip         99999 non-null  object\n",
      " 13  device_model      99999 non-null  object\n",
      " 14  device_type       99999 non-null  int64 \n",
      " 15  device_conn_type  99999 non-null  int64 \n",
      " 16  C14               99999 non-null  int64 \n",
      " 17  C15               99999 non-null  int64 \n",
      " 18  C16               99999 non-null  int64 \n",
      " 19  C17               99999 non-null  int64 \n",
      " 20  C18               99999 non-null  int64 \n",
      " 21  C19               99999 non-null  int64 \n",
      " 22  C20               99999 non-null  int64 \n",
      " 23  C21               99999 non-null  int64 \n",
      "dtypes: int64(14), object(9), uint64(1)\n",
      "memory usage: 18.3+ MB\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.read_csv('train.gz', header=0, skiprows=skip)\n",
    "\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "0c2438be",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>click</th>\n",
       "      <th>hour</th>\n",
       "      <th>C1</th>\n",
       "      <th>banner_pos</th>\n",
       "      <th>site_id</th>\n",
       "      <th>site_domain</th>\n",
       "      <th>site_category</th>\n",
       "      <th>app_id</th>\n",
       "      <th>app_domain</th>\n",
       "      <th>...</th>\n",
       "      <th>device_type</th>\n",
       "      <th>device_conn_type</th>\n",
       "      <th>C14</th>\n",
       "      <th>C15</th>\n",
       "      <th>C16</th>\n",
       "      <th>C17</th>\n",
       "      <th>C18</th>\n",
       "      <th>C19</th>\n",
       "      <th>C20</th>\n",
       "      <th>C21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10000679056417042096</td>\n",
       "      <td>0</td>\n",
       "      <td>14102100</td>\n",
       "      <td>1005</td>\n",
       "      <td>1</td>\n",
       "      <td>fe8cc448</td>\n",
       "      <td>9166c161</td>\n",
       "      <td>0569f928</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>18993</td>\n",
       "      <td>320</td>\n",
       "      <td>50</td>\n",
       "      <td>2161</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>-1</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1006958220126877439</td>\n",
       "      <td>0</td>\n",
       "      <td>14102100</td>\n",
       "      <td>1005</td>\n",
       "      <td>0</td>\n",
       "      <td>1fbe01fe</td>\n",
       "      <td>f3845767</td>\n",
       "      <td>28905ebd</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>15708</td>\n",
       "      <td>320</td>\n",
       "      <td>50</td>\n",
       "      <td>1722</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>-1</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10091458153276608983</td>\n",
       "      <td>0</td>\n",
       "      <td>14102100</td>\n",
       "      <td>1005</td>\n",
       "      <td>0</td>\n",
       "      <td>1fbe01fe</td>\n",
       "      <td>f3845767</td>\n",
       "      <td>28905ebd</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>15708</td>\n",
       "      <td>320</td>\n",
       "      <td>50</td>\n",
       "      <td>1722</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>-1</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>103250964972173685</td>\n",
       "      <td>0</td>\n",
       "      <td>14102100</td>\n",
       "      <td>1005</td>\n",
       "      <td>0</td>\n",
       "      <td>6c5b482c</td>\n",
       "      <td>7687a86e</td>\n",
       "      <td>3e814130</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>17653</td>\n",
       "      <td>300</td>\n",
       "      <td>250</td>\n",
       "      <td>1994</td>\n",
       "      <td>2</td>\n",
       "      <td>39</td>\n",
       "      <td>-1</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10447359675376765576</td>\n",
       "      <td>0</td>\n",
       "      <td>14102100</td>\n",
       "      <td>1005</td>\n",
       "      <td>0</td>\n",
       "      <td>543a539e</td>\n",
       "      <td>c7ca3108</td>\n",
       "      <td>3e814130</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20362</td>\n",
       "      <td>320</td>\n",
       "      <td>50</td>\n",
       "      <td>2333</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>-1</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     id  click      hour    C1  banner_pos   site_id  \\\n",
       "0  10000679056417042096      0  14102100  1005           1  fe8cc448   \n",
       "1   1006958220126877439      0  14102100  1005           0  1fbe01fe   \n",
       "2  10091458153276608983      0  14102100  1005           0  1fbe01fe   \n",
       "3    103250964972173685      0  14102100  1005           0  6c5b482c   \n",
       "4  10447359675376765576      0  14102100  1005           0  543a539e   \n",
       "\n",
       "  site_domain site_category    app_id app_domain  ... device_type  \\\n",
       "0    9166c161      0569f928  ecad2386   7801e8d9  ...           1   \n",
       "1    f3845767      28905ebd  ecad2386   7801e8d9  ...           1   \n",
       "2    f3845767      28905ebd  ecad2386   7801e8d9  ...           1   \n",
       "3    7687a86e      3e814130  ecad2386   7801e8d9  ...           1   \n",
       "4    c7ca3108      3e814130  ecad2386   7801e8d9  ...           1   \n",
