Projekt_JZ.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Jakub Zbrzezny, nr indeksu: 286689\n",
    "\n",
    "# Projekt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pip install keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from pprint import pprint\n",
    "\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "import statsmodels.api as sm\n",
    "import statsmodels.formula.api as smf\n",
    "\n",
    "# Import kerasa\n",
    "import keras\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Flatten\n",
    "\n",
    "from tensorflow.keras.datasets import mnist\n",
    "\n",
    "import sklearn.metrics # Importowanie funkcji sklearn.metrics.roc_auc_score.\n",
    "# Ta funkcja liczy nam AUC.\n",
    "\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Embedding, LSTM, GRU, SimpleRNN, Dense\n",
    "\n",
    "from tensorflow import feature_column\n",
    "from tensorflow.keras import layers\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Wczytanie danych"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Mały zbiór treningowy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dane .gz rozpakowałem.\n",
    "D100k = pd.read_csv(\"D100k.tsv\", sep = \"\\t\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>Click</th>\n",
       "      <th>DisplayURL</th>\n",
       "      <th>AdId</th>\n",
       "      <th>AdvertiserId</th>\n",
       "      <th>Depth</th>\n",
       "      <th>Position</th>\n",
       "      <th>UserID</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Age</th>\n",
       "      <th>AdKeyword_tokens</th>\n",
       "      <th>AdTitle_tokens</th>\n",
       "      <th>AdDescription_tokens</th>\n",
       "      <th>Query_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686695</td>\n",
       "      <td>385</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4133|95|17</td>\n",
       "      <td>4133|95|17|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>4133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>13677630321509009335</td>\n",
       "      <td>3517124</td>\n",
       "      <td>23778</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4133</td>\n",
       "      <td>145|65|3927|832|93</td>\n",
       "      <td>3683|4990|2793|11589|21|10741|26|16044|26|3168...</td>\n",
       "      <td>4133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>11689327222955583742</td>\n",
       "      <td>21021375</td>\n",
       "      <td>27701</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4133</td>\n",
       "      <td>4133|95|1|339|125|21|83093</td>\n",
       "      <td>726|50|2218|1533|2275|4133|1299|509|95|2072|1|...</td>\n",
       "      <td>4133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686695</td>\n",
       "      <td>385</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16960371</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>4133|95|17</td>\n",
       "      <td>4133|95|17|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>4133|4942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>15132506310926074459</td>\n",
       "      <td>4424000</td>\n",
       "      <td>20940</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3524325</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>121|4133|95</td>\n",
       "      <td>121|4133|95|8762|3957|4563|2233|192|28|138|3</td>\n",
       "      <td>62|1162|570|8|4133|95|1|81|102|1155|650|1255|1...</td>\n",
       "      <td>4133|4942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686888</td>\n",
       "      <td>385</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>11198634</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>14530|882|167</td>\n",
       "      <td>14530|882|167|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>14530|95|1156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0</td>\n",
       "      <td>13883743923141531481</td>\n",
       "      <td>20568201</td>\n",
       "      <td>29921</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>11198634</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>14530|95|1156</td>\n",
       "      <td>14530|95|1156|0|42|45|14530|95|1156</td>\n",
       "      <td>14530|95|1156|36|14530|4557|6|1|92|28|81|42|45...</td>\n",
       "      <td>14530|95|1156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686578</td>\n",
       "      <td>385</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16905873</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>121|3337|95|6</td>\n",
       "      <td>121|3337|95|6|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>3337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7687047</td>\n",
       "      <td>385</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>5077836</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1970|95|17</td>\n",
       "      <td>1970|95|17|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>1970|4942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686572</td>\n",
       "      <td>385</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3519270</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3337|95|126</td>\n",
       "      <td>3337|95|126|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>304|20|3337|95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>514151146628080502</td>\n",
       "      <td>3691703</td>\n",
       "      <td>22478</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3519270</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3337|95</td>\n",
       "      <td>3337|95|1|380|11435|5523|149</td>\n",
       "      <td>3337|95|1|1007|418|353|1|3337|70|270|89|130|68...</td>\n",
       "      <td>304|20|3337|95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686612</td>\n",
       "      <td>385</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>6247119</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>6047|60</td>\n",
       "      <td>6047|60|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>6047|1098</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0</td>\n",
       "      <td>694915634177245725</td>\n",
       "      <td>9592698</td>\n",
       "      <td>5079</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>6247119</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>211|6047|60</td>\n",
       "      <td>211|6047|60|15|277|60|23|3</td>\n",
       "      <td>40|211|60|1|20|60|580|189|58|6469|4034|14067|1...</td>\n",
       "      <td>6047|1098</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7687144</td>\n",
       "      <td>385</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>4144358</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>121|7894|95|6</td>\n",
       "      <td>121|7894|95|6|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>1467|7894|95|1098|601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>12057878999086460853</td>\n",
       "      <td>20164346</td>\n",
       "      <td>27961</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4144358</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1467|7894</td>\n",
       "      <td>51|0|1467|7894|8|183|94</td>\n",
       "      <td>36|1467|7894|37|51|1|173|32|26|214|154|26|234|...</td>\n",
       "      <td>1467|7894|95|1098|601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0</td>\n",
       "      <td>14347859242938208407</td>\n",
       "      <td>21963776</td>\n",
       "      <td>33931</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4144358</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1467|7894|95</td>\n",
       "      <td>1467|7894|95|1017|5003</td>\n",
       "      <td>1187|734|2|12872|1467|7894|95|1771|1|213|187|1...</td>\n",