       "\n",
       "  device_conn_type    C14  C15  C16   C17  C18  C19  C20  C21  \n",
       "0                0  18993  320   50  2161    0   35   -1  157  \n",
       "1                0  15708  320   50  1722    0   35   -1   79  \n",
       "2                0  15708  320   50  1722    0   35   -1   79  \n",
       "3                0  17653  300  250  1994    2   39   -1   33  \n",
       "4                0  20362  320   50  2333    0   39   -1  157  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "8f3a74f5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14102209    1088\n",
       "14102813    1065\n",
       "14102210    1059\n",
       "14102814     984\n",
       "14102212     956\n",
       "            ... \n",
       "14102502      90\n",
       "14102420      70\n",
       "14102423      62\n",
       "14102421      47\n",
       "14102422      38\n",
       "Name: hour, Length: 240, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['hour'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "eccc7685",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>unique</th>\n",
       "      <th>top</th>\n",
       "      <th>freq</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9225507433603274752.0</td>\n",
       "      <td>5317900022276816896.0</td>\n",
       "      <td>82848986010454.0</td>\n",
       "      <td>4627864442575608832.0</td>\n",
       "      <td>9215817656371994624.0</td>\n",
       "      <td>13837361437098051584.0</td>\n",
       "      <td>18446653833922121728.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>click</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.167622</td>\n",
       "      <td>0.373532</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hour</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14102558.587186</td>\n",
       "      <td>296.052779</td>\n",
       "      <td>14102100.0</td>\n",
       "      <td>14102304.0</td>\n",
       "      <td>14102602.0</td>\n",
       "      <td>14102814.0</td>\n",
       "      <td>14103023.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C1</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1004.96104</td>\n",
       "      <td>1.08709</td>\n",
       "      <td>1001.0</td>\n",
       "      <td>1005.0</td>\n",
       "      <td>1005.0</td>\n",
       "      <td>1005.0</td>\n",
       "      <td>1012.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>banner_pos</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.289993</td>\n",
       "      <td>0.513069</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>site_id</th>\n",
       "      <td>99999</td>\n",
       "      <td>1458</td>\n",
       "      <td>85f751fd</td>\n",
       "      <td>35894</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>site_domain</th>\n",
       "      <td>99999</td>\n",
       "      <td>1352</td>\n",
       "      <td>c4e18dd6</td>\n",
       "      <td>37295</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>site_category</th>\n",
       "      <td>99999</td>\n",
       "      <td>19</td>\n",
       "      <td>50e219e0</td>\n",
       "      <td>40787</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app_id</th>\n",
       "      <td>99999</td>\n",
       "      <td>1316</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>64105</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app_domain</th>\n",
       "      <td>99999</td>\n",
       "      <td>99</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>67440</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app_category</th>\n",
       "      <td>99999</td>\n",
       "      <td>22</td>\n",
       "      <td>07d7df22</td>\n",
       "      <td>64909</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_id</th>\n",
       "      <td>99999</td>\n",
       "      <td>16737</td>\n",
       "      <td>a99f214a</td>\n",
       "      <td>82637</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_ip</th>\n",
       "      <td>99999</td>\n",
       "      <td>78029</td>\n",
       "      <td>6b9769f2</td>\n",
       "      <td>509</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_model</th>\n",
       "      <td>99999</td>\n",
       "      <td>3168</td>\n",
       "      <td>8a4875bd</td>\n",
       "      <td>6104</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_type</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.01178</td>\n",
       "      <td>0.520006</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_conn_type</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.329093</td>\n",
       "      <td>0.85418</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C14</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18842.314193</td>\n",
       "      <td>4954.643941</td>\n",
       "      <td>375.0</td>\n",
       "      <td>16920.0</td>\n",
       "      <td>20346.0</td>\n",
       "      <td>21893.0</td>\n",
       "      <td>24043.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C15</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>318.858989</td>\n",
       "      <td>21.259143</td>\n",
       "      <td>120.0</td>\n",
       "      <td>320.0</td>\n",
       "      <td>320.0</td>\n",
       "      <td>320.0</td>\n",
       "      <td>1024.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C16</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60.187922</td>\n",
       "      <td>47.532918</td>\n",
       "      <td>20.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>1024.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C17</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2113.374504</td>\n",
       "      <td>608.565903</td>\n",
       "      <td>112.0</td>\n",
       "      <td>1863.0</td>\n",
       "      <td>2323.0</td>\n",
       "      <td>2526.0</td>\n",
       "      <td>2757.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C18</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.438634</td>\n",
       "      <td>1.325951</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C19</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>227.098661</td>\n",
       "      <td>350.808479</td>\n",
       "      <td>33.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>39.0</td>\n",
       "      <td>171.0</td>\n",
       "      <td>1839.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C20</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>53305.816328</td>\n",
       "      <td>49951.241328</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>100049.0</td>\n",
       "      <td>100090.0</td>\n",
       "      <td>100248.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C21</th>\n",
       "      <td>99999.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>82.97601</td>\n",
       "      <td>70.09222</td>\n",
       "      <td>1.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>255.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    count unique       top   freq                   mean  \\\n",
       "id                99999.0    NaN       NaN    NaN  9225507433603274752.0   \n",
       "click             99999.0    NaN       NaN    NaN               0.167622   \n",
       "hour              99999.0    NaN       NaN    NaN        14102558.587186   \n",
       "C1                99999.0    NaN       NaN    NaN             1004.96104   \n",
       "banner_pos        99999.0    NaN       NaN    NaN               0.289993   \n",
       "site_id             99999   1458  85f751fd  35894                    NaN   \n",
       "site_domain         99999   1352  c4e18dd6  37295                    NaN   \n",
       "site_category       99999     19  50e219e0  40787                    NaN   \n",
       "app_id              99999   1316  ecad2386  64105                    NaN   \n",
       "app_domain          99999     99  7801e8d9  67440                    NaN   \n",
       "app_category        99999     22  07d7df22  64909                    NaN   \n",
       "device_id           99999  16737  a99f214a  82637                    NaN   \n",
       "device_ip           99999  78029  6b9769f2    509                    NaN   \n",
       "device_model        99999   3168  8a4875bd   6104                    NaN   \n",
       "device_type       99999.0    NaN       NaN    NaN                1.01178   \n",
       "device_conn_type  99999.0    NaN       NaN    NaN               0.329093   \n",
       "C14               99999.0    NaN       NaN    NaN           18842.314193   \n",
       "C15               99999.0    NaN       NaN    NaN             318.858989   \n",
       "C16               99999.0    NaN       NaN    NaN              60.187922   \n",
       "C17               99999.0    NaN       NaN    NaN            2113.374504   \n",
       "C18               99999.0    NaN       NaN    NaN               1.438634   \n",
       "C19               99999.0    NaN       NaN    NaN             227.098661   \n",
       "C20               99999.0    NaN       NaN    NaN           53305.816328   \n",
       "C21               99999.0    NaN       NaN    NaN               82.97601   \n",
       "\n",
       "                                    std               min  \\\n",
       "id                5317900022276816896.0  82848986010454.0   \n",
       "click                          0.373532               0.0   \n",
       "hour                         296.052779        14102100.0   \n",
       "C1                              1.08709            1001.0   \n",
       "banner_pos                     0.513069               0.0   \n",
       "site_id                             NaN               NaN   \n",
       "site_domain                         NaN               NaN   \n",
       "site_category                       NaN               NaN   \n",
       "app_id                              NaN               NaN   \n",
       "app_domain                          NaN               NaN   \n",
       "app_category                        NaN               NaN   \n",
       "device_id                           NaN               NaN   \n",
       "device_ip                           NaN               NaN   \n",
       "device_model                        NaN               NaN   \n",
       "device_type                    0.520006               0.0   \n",
       "device_conn_type                0.85418               0.0   \n",
       "C14                         4954.643941             375.0   \n",