       "      <td>1467|7894|95|1098|601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>4860571499428580850</td>\n",
       "      <td>21560664</td>\n",
       "      <td>37484</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>300|17</td>\n",
       "      <td>300|17|15|29|24|4849|1925|6302|350|1591</td>\n",
       "      <td>1925|6302|350|1591|300|17|1371|4493|3123|17458...</td>\n",
       "      <td>300|202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0</td>\n",
       "      <td>10664591050012752564</td>\n",
       "      <td>20620168</td>\n",
       "      <td>30128</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>300|227</td>\n",
       "      <td>1323|1803|300|227|457|8|163|94|15320|7132</td>\n",
       "      <td>300|227|26889|9442|163|0|620|3123|0|300|176|23...</td>\n",
       "      <td>300|202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>4860571499428580850</td>\n",
       "      <td>21560664</td>\n",
       "      <td>37484</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>302819</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>300|17</td>\n",
       "      <td>300|17|15|29|24|4849|1925|6302|350|1591</td>\n",
       "      <td>1925|6302|350|1591|300|17|1371|4493|3123|17458...</td>\n",
       "      <td>7649|75|50|300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0</td>\n",
       "      <td>10664591050012752564</td>\n",
       "      <td>21443408</td>\n",
       "      <td>30128</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>302819</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>300|17</td>\n",
       "      <td>300|17|54|1323|14572|1591|15|457|8|163|94</td>\n",
       "      <td>300|227|4158|1591|1340|1913|768|641|15770|5751...</td>\n",
       "      <td>7649|75|50|300</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Click            DisplayURL      AdId  AdvertiserId  Depth  Position  \\\n",
       "0       0   4298118681424644510   7686695           385      3         3   \n",
       "1       0  13677630321509009335   3517124         23778      3         1   \n",
       "2       0  11689327222955583742  21021375         27701      3         2   \n",
       "3       0   4298118681424644510   7686695           385      1         1   \n",
       "4       0  15132506310926074459   4424000         20940      1         1   \n",
       "5       0   4298118681424644510   7686888           385      2         2   \n",
       "6       0  13883743923141531481  20568201         29921      2         1   \n",
       "7       0   4298118681424644510   7686578           385      1         1   \n",
       "8       0   4298118681424644510   7687047           385      1         1   \n",
       "9       0   4298118681424644510   7686572           385      2         2   \n",
       "10      0    514151146628080502   3691703         22478      2         1   \n",
       "11      0   4298118681424644510   7686612           385      2         2   \n",
       "12      0    694915634177245725   9592698          5079      2         1   \n",
       "13      0   4298118681424644510   7687144           385      3         3   \n",
       "14      0  12057878999086460853  20164346         27961      3         1   \n",
       "15      0  14347859242938208407  21963776         33931      3         2   \n",
       "16      0   4860571499428580850  21560664         37484      2         2   \n",
       "17      0  10664591050012752564  20620168         30128      2         1   \n",
       "18      0   4860571499428580850  21560664         37484      3         3   \n",
       "19      0  10664591050012752564  21443408         30128      3         1   \n",
       "\n",
       "      UserID  Gender  Age AdKeyword_tokens  \\\n",
       "0     490234       1    3       4133|95|17   \n",
       "1     490234       1    3             4133   \n",
       "2     490234       1    3             4133   \n",
       "3   16960371       2    2       4133|95|17   \n",
       "4    3524325       1    3      121|4133|95   \n",
       "5   11198634       2    3    14530|882|167   \n",
       "6   11198634       2    3    14530|95|1156   \n",
       "7   16905873       2    4    121|3337|95|6   \n",
       "8    5077836       2    3       1970|95|17   \n",
       "9    3519270       2    1      3337|95|126   \n",
       "10   3519270       2    1          3337|95   \n",
       "11   6247119       2    6          6047|60   \n",
       "12   6247119       2    6      211|6047|60   \n",
       "13   4144358       2    4    121|7894|95|6   \n",
       "14   4144358       2    4        1467|7894   \n",
       "15   4144358       2    4     1467|7894|95   \n",
       "16    490234       1    3           300|17   \n",
       "17    490234       1    3          300|227   \n",
       "18    302819       0    1           300|17   \n",
       "19    302819       0    1           300|17   \n",
       "\n",
       "                                  AdTitle_tokens  \\\n",
       "0                  4133|95|17|0|4732|95|146|4079   \n",
       "1                             145|65|3927|832|93   \n",
       "2                     4133|95|1|339|125|21|83093   \n",
       "3                  4133|95|17|0|4732|95|146|4079   \n",
       "4   121|4133|95|8762|3957|4563|2233|192|28|138|3   \n",
       "5               14530|882|167|0|4732|95|146|4079   \n",
       "6            14530|95|1156|0|42|45|14530|95|1156   \n",
       "7               121|3337|95|6|0|4732|95|146|4079   \n",
       "8                  1970|95|17|0|4732|95|146|4079   \n",
       "9                 3337|95|126|0|4732|95|146|4079   \n",
       "10                  3337|95|1|380|11435|5523|149   \n",
       "11                    6047|60|0|4732|95|146|4079   \n",
       "12                    211|6047|60|15|277|60|23|3   \n",
       "13              121|7894|95|6|0|4732|95|146|4079   \n",
       "14                       51|0|1467|7894|8|183|94   \n",
       "15                        1467|7894|95|1017|5003   \n",
       "16       300|17|15|29|24|4849|1925|6302|350|1591   \n",
       "17     1323|1803|300|227|457|8|163|94|15320|7132   \n",
       "18       300|17|15|29|24|4849|1925|6302|350|1591   \n",
       "19     300|17|54|1323|14572|1591|15|457|8|163|94   \n",
       "\n",
       "                                 AdDescription_tokens           Query_tokens  \n",
       "0   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...                   4133  \n",
       "1   3683|4990|2793|11589|21|10741|26|16044|26|3168...                   4133  \n",
       "2   726|50|2218|1533|2275|4133|1299|509|95|2072|1|...                   4133  \n",
       "3   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...              4133|4942  \n",
       "4   62|1162|570|8|4133|95|1|81|102|1155|650|1255|1...              4133|4942  \n",
       "5   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...          14530|95|1156  \n",
       "6   14530|95|1156|36|14530|4557|6|1|92|28|81|42|45...          14530|95|1156  \n",
       "7   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...                   3337  \n",
       "8   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...              1970|4942  \n",
       "9   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...         304|20|3337|95  \n",
       "10  3337|95|1|1007|418|353|1|3337|70|270|89|130|68...         304|20|3337|95  \n",
       "11  8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...              6047|1098  \n",
       "12  40|211|60|1|20|60|580|189|58|6469|4034|14067|1...              6047|1098  \n",
       "13  8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...  1467|7894|95|1098|601  \n",
       "14  36|1467|7894|37|51|1|173|32|26|214|154|26|234|...  1467|7894|95|1098|601  \n",
       "15  1187|734|2|12872|1467|7894|95|1771|1|213|187|1...  1467|7894|95|1098|601  \n",
       "16  1925|6302|350|1591|300|17|1371|4493|3123|17458...                300|202  \n",
       "17  300|227|26889|9442|163|0|620|3123|0|300|176|23...                300|202  \n",
       "18  1925|6302|350|1591|300|17|1371|4493|3123|17458...         7649|75|50|300  \n",
       "19  300|227|4158|1591|1340|1913|768|641|15770|5751...         7649|75|50|300  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "D100k.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Średni zbiór treningowy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "D10M = pd.read_csv(\"D10M.tsv\", sep = \"\\t\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>Click</th>\n",
       "      <th>DisplayURL</th>\n",
       "      <th>AdId</th>\n",
       "      <th>AdvertiserId</th>\n",
       "      <th>Depth</th>\n",
       "      <th>Position</th>\n",
       "      <th>UserID</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Age</th>\n",
       "      <th>AdKeyword_tokens</th>\n",
       "      <th>AdTitle_tokens</th>\n",
       "      <th>AdDescription_tokens</th>\n",
       "      <th>Query_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686695</td>\n",
       "      <td>385</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4133|95|17</td>\n",
       "      <td>4133|95|17|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>4133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>13677630321509009335</td>\n",
       "      <td>3517124</td>\n",
       "      <td>23778</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4133</td>\n",
       "      <td>145|65|3927|832|93</td>\n",
       "      <td>3683|4990|2793|11589|21|10741|26|16044|26|3168...</td>\n",
       "      <td>4133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>11689327222955583742</td>\n",
       "      <td>21021375</td>\n",
       "      <td>27701</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4133</td>\n",
       "      <td>4133|95|1|339|125|21|83093</td>\n",
       "      <td>726|50|2218|1533|2275|4133|1299|509|95|2072|1|...</td>\n",
       "      <td>4133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686695</td>\n",
       "      <td>385</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16960371</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>4133|95|17</td>\n",
       "      <td>4133|95|17|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>4133|4942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>15132506310926074459</td>\n",
       "      <td>4424000</td>\n",
       "      <td>20940</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3524325</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>121|4133|95</td>\n",
       "      <td>121|4133|95|8762|3957|4563|2233|192|28|138|3</td>\n",
       "      <td>62|1162|570|8|4133|95|1|81|102|1155|650|1255|1...</td>\n",
       "      <td>4133|4942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686888</td>\n",
       "      <td>385</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>11198634</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>14530|882|167</td>\n",
       "      <td>14530|882|167|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>14530|95|1156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0</td>\n",
       "      <td>13883743923141531481</td>\n",
       "      <td>20568201</td>\n",
       "      <td>29921</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>11198634</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>14530|95|1156</td>\n",
       "      <td>14530|95|1156|0|42|45|14530|95|1156</td>\n",
       "      <td>14530|95|1156|36|14530|4557|6|1|92|28|81|42|45...</td>\n",
       "      <td>14530|95|1156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686578</td>\n",
       "      <td>385</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16905873</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>121|3337|95|6</td>\n",
       "      <td>121|3337|95|6|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>3337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7687047</td>\n",
       "      <td>385</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>5077836</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1970|95|17</td>\n",
       "      <td>1970|95|17|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>1970|4942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686572</td>\n",
       "      <td>385</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3519270</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3337|95|126</td>\n",
       "      <td>3337|95|126|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>304|20|3337|95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>514151146628080502</td>\n",
       "      <td>3691703</td>\n",
       "      <td>22478</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3519270</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3337|95</td>\n",
       "      <td>3337|95|1|380|11435|5523|149</td>\n",
       "      <td>3337|95|1|1007|418|353|1|3337|70|270|89|130|68...</td>\n",
       "      <td>304|20|3337|95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7686612</td>\n",
       "      <td>385</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>6247119</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>6047|60</td>\n",
       "      <td>6047|60|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>6047|1098</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0</td>\n",
       "      <td>694915634177245725</td>\n",
       "      <td>9592698</td>\n",
       "      <td>5079</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>6247119</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>211|6047|60</td>\n",
       "      <td>211|6047|60|15|277|60|23|3</td>\n",
       "      <td>40|211|60|1|20|60|580|189|58|6469|4034|14067|1...</td>\n",
       "      <td>6047|1098</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>4298118681424644510</td>\n",
       "      <td>7687144</td>\n",
       "      <td>385</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>4144358</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>121|7894|95|6</td>\n",
       "      <td>121|7894|95|6|0|4732|95|146|4079</td>\n",
       "      <td>8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...</td>\n",
       "      <td>1467|7894|95|1098|601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>12057878999086460853</td>\n",
       "      <td>20164346</td>\n",
       "      <td>27961</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4144358</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1467|7894</td>\n",
       "      <td>51|0|1467|7894|8|183|94</td>\n",
       "      <td>36|1467|7894|37|51|1|173|32|26|214|154|26|234|...</td>\n",
       "      <td>1467|7894|95|1098|601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0</td>\n",
       "      <td>14347859242938208407</td>\n",
       "      <td>21963776</td>\n",
       "      <td>33931</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4144358</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1467|7894|95</td>\n",
       "      <td>1467|7894|95|1017|5003</td>\n",
       "      <td>1187|734|2|12872|1467|7894|95|1771|1|213|187|1...</td>\n",
       "      <td>1467|7894|95|1098|601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>4860571499428580850</td>\n",
       "      <td>21560664</td>\n",