       "C15                           21.259143             120.0   \n",
       "C16                           47.532918              20.0   \n",
       "C17                          608.565903             112.0   \n",
       "C18                            1.325951               0.0   \n",
       "C19                          350.808479              33.0   \n",
       "C20                        49951.241328              -1.0   \n",
       "C21                            70.09222               1.0   \n",
       "\n",
       "                                    25%                    50%  \\\n",
       "id                4627864442575608832.0  9215817656371994624.0   \n",
       "click                               0.0                    0.0   \n",
       "hour                         14102304.0             14102602.0   \n",
       "C1                               1005.0                 1005.0   \n",
       "banner_pos                          0.0                    0.0   \n",
       "site_id                             NaN                    NaN   \n",
       "site_domain                         NaN                    NaN   \n",
       "site_category                       NaN                    NaN   \n",
       "app_id                              NaN                    NaN   \n",
       "app_domain                          NaN                    NaN   \n",
       "app_category                        NaN                    NaN   \n",
       "device_id                           NaN                    NaN   \n",
       "device_ip                           NaN                    NaN   \n",
       "device_model                        NaN                    NaN   \n",
       "device_type                         1.0                    1.0   \n",
       "device_conn_type                    0.0                    0.0   \n",
       "C14                             16920.0                20346.0   \n",
       "C15                               320.0                  320.0   \n",
       "C16                                50.0                   50.0   \n",
       "C17                              1863.0                 2323.0   \n",
       "C18                                 0.0                    2.0   \n",
       "C19                                35.0                   39.0   \n",
       "C20                                -1.0               100049.0   \n",
       "C21                                23.0                   61.0   \n",
       "\n",
       "                                     75%                     max  \n",
       "id                13837361437098051584.0  18446653833922121728.0  \n",
       "click                                0.0                     1.0  \n",
       "hour                          14102814.0              14103023.0  \n",
       "C1                                1005.0                  1012.0  \n",
       "banner_pos                           1.0                     7.0  \n",
       "site_id                              NaN                     NaN  \n",
       "site_domain                          NaN                     NaN  \n",
       "site_category                        NaN                     NaN  \n",
       "app_id                               NaN                     NaN  \n",
       "app_domain                           NaN                     NaN  \n",
       "app_category                         NaN                     NaN  \n",
       "device_id                            NaN                     NaN  \n",
       "device_ip                            NaN                     NaN  \n",
       "device_model                         NaN                     NaN  \n",
       "device_type                          1.0                     5.0  \n",
       "device_conn_type                     0.0                     5.0  \n",
       "C14                              21893.0                 24043.0  \n",
       "C15                                320.0                  1024.0  \n",
       "C16                                 50.0                  1024.0  \n",
       "C17                               2526.0                  2757.0  \n",
       "C18                                  3.0                     3.0  \n",
       "C19                                171.0                  1839.0  \n",
       "C20                             100090.0                100248.0  \n",
       "C21                                101.0                   255.0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe(include='all').T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e8587d50",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <td>10000679056417042096</td>\n",
       "      <td>1006958220126877439</td>\n",
       "      <td>10091458153276608983</td>\n",
       "      <td>103250964972173685</td>\n",
       "      <td>10447359675376765576</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>click</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hour</th>\n",
       "      <td>14102100</td>\n",
       "      <td>14102100</td>\n",
       "      <td>14102100</td>\n",
       "      <td>14102100</td>\n",
       "      <td>14102100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C1</th>\n",
       "      <td>1005</td>\n",
       "      <td>1005</td>\n",
       "      <td>1005</td>\n",
       "      <td>1005</td>\n",
       "      <td>1005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>banner_pos</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>site_id</th>\n",
       "      <td>fe8cc448</td>\n",
       "      <td>1fbe01fe</td>\n",
       "      <td>1fbe01fe</td>\n",
       "      <td>6c5b482c</td>\n",
       "      <td>543a539e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>site_domain</th>\n",
       "      <td>9166c161</td>\n",
       "      <td>f3845767</td>\n",
       "      <td>f3845767</td>\n",
       "      <td>7687a86e</td>\n",
       "      <td>c7ca3108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>site_category</th>\n",
       "      <td>0569f928</td>\n",
       "      <td>28905ebd</td>\n",
       "      <td>28905ebd</td>\n",
       "      <td>3e814130</td>\n",
       "      <td>3e814130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app_id</th>\n",
       "      <td>ecad2386</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>ecad2386</td>\n",
       "      <td>ecad2386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app_domain</th>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>7801e8d9</td>\n",
       "      <td>7801e8d9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app_category</th>\n",
       "      <td>07d7df22</td>\n",
       "      <td>07d7df22</td>\n",
       "      <td>07d7df22</td>\n",
       "      <td>07d7df22</td>\n",
       "      <td>07d7df22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_id</th>\n",
       "      <td>a99f214a</td>\n",
       "      <td>a99f214a</td>\n",
       "      <td>a99f214a</td>\n",
       "      <td>a99f214a</td>\n",
       "      <td>a99f214a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_ip</th>\n",
       "      <td>9644d0bf</td>\n",
       "      <td>6b9769f2</td>\n",
       "      <td>c17d35ae</td>\n",
       "      <td>b271bd3c</td>\n",
       "      <td>57bce7c6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_model</th>\n",
       "      <td>779d90c2</td>\n",
       "      <td>f39b265e</td>\n",
       "      <td>36d749e5</td>\n",
       "      <td>8571106b</td>\n",
       "      <td>d787e91b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_type</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_conn_type</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C14</th>\n",
       "      <td>18993</td>\n",
       "      <td>15708</td>\n",
       "      <td>15708</td>\n",
       "      <td>17653</td>\n",
       "      <td>20362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C15</th>\n",
       "      <td>320</td>\n",
       "      <td>320</td>\n",
       "      <td>320</td>\n",
       "      <td>300</td>\n",
       "      <td>320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C16</th>\n",
       "      <td>50</td>\n",
       "      <td>50</td>\n",
       "      <td>50</td>\n",
       "      <td>250</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C17</th>\n",
       "      <td>2161</td>\n",
       "      <td>1722</td>\n",
       "      <td>1722</td>\n",
       "      <td>1994</td>\n",
       "      <td>2333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C18</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C19</th>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>39</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C20</th>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C21</th>\n",
       "      <td>157</td>\n",
       "      <td>79</td>\n",
       "      <td>79</td>\n",
       "      <td>33</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     0                    1  \\\n",
       "id                10000679056417042096  1006958220126877439   \n",
       "click                                0                    0   \n",
       "hour                          14102100             14102100   \n",
       "C1                                1005                 1005   \n",
       "banner_pos                           1                    0   \n",
       "site_id                       fe8cc448             1fbe01fe   \n",
       "site_domain                   9166c161             f3845767   \n",
       "site_category                 0569f928             28905ebd   \n",
       "app_id                        ecad2386             ecad2386   \n",
       "app_domain                    7801e8d9             7801e8d9   \n",
       "app_category                  07d7df22             07d7df22   \n",
       "device_id                     a99f214a             a99f214a   \n",
       "device_ip                     9644d0bf             6b9769f2   \n",
       "device_model                  779d90c2             f39b265e   \n",
       "device_type                          1                    1   \n",
       "device_conn_type                     0                    0   \n",
       "C14                              18993                15708   \n",
       "C15                                320                  320   \n",
       "C16                                 50                   50   \n",
       "C17                               2161                 1722   \n",
       "C18                                  0                    0   \n",