       "      <td>37484</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>300|17</td>\n",
       "      <td>300|17|15|29|24|4849|1925|6302|350|1591</td>\n",
       "      <td>1925|6302|350|1591|300|17|1371|4493|3123|17458...</td>\n",
       "      <td>300|202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0</td>\n",
       "      <td>10664591050012752564</td>\n",
       "      <td>20620168</td>\n",
       "      <td>30128</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>490234</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>300|227</td>\n",
       "      <td>1323|1803|300|227|457|8|163|94|15320|7132</td>\n",
       "      <td>300|227|26889|9442|163|0|620|3123|0|300|176|23...</td>\n",
       "      <td>300|202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>4860571499428580850</td>\n",
       "      <td>21560664</td>\n",
       "      <td>37484</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>302819</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>300|17</td>\n",
       "      <td>300|17|15|29|24|4849|1925|6302|350|1591</td>\n",
       "      <td>1925|6302|350|1591|300|17|1371|4493|3123|17458...</td>\n",
       "      <td>7649|75|50|300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0</td>\n",
       "      <td>10664591050012752564</td>\n",
       "      <td>21443408</td>\n",
       "      <td>30128</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>302819</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>300|17</td>\n",
       "      <td>300|17|54|1323|14572|1591|15|457|8|163|94</td>\n",
       "      <td>300|227|4158|1591|1340|1913|768|641|15770|5751...</td>\n",
       "      <td>7649|75|50|300</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Click            DisplayURL      AdId  AdvertiserId  Depth  Position  \\\n",
       "0       0   4298118681424644510   7686695           385      3         3   \n",
       "1       0  13677630321509009335   3517124         23778      3         1   \n",
       "2       0  11689327222955583742  21021375         27701      3         2   \n",
       "3       0   4298118681424644510   7686695           385      1         1   \n",
       "4       0  15132506310926074459   4424000         20940      1         1   \n",
       "5       0   4298118681424644510   7686888           385      2         2   \n",
       "6       0  13883743923141531481  20568201         29921      2         1   \n",
       "7       0   4298118681424644510   7686578           385      1         1   \n",
       "8       0   4298118681424644510   7687047           385      1         1   \n",
       "9       0   4298118681424644510   7686572           385      2         2   \n",
       "10      0    514151146628080502   3691703         22478      2         1   \n",
       "11      0   4298118681424644510   7686612           385      2         2   \n",
       "12      0    694915634177245725   9592698          5079      2         1   \n",
       "13      0   4298118681424644510   7687144           385      3         3   \n",
       "14      0  12057878999086460853  20164346         27961      3         1   \n",
       "15      0  14347859242938208407  21963776         33931      3         2   \n",
       "16      0   4860571499428580850  21560664         37484      2         2   \n",
       "17      0  10664591050012752564  20620168         30128      2         1   \n",
       "18      0   4860571499428580850  21560664         37484      3         3   \n",
       "19      0  10664591050012752564  21443408         30128      3         1   \n",
       "\n",
       "      UserID  Gender  Age AdKeyword_tokens  \\\n",
       "0     490234       1    3       4133|95|17   \n",
       "1     490234       1    3             4133   \n",
       "2     490234       1    3             4133   \n",
       "3   16960371       2    2       4133|95|17   \n",
       "4    3524325       1    3      121|4133|95   \n",
       "5   11198634       2    3    14530|882|167   \n",
       "6   11198634       2    3    14530|95|1156   \n",
       "7   16905873       2    4    121|3337|95|6   \n",
       "8    5077836       2    3       1970|95|17   \n",
       "9    3519270       2    1      3337|95|126   \n",
       "10   3519270       2    1          3337|95   \n",
       "11   6247119       2    6          6047|60   \n",
       "12   6247119       2    6      211|6047|60   \n",
       "13   4144358       2    4    121|7894|95|6   \n",
       "14   4144358       2    4        1467|7894   \n",
       "15   4144358       2    4     1467|7894|95   \n",
       "16    490234       1    3           300|17   \n",
       "17    490234       1    3          300|227   \n",
       "18    302819       0    1           300|17   \n",
       "19    302819       0    1           300|17   \n",
       "\n",
       "                                  AdTitle_tokens  \\\n",
       "0                  4133|95|17|0|4732|95|146|4079   \n",
       "1                             145|65|3927|832|93   \n",
       "2                     4133|95|1|339|125|21|83093   \n",
       "3                  4133|95|17|0|4732|95|146|4079   \n",
       "4   121|4133|95|8762|3957|4563|2233|192|28|138|3   \n",
       "5               14530|882|167|0|4732|95|146|4079   \n",
       "6            14530|95|1156|0|42|45|14530|95|1156   \n",
       "7               121|3337|95|6|0|4732|95|146|4079   \n",
       "8                  1970|95|17|0|4732|95|146|4079   \n",
       "9                 3337|95|126|0|4732|95|146|4079   \n",
       "10                  3337|95|1|380|11435|5523|149   \n",
       "11                    6047|60|0|4732|95|146|4079   \n",
       "12                    211|6047|60|15|277|60|23|3   \n",
       "13              121|7894|95|6|0|4732|95|146|4079   \n",
       "14                       51|0|1467|7894|8|183|94   \n",
       "15                        1467|7894|95|1017|5003   \n",
       "16       300|17|15|29|24|4849|1925|6302|350|1591   \n",
       "17     1323|1803|300|227|457|8|163|94|15320|7132   \n",
       "18       300|17|15|29|24|4849|1925|6302|350|1591   \n",
       "19     300|17|54|1323|14572|1591|15|457|8|163|94   \n",
       "\n",
       "                                 AdDescription_tokens           Query_tokens  \n",
       "0   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...                   4133  \n",
       "1   3683|4990|2793|11589|21|10741|26|16044|26|3168...                   4133  \n",
       "2   726|50|2218|1533|2275|4133|1299|509|95|2072|1|...                   4133  \n",
       "3   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...              4133|4942  \n",
       "4   62|1162|570|8|4133|95|1|81|102|1155|650|1255|1...              4133|4942  \n",
       "5   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...          14530|95|1156  \n",
       "6   14530|95|1156|36|14530|4557|6|1|92|28|81|42|45...          14530|95|1156  \n",
       "7   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...                   3337  \n",
       "8   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...              1970|4942  \n",
       "9   8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...         304|20|3337|95  \n",
       "10  3337|95|1|1007|418|353|1|3337|70|270|89|130|68...         304|20|3337|95  \n",
       "11  8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...              6047|1098  \n",
       "12  40|211|60|1|20|60|580|189|58|6469|4034|14067|1...              6047|1098  \n",
       "13  8|81|123|205|2|95|26|95|60|32|1|17|146|1|991|3...  1467|7894|95|1098|601  \n",
       "14  36|1467|7894|37|51|1|173|32|26|214|154|26|234|...  1467|7894|95|1098|601  \n",