       "C19                                 35                   35   \n",
       "C20                                 -1                   -1   \n",
       "C21                                157                   79   \n",
       "\n",
       "                                     2                   3  \\\n",
       "id                10091458153276608983  103250964972173685   \n",
       "click                                0                   0   \n",
       "hour                          14102100            14102100   \n",
       "C1                                1005                1005   \n",
       "banner_pos                           0                   0   \n",
       "site_id                       1fbe01fe            6c5b482c   \n",
       "site_domain                   f3845767            7687a86e   \n",
       "site_category                 28905ebd            3e814130   \n",
       "app_id                        ecad2386            ecad2386   \n",
       "app_domain                    7801e8d9            7801e8d9   \n",
       "app_category                  07d7df22            07d7df22   \n",
       "device_id                     a99f214a            a99f214a   \n",
       "device_ip                     c17d35ae            b271bd3c   \n",
       "device_model                  36d749e5            8571106b   \n",
       "device_type                          1                   1   \n",
       "device_conn_type                     0                   0   \n",
       "C14                              15708               17653   \n",
       "C15                                320                 300   \n",
       "C16                                 50                 250   \n",
       "C17                               1722                1994   \n",
       "C18                                  0                   2   \n",
       "C19                                 35                  39   \n",
       "C20                                 -1                  -1   \n",
       "C21                                 79                  33   \n",
       "\n",
       "                                     4  \n",
       "id                10447359675376765576  \n",
       "click                                0  \n",
       "hour                          14102100  \n",
       "C1                                1005  \n",
       "banner_pos                           0  \n",
       "site_id                       543a539e  \n",
       "site_domain                   c7ca3108  \n",
       "site_category                 3e814130  \n",
       "app_id                        ecad2386  \n",
       "app_domain                    7801e8d9  \n",
       "app_category                  07d7df22  \n",
       "device_id                     a99f214a  \n",
       "device_ip                     57bce7c6  \n",
       "device_model                  d787e91b  \n",
       "device_type                          1  \n",
       "device_conn_type                     0  \n",
       "C14                              20362  \n",
       "C15                                320  \n",
       "C16                                 50  \n",
       "C17                               2333  \n",
       "C18                                  0  \n",
       "C19                                 39  \n",
       "C20                                 -1  \n",
       "C21                                157  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head().T"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb0fcf9e",
   "metadata": {},
   "source": [
    "#### Dropping text columns, target variable, and seemingly useless features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b2d3b6e3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(99999, 13)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df.select_dtypes('number').drop(['id','click'],axis=1) #,'device_type'\n",
    "y = df['click']\n",
    "\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3e6543a4",
   "metadata": {},
   "source": [
    "#### Normalizing dataset using Z-Score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "9405dfb5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hour</th>\n",
       "      <th>C1</th>\n",
       "      <th>banner_pos</th>\n",
       "      <th>device_type</th>\n",
       "      <th>device_conn_type</th>\n",
       "      <th>C14</th>\n",
       "      <th>C15</th>\n",
       "      <th>C16</th>\n",
       "      <th>C17</th>\n",
       "      <th>C18</th>\n",
       "      <th>C19</th>\n",
       "      <th>C20</th>\n",
       "      <th>C21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.549005</td>\n",
       "      <td>0.035839</td>\n",
       "      <td>1.383842</td>\n",
       "      <td>-0.022654</td>\n",
       "      <td>-0.385274</td>\n",
       "      <td>0.030413</td>\n",
       "      <td>0.053672</td>\n",
       "      <td>-0.214334</td>\n",
       "      <td>0.078259</td>\n",
       "      <td>-1.084983</td>\n",
       "      <td>-0.547588</td>\n",
       "      <td>-1.067177</td>\n",
       "      <td>1.056094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.549005</td>\n",
       "      <td>0.035839</td>\n",
       "      <td>-0.565212</td>\n",
       "      <td>-0.022654</td>\n",