       "15  1187|734|2|12872|1467|7894|95|1771|1|213|187|1...  1467|7894|95|1098|601  \n",
       "16  1925|6302|350|1591|300|17|1371|4493|3123|17458...                300|202  \n",
       "17  300|227|26889|9442|163|0|620|3123|0|300|176|23...                300|202  \n",
       "18  1925|6302|350|1591|300|17|1371|4493|3123|17458...         7649|75|50|300  \n",
       "19  300|227|4158|1591|1340|1913|768|641|15770|5751...         7649|75|50|300  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "D10M.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Zbiór testowy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = pd.read_csv(\"D5M_test_x.tsv\", sep = \"\\t\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>Click</th>\n",
       "      <th>DisplayURL</th>\n",
       "      <th>AdId</th>\n",
       "      <th>AdvertiserId</th>\n",
       "      <th>Depth</th>\n",
       "      <th>Position</th>\n",
       "      <th>UserID</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Age</th>\n",
       "      <th>AdKeyword_tokens</th>\n",
       "      <th>AdTitle_tokens</th>\n",
       "      <th>AdDescription_tokens</th>\n",
       "      <th>Query_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>9860713142308876126</td>\n",
       "      <td>21197821</td>\n",
       "      <td>35826</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>9162163</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>143</td>\n",
       "      <td>162|5059|61|143|4|15|4185|462|49|3</td>\n",
       "      <td>143|1556|61|143|134|33|49|1|1157|1044|61|1|579...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>9860713142308876126</td>\n",
       "      <td>21197821</td>\n",
       "      <td>35826</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>13329036</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>143</td>\n",
       "      <td>162|5059|61|143|4|15|4185|462|49|3</td>\n",
       "      <td>143|1556|61|143|134|33|49|1|1157|1044|61|1|579...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>9860713142308876126</td>\n",
       "      <td>21197821</td>\n",
       "      <td>35826</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>295270</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>143</td>\n",
       "      <td>162|5059|61|143|4|15|4185|462|49|3</td>\n",
       "      <td>143|1556|61|143|134|33|49|1|1157|1044|61|1|579...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>6868758</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2447751</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>13308359</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>13338529</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>6893672</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>9183599</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>6882275</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>13338226</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>13351842</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0</td>\n",
       "      <td>12892941488276077458</td>\n",
       "      <td>21460946</td>\n",
       "      <td>34043</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>6893672</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>143</td>\n",
       "      <td>143|1876|1275|6057|1030</td>\n",
       "      <td>6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>9131601</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5503672</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5504928</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>9137680</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>9139943</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5507350</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0</td>\n",
       "      <td>14202400320781967402</td>\n",
       "      <td>20890861</td>\n",
       "      <td>27503</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>13326171</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>143</td>\n",
       "      <td>1878|143|123|213|17|454|3</td>\n",
       "      <td>1878|390|1354|149|1|2457|139|124|28|257|178</td>\n",
       "      <td>143</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Click            DisplayURL      AdId  AdvertiserId  Depth  Position  \\\n",
       "0       0   9860713142308876126  21197821         35826      2         1   \n",
       "1       0   9860713142308876126  21197821         35826      2         1   \n",
       "2       0   9860713142308876126  21197821         35826      2         1   \n",
       "3       0  12892941488276077458  21460946         34043      2         2   \n",
       "4       0  12892941488276077458  21460946         34043      2         2   \n",
       "5       0  12892941488276077458  21460946         34043      2         1   \n",
       "6       0  12892941488276077458  21460946         34043      2         2   \n",
       "7       0  12892941488276077458  21460946         34043      2         1   \n",
       "8       0  12892941488276077458  21460946         34043      3         3   \n",
       "9       0  12892941488276077458  21460946         34043      2         1   \n",
       "10      0  12892941488276077458  21460946         34043      2         2   \n",
       "11      0  12892941488276077458  21460946         34043      2         1   \n",
       "12      0  12892941488276077458  21460946         34043      2         1   \n",
       "13      0  14202400320781967402  20890861         27503      2         2   \n",
       "14      0  14202400320781967402  20890861         27503      2         2   \n",
       "15      0  14202400320781967402  20890861         27503      2         2   \n",
       "16      0  14202400320781967402  20890861         27503      2         2   \n",
       "17      0  14202400320781967402  20890861         27503      2         2   \n",
       "18      0  14202400320781967402  20890861         27503      2         2   \n",
       "19      0  14202400320781967402  20890861         27503      2         1   \n",
       "\n",
       "      UserID  Gender  Age AdKeyword_tokens  \\\n",
       "0    9162163       2    1              143   \n",
       "1   13329036       2    2              143   \n",
       "2     295270       1    3              143   \n",
       "3    6868758       2    5              143   \n",
       "4    2447751       1    1              143   \n",
       "5   13308359       2    5              143   \n",
       "6   13338529       1    6              143   \n",
       "7    6893672       1    2              143   \n",
       "8    9183599       1    1              143   \n",
       "9    6882275       2    1              143   \n",
       "10  13338226       1    2              143   \n",
       "11  13351842       2    2              143   \n",
       "12   6893672       1    2              143   \n",
       "13   9131601       0    1              143   \n",
       "14   5503672       1    3              143   \n",
       "15   5504928       1    3              143   \n",
       "16   9137680       1    3              143   \n",
       "17   9139943       1    4              143   \n",
       "18   5507350       2    2              143   \n",
       "19  13326171       1    3              143   \n",
       "\n",
       "                        AdTitle_tokens  \\\n",
       "0   162|5059|61|143|4|15|4185|462|49|3   \n",
       "1   162|5059|61|143|4|15|4185|462|49|3   \n",
       "2   162|5059|61|143|4|15|4185|462|49|3   \n",
       "3              143|1876|1275|6057|1030   \n",
       "4              143|1876|1275|6057|1030   \n",
       "5              143|1876|1275|6057|1030   \n",
       "6              143|1876|1275|6057|1030   \n",
       "7              143|1876|1275|6057|1030   \n",
       "8              143|1876|1275|6057|1030   \n",
       "9              143|1876|1275|6057|1030   \n",
       "10             143|1876|1275|6057|1030   \n",
       "11             143|1876|1275|6057|1030   \n",
       "12             143|1876|1275|6057|1030   \n",
       "13           1878|143|123|213|17|454|3   \n",
       "14           1878|143|123|213|17|454|3   \n",
       "15           1878|143|123|213|17|454|3   \n",
       "16           1878|143|123|213|17|454|3   \n",
       "17           1878|143|123|213|17|454|3   \n",
       "18           1878|143|123|213|17|454|3   \n",
       "19           1878|143|123|213|17|454|3   \n",
       "\n",
       "                                 AdDescription_tokens Query_tokens  \n",
       "0   143|1556|61|143|134|33|49|1|1157|1044|61|1|579...          143  \n",
       "1   143|1556|61|143|134|33|49|1|1157|1044|61|1|579...          143  \n",
       "2   143|1556|61|143|134|33|49|1|1157|1044|61|1|579...          143  \n",
       "3   143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...          143  \n",
       "4   143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...          143  \n",
       "5   143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...          143  \n",
       "6   143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...          143  \n",
       "7   143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...          143  \n",
       "8   143|1|6681|2|242|7277|2223|1|137|92|2312|1308|...          143  \n",
       "9   6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...          143  \n",
       "10  6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...          143  \n",
       "11  6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...          143  \n",
       "12  6681|2|242|7277|2223|1|137|92|2312|1308|1353|8...          143  \n",
       "13        1878|390|1354|149|1|2457|139|124|28|257|178          143  \n",
       "14        1878|390|1354|149|1|2457|139|124|28|257|178          143  \n",
       "15        1878|390|1354|149|1|2457|139|124|28|257|178          143  \n",
       "16        1878|390|1354|149|1|2457|139|124|28|257|178          143  \n",
       "17        1878|390|1354|149|1|2457|139|124|28|257|178          143  \n",
       "18        1878|390|1354|149|1|2457|139|124|28|257|178          143  \n",
       "19        1878|390|1354|149|1|2457|139|124|28|257|178          143  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modelowanie"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Chcemy przewidzieć, czy użytkownik kliknie na daną reklamę, czy nie. Wiemy, że w zbiorach D100k.tsv, D10M.tsv oraz D80M.tsv, było średnio 5% kliknięć użytkowników na reklamę."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zmienną objaśnianą jest zmienna \"Click\"."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Trenowałem modele na zbiorach D100k.tsv oraz D10M.tsv. Dla każdych modeli oraz zbiorów, dokonałem podziału na zbiór treningowy i walidacyjny w stosunku 80:20.\n",
    "\n",
    "#### Dla modelu, który uznałem za najlepszy, tzn. były najlepsze wyniki AUC dla D100k i D10M, dokonałem predykcji na zbiorze testowym D5M_test_x.tsv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modele regresji liniowej"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Model z samą zmienną Age."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_100k_1 = D100k.iloc[0:80000, :] \n",
    "test_100k_1 = D100k.iloc[80000:, :]    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_maly_1 = smf.ols('Click ~ Age', data = train_100k_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_maly_1 = model_maly_1.fit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                  Click   R-squared:                       0.000\n",
      "Model:                            OLS   Adj. R-squared:                  0.000\n",
      "Method:                 Least Squares   F-statistic:                     9.370\n",
      "Date:                Tue, 11 May 2021   Prob (F-statistic):            0.00221\n",
      "Time:                        21:43:10   Log-Likelihood:                 8570.1\n",
      "No. Observations:               80000   AIC:                        -1.714e+04\n",
      "Df Residuals:                   79998   BIC:                        -1.712e+04\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.0445      0.002     23.613      0.000       0.041       0.048\n",
      "Age            0.0017      0.001      3.061      0.002       0.001       0.003\n",
      "==============================================================================\n",
      "Omnibus:                    65325.429   Durbin-Watson:                   1.848\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):           994205.795\n",
      "Skew:                           4.141   Prob(JB):                         0.00\n",
      "Kurtosis:                      18.154   Cond. No.                         9.05\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "model_maly_1_summary = model_small_1.summary()\n",
    "print(model_maly_1_summary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "statsmodels.regression.linear_model.RegressionResultsWrapper"
      ]
     },
     "execution_count": 200,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(model_maly_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       -0.049487\n",
       "1       -0.049487\n",
       "2       -0.049487\n",
       "3       -0.047816\n",
       "4       -0.049487\n",
       "           ...   \n",
       "79995   -0.049487\n",
       "79996   -0.049487\n",
       "79997   -0.049487\n",
       "79998   -0.049487\n",
       "79999   -0.049487\n",
       "Length: 80000, dtype: float64"
      ]
     },
     "execution_count": 201,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_1.resid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "80000    0.049487\n",
       "80001    0.049487\n",
       "80002    0.049487\n",
       "80003    0.049487\n",
       "80004    0.049487\n",
       "           ...   \n",
       "99995    0.049487\n",
       "99996    0.049487\n",
       "99997    0.052829\n",
       "99998    0.047816\n",
       "99999    0.051158\n",
       "Length: 20000, dtype: float64"
      ]
     },
     "execution_count": 202,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_1 = model_small_1.predict(test_100k_1.loc[:, [\"Click\", \"Age\"]])\n",
    "pred_maly_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.04987362976486753"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_1.mean()\n",
    "# 0.04987362976486753"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Czyli średnia predykcji wynosi prawie 0.05, czyli 5%, a więc się zgadza z średnią liczbą kliknięć użytkownika na reklamę."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5202668887316323"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_100k_1.loc[:,\"Click\"], pred_maly_1)\n",
    "# 0.5202668887316323"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zatem wartość AUC wynosi 0.520."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_10M_1 = D10M.iloc[0:80000, :] \n",
    "test_10M_1 = D10M.iloc[80000:, :]    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_sredni_1 = smf.ols('Click ~ Age', data = train_10M_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_sredni_1 = model_sredni_1.fit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       -0.049487\n",
       "1       -0.049487\n",
       "2       -0.049487\n",
       "3       -0.047816\n",