       "      <td>-0.385274</td>\n",
       "      <td>-0.632601</td>\n",
       "      <td>0.053672</td>\n",
       "      <td>-0.214334</td>\n",
       "      <td>-0.643109</td>\n",
       "      <td>-1.084983</td>\n",
       "      <td>-0.547588</td>\n",
       "      <td>-1.067177</td>\n",
       "      <td>-0.056725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1.549005</td>\n",
       "      <td>0.035839</td>\n",
       "      <td>-0.565212</td>\n",
       "      <td>-0.022654</td>\n",
       "      <td>-0.385274</td>\n",
       "      <td>-0.632601</td>\n",
       "      <td>0.053672</td>\n",
       "      <td>-0.214334</td>\n",
       "      <td>-0.643109</td>\n",
       "      <td>-1.084983</td>\n",
       "      <td>-0.547588</td>\n",
       "      <td>-1.067177</td>\n",
       "      <td>-0.056725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1.549005</td>\n",
       "      <td>0.035839</td>\n",
       "      <td>-0.565212</td>\n",
       "      <td>-0.022654</td>\n",
       "      <td>-0.385274</td>\n",
       "      <td>-0.240040</td>\n",
       "      <td>-0.887100</td>\n",
       "      <td>3.993276</td>\n",
       "      <td>-0.196157</td>\n",
       "      <td>0.423368</td>\n",
       "      <td>-0.536186</td>\n",
       "      <td>-1.067177</td>\n",
       "      <td>-0.713004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.549005</td>\n",
       "      <td>0.035839</td>\n",
       "      <td>-0.565212</td>\n",
       "      <td>-0.022654</td>\n",
       "      <td>-0.385274</td>\n",
       "      <td>0.306719</td>\n",
       "      <td>0.053672</td>\n",
       "      <td>-0.214334</td>\n",
       "      <td>0.360890</td>\n",
       "      <td>-1.084983</td>\n",
       "      <td>-0.536186</td>\n",
       "      <td>-1.067177</td>\n",
       "      <td>1.056094</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       hour        C1  banner_pos  device_type  device_conn_type       C14  \\\n",
       "0 -1.549005  0.035839    1.383842    -0.022654         -0.385274  0.030413   \n",
       "1 -1.549005  0.035839   -0.565212    -0.022654         -0.385274 -0.632601   \n",
       "2 -1.549005  0.035839   -0.565212    -0.022654         -0.385274 -0.632601   \n",
       "3 -1.549005  0.035839   -0.565212    -0.022654         -0.385274 -0.240040   \n",
       "4 -1.549005  0.035839   -0.565212    -0.022654         -0.385274  0.306719   \n",
       "\n",
       "        C15       C16       C17       C18       C19       C20       C21  \n",
       "0  0.053672 -0.214334  0.078259 -1.084983 -0.547588 -1.067177  1.056094  \n",
       "1  0.053672 -0.214334 -0.643109 -1.084983 -0.547588 -1.067177 -0.056725  \n",
       "2  0.053672 -0.214334 -0.643109 -1.084983 -0.547588 -1.067177 -0.056725  \n",
       "3 -0.887100  3.993276 -0.196157  0.423368 -0.536186 -1.067177 -0.713004  \n",
       "4  0.053672 -0.214334  0.360890 -1.084983 -0.536186 -1.067177  1.056094  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = (X - X.mean()) / X.std()\n",
    "\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67810db9",
   "metadata": {},
   "source": [
    "#### Evaluating imbalance of dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "5018bb8a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.832378\n",
       "1    0.167622\n",
       "Name: click, dtype: float64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "605c953a",
   "metadata": {},
   "source": [
    "#### Instantiating, Training, and calling feature_importances_ attribute on RandomForest model to evaluate our features' predictive power"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ba83b819",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 25.2 s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Importances</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>hour</th>\n",
       "      <td>0.576899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C14</th>\n",
       "      <td>0.129982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C20</th>\n",
       "      <td>0.093384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C17</th>\n",
       "      <td>0.039228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C21</th>\n",
       "      <td>0.033684</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C19</th>\n",
       "      <td>0.024802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_conn_type</th>\n",
       "      <td>0.024324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C16</th>\n",
       "      <td>0.019182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C18</th>\n",
       "      <td>0.018030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>banner_pos</th>\n",
       "      <td>0.016442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C15</th>\n",
       "      <td>0.009877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C1</th>\n",
       "      <td>0.007405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device_type</th>\n",
       "      <td>0.006762</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Feature Importances\n",
       "hour                         0.576899\n",
       "C14                          0.129982\n",
       "C20                          0.093384\n",
       "C17                          0.039228\n",