       "4       -0.049487\n",
       "           ...   \n",
       "79995   -0.049487\n",
       "79996   -0.049487\n",
       "79997   -0.049487\n",
       "79998   -0.049487\n",
       "79999   -0.049487\n",
       "Length: 80000, dtype: float64"
      ]
     },
     "execution_count": 192,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_1.resid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "80000      0.049487\n",
       "80001      0.049487\n",
       "80002      0.049487\n",
       "80003      0.049487\n",
       "80004      0.049487\n",
       "             ...   \n",
       "9999995    0.046145\n",
       "9999996    0.046145\n",
       "9999997    0.049487\n",
       "9999998    0.052829\n",
       "9999999    0.052829\n",
       "Length: 9920000, dtype: float64"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_1 = model_sredni_1.predict(test_10M_1.loc[:, [\"Click\", \"Age\"]])\n",
    "pred_sredni_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.050137931819387686"
      ]
     },
     "execution_count": 194,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_1.mean()\n",
    "# 0.050137931819387686"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5094500032414795"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_10M_1.loc[:,\"Click\"], pred_sredni_1)\n",
    "# 0.5094500032414795"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.509"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. Model ze zmienną \"Age\" i \"Gender\"."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Średni zbiór "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_maly_2 = smf.ols('Click ~ Age + Gender', data = train_100k_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_maly_2 = model_maly_2.fit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                  Click   R-squared:                       0.000\n",
      "Model:                            OLS   Adj. R-squared:                  0.000\n",
      "Method:                 Least Squares   F-statistic:                     4.693\n",
      "Date:                Tue, 11 May 2021   Prob (F-statistic):            0.00916\n",
      "Time:                        21:43:58   Log-Likelihood:                 8570.1\n",
      "No. Observations:               80000   AIC:                        -1.713e+04\n",
      "Df Residuals:                   79997   BIC:                        -1.711e+04\n",
      "Df Model:                           2                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.0447      0.003     16.707      0.000       0.039       0.050\n",
      "Age            0.0017      0.001      3.058      0.002       0.001       0.003\n",
      "Gender        -0.0002      0.001     -0.129      0.897      -0.003       0.003\n",
      "==============================================================================\n",
      "Omnibus:                    65325.408   Durbin-Watson:                   1.848\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):           994204.699\n",
      "Skew:                           4.141   Prob(JB):                         0.00\n",
      "Kurtosis:                      18.154   Cond. No.                         14.5\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "model_maly_2_summary = model_maly_2.summary()\n",
    "print(model_maly_2_summary)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Widzimy, że obie zmienne są istotne."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       -0.049566\n",
       "1       -0.049566\n",
       "2       -0.049566\n",
       "3       -0.047700\n",
       "4       -0.049566\n",
       "           ...   \n",
       "79995   -0.049566\n",
       "79996   -0.049566\n",
       "79997   -0.049566\n",
       "79998   -0.049566\n",
       "79999   -0.049566\n",
       "Length: 80000, dtype: float64"
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_2.resid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "80000    0.049566\n",
       "80001    0.049566\n",
       "80002    0.049566\n",
       "80003    0.049566\n",
       "80004    0.049566\n",
       "           ...   \n",
       "99995    0.049566\n",
       "99996    0.049566\n",
       "99997    0.052735\n",
       "99998    0.047700\n",
       "99999    0.051245\n",
       "Length: 20000, dtype: float64"
      ]
     },
     "execution_count": 209,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_2 = model_maly_2.predict(test_100k_1.loc[:, [\"Click\", \"Age\", \"Gender\"]])\n",
    "pred_maly_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.049866805068917774"
      ]
     },
     "execution_count": 210,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_2.mean()\n",
    "# 0.049866805068917774"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5199692471759081"
      ]
     },
     "execution_count": 211,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_100k_1.loc[:,\"Click\"], pred_maly_2)\n",
    "# 0.5094500032414795"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.520"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.05014049764727051"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_2 = smf.ols('Click ~ Age + Gender', data = train_10M_1)\n",
    "model_sredni_2 = model_sredni_2.fit()\n",
    "pred_sredni_2 = model_sredni_2.predict(test_10M_1.loc[:, [\"Click\", \"Age\", \"Gender\"]])\n",
    "pred_sredni_2.mean()\n",
    "# 0.05014049764727051"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5084655293973371"
      ]
     },
     "execution_count": 214,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_10M_1.loc[:,\"Click\"], pred_sredni_2)\n",
    "# 0.5084655293973371"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.508"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3. Model z samą zmienną \"Gender\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.04974794387199179"
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_3 = smf.ols('Click ~ Gender', data = train_100k_1)\n",
    "model_maly_3 = model_maly_3.fit()\n",
    "pred_maly_3 = model_maly_3.predict(test_100k_1.loc[:, [\"Click\", \"Gender\"]])\n",
    "pred_maly_3.mean()\n",
    "# 0.04974794387199179"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5046949656571331"
      ]
     },
     "execution_count": 220,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_100k_1.loc[:,\"Click\"], pred_maly_3)\n",
    "# 0.5046949656571331"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.505"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.04973632202593676"
      ]
     },
     "execution_count": 222,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_3 = smf.ols('Click ~ Gender', data = train_10M_1)\n",
    "model_sredni_3 = model_sredni_3.fit()\n",
    "pred_sredni_3 = model_sredni_3.predict(test_10M_1.loc[:, [\"Click\", \"Gender\"]])\n",
    "pred_sredni_3.mean()\n",
    "# 0.04973632202593676"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5035258805338853"
      ]
     },
     "execution_count": 223,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_10M_1.loc[:,\"Click\"], pred_sredni_3)\n",
    "# 0.5035258805338853"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.504"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. Model z wszystkimi małymi zmiennymi: \"Age\", \"Gender\", \"Depth\", \"Position\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 398,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.04951705040010446"
      ]
     },
     "execution_count": 398,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_4 = smf.ols('Click ~ Age + Gender + Depth + Position', data = train_100k_1)\n",
    "model_maly_4 = model_maly_4.fit()\n",
    "pred_maly_4 = model_maly_4.predict(test_100k_1.loc[:, [\"Click\", \"Age\", \"Gender\", \"Depth\", \"Position\"]])\n",
    "pred_maly_4.mean()\n",
    "# 0.04951705040010446"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 399,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6187734984617962"
      ]
     },
     "execution_count": 399,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_100k_1.loc[:,\"Click\"], pred_maly_4)\n",
    "# 0.6187734984617962"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.619"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Przy uwzględnieniu wszystkich małych zmiennych widać więc wyraźną poprawę."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 400,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.05009517168821361"
      ]
     },
     "execution_count": 400,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_4 = smf.ols('Click ~ Age + Gender + Depth + Position', data = train_10M_1)\n",
    "model_sredni_4 = model_sredni_4.fit()\n",
    "pred_sredni_4 = model_sredni_4.predict(test_10M_1.loc[:, [\"Click\", \"Age\", \"Gender\", \"Depth\", \"Position\"]])\n",
    "pred_sredni_4.mean()\n",
    "# 0.05009517168821361"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 401,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5981106235060497"
      ]
     },
     "execution_count": 401,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(test_10M_1.loc[:,\"Click\"], pred_sredni_4)\n",
    "# 0.5981106235060497"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.598"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### A więc w przypadku zbioru średniego, również widać wyraźną poprawę."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Generalnie, AUC na zbiorze walidacyjnym ze zbioru o 10 milionach wierszy wychodziło niższe niż AUC na zbiorze walidacyjnym ze zbioru o 100 tysiącach wierszy. Wynika to z tego, że w przypadku zbioru D10M, testujemy na dużo większym zbiorze."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modele z użyciem pakietu tensorflow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5. Model z wszystkimi małymi zmiennymi: \"Age\", \"Gender\", Depth\", \"Position\" - 1 warstwa z aktywacją za pomocą funkcji ReLU. Kompilowałem model z wykorzystaniem klasyfikatora stochastycznego spadku gradientu SGD z binarną funkcją straty krzyżowej entropii.\n",
    "\n",
    "#### Liczba epok = 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_100k_5 = D100k.iloc[0:80000, :] \n",
    "test_100k_5 = D100k.iloc[80000:, :]\n",
    "\n",
    "X_train_maly_5 = train_100k_5.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\"]]\n",
    "Y_train_maly_5 = train_100k_5.loc[:, \"Click\"]\n",
    "                                        \n",
    "X_test_maly_5 = test_100k_5.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\"]]\n",
    "Y_test_maly_5 = test_100k_5.loc[:, \"Click\"]                                        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_maly_5 = tf.keras.models.Sequential()\n",
    "model_maly_5.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_maly_5.add(tf.keras.layers.Dense(1, activation = \"relu\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "2000/2000 [==============================] - 2s 653us/step - loss: 0.9809 - val_loss: 0.8329\n",
      "Epoch 2/5\n",
      "2000/2000 [==============================] - 1s 639us/step - loss: 0.7447 - val_loss: 0.8329\n",
      "Epoch 3/5\n",
      "2000/2000 [==============================] - 1s 603us/step - loss: 0.7400 - val_loss: 0.8329\n",
      "Epoch 4/5\n",
      "2000/2000 [==============================] - 1s 596us/step - loss: 0.7405 - val_loss: 0.8329\n",
      "Epoch 5/5\n",
      "2000/2000 [==============================] - 1s 593us/step - loss: 0.7447 - val_loss: 0.8329\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290dea6f190>"
      ]
     },
     "execution_count": 299,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_5.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_5.fit(X_train_maly_5, Y_train_maly_5, epochs = 5, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 313,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 313,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_maly_5.predict(X_test_maly_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 301,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_maly_5 = model_maly_5.predict(X_test_maly_5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.],\n",
       "       [0.],\n",
       "       [0.],\n",
       "       ...,\n",
       "       [0.],\n",
       "       [0.],\n",
       "       [0.]], dtype=float32)"
      ]
     },
     "execution_count": 302,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 303,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 303,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(Y_test_maly_5, pred_maly_5)\n",
    "# 0.5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.500"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_10M_5 = D10M.iloc[0:80000, :] \n",
    "test_10M_5 = D10M.iloc[80000:, :]\n",
    "\n",
    "X_train_sredni_5 = train_10M_5.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\"]]\n",
    "Y_train_sredni_5 = train_10M_5.loc[:, \"Click\"]\n",
    "                                        \n",
    "X_test_sredni_5 = test_10M_5.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\"]]\n",
    "Y_test_sredni_5 = test_10M_5.loc[:, \"Click\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 305,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_sredni_5 = tf.keras.models.Sequential()\n",
    "model_sredni_5.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_sredni_5.add(tf.keras.layers.Dense(1, activation = \"relu\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "2000/2000 [==============================] - 2s 641us/step - loss: 0.7342 - val_loss: 0.8329\n",
      "Epoch 2/5\n",
      "2000/2000 [==============================] - 1s 595us/step - loss: 0.7327 - val_loss: 0.8329\n",
      "Epoch 3/5\n",
      "2000/2000 [==============================] - 1s 595us/step - loss: 0.7577 - val_loss: 0.8329\n",
      "Epoch 4/5\n",
      "2000/2000 [==============================] - 1s 627us/step - loss: 0.7444 - val_loss: 0.8329\n",
      "Epoch 5/5\n",
      "2000/2000 [==============================] - 1s 628us/step - loss: 0.7380 - val_loss: 0.8329\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290df7fa130>"
      ]
     },
     "execution_count": 306,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_5.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_5.fit(X_train_sredni_5, Y_train_sredni_5, epochs = 5, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 307,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 307,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_sredni_5.predict(X_test_sredni_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 308,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_sredni_5 = model_sredni_5.predict(X_test_sredni_5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 310,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(Y_test_sredni_5, pred_sredni_5)\n",
    "# 0.5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.500"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 6. Model przy zwiększonej liczbie warstw z funkcją ReLU do 2 oraz dodatkowo z jedną warstwą przy aktywacji za pomocą funkcji sigmoidalnej."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 336,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "2000/2000 [==============================] - 2s 691us/step - loss: 0.4210 - val