       "C21                          0.033684\n",
       "C19                          0.024802\n",
       "device_conn_type             0.024324\n",
       "C16                          0.019182\n",
       "C18                          0.018030\n",
       "banner_pos                   0.016442\n",
       "C15                          0.009877\n",
       "C1                           0.007405\n",
       "device_type                  0.006762"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "rf = RandomForestClassifier(class_weight='balanced')\n",
    "\n",
    "rf_feature_importances = pd.DataFrame(rf.fit(X,y).feature_importances_,index=[i for i in X.columns],columns=['Feature Importances']).sort_values(by='Feature Importances', ascending=False)\n",
    "rf_feature_importances[:15]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "07147769",
   "metadata": {},
   "source": [
    "#### Reducing feature dimension using RandomForest Feature Importance Attribute\n",
    "* Decided to omit this step because simply dropping text columns reduced feature dimension significantly enough."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "e20214f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# features = rf_feature_importances[rf_feature_importances['Feature Importances'] > 0.01].index\n",
    "# print(features)\n",
    "# X = X[features]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1441d8e",
   "metadata": {},
   "source": [
    "#### Training and Score Plotting various models on dataset \n",
    "* Default/Minimal parameter tuning."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "fc89a0ec",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 9min 34s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1eb00bfb610>,\n",
       " <matplotlib.lines.Line2D at 0x1eb00bfb6a0>,\n",
       " <matplotlib.lines.Line2D at 0x1eb00bfb490>]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# Don't forget a StackClassifier using top models\n",
    "\n",
    "log = LogisticRegression()\n",
    "rf = RandomForestClassifier(class_weight='balanced')\n",
    "knn = KNeighborsClassifier()\n",
    "ada = AdaBoostClassifier()\n",
    "xgb = XGBClassifier(use_label_encoder=False,verbosity=0)\n",
    "\n",
    "models = [rf, log, knn, ada, xgb]\n",
    "metrics = ['f1', 'precision', 'recall']\n",
    "\n",
    "# Function to score an array of machine learning algorithms on the dataset\n",
    "def score_models(X, y, models, metrics, folds): \n",
    "    \n",
    "    # plots scorings\n",
    "    # I want this function to ALSO return confusion matrix object, roc/auc plot object,\n",
    "    # precision/recall plot object, [tp,fp,tn,fn] or rates, \n",
    "    # recall, precision, f1, Matthews correlation coeficient\n",
    "    \n",
    "    p = np.random.permutation(len(X))\n",
    "    X, y = X.iloc[p], y.iloc[p]\n",
    "    \n",
    "    model_scores = {}\n",
    "    \n",
    "    for model in models: #replace with cross_val_score for metric in metrics\n",
    "        \n",
    "        scores = []\n",
    "        \n",
    "        for metric in metrics:\n",
    "            score = cross_val_score(model, X, y, cv = folds, scoring = metric) #returns avg score per metric\n",
    "            # after this runs, save that score before next iteration\n",
    "        \n",
    "            scores.append(score.mean())\n",
    "            # saves score for that metric to dictionary where metric is key, score is value\n",
    "        \n",
    "        model_scores[str(model)] = scores\n",
    "    \n",
    "    plt.figure(figsize=(12,8))\n",
    "    score_plot = plt.plot([x.split('(')[0] for x in model_scores.keys()], model_scores.values(), marker='*')\n",
    "    plt.legend(metrics)\n",
    "    plt.title('Score Plot')\n",
    "    plt.xlabel('Models')\n",
    "    plt.ylabel('Scores')\n",
    "    \n",
    "    return score_plot\n",
    "\n",
    "score_models(X, y, models, metrics, 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11a3eb56",
   "metadata": {},
   "source": [
    "#### GridSearching best model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "09b85131",
   "metadata": {},
   "outputs": [],
   "source": [
    "# GridSearch Here"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b3fd6385",
   "metadata": {},
   "source": [
    "#### Scoreplotting hyperparameter tuned best model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "1a43a8c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scoreplot hyperparameter tuned best model here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "3e3821c2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Done for the day 10/22/2021 \n",
    "\n",
    "# Maybe try to handle more data?\n",
    "# Can you do some feature engineering?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "242b23d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compare other models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4502ff8f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reformat hour column to prevent overfitting\n",
    "# Break hour column into year, month, day,\n",
    "# Break day into morning/afternoon/night etc\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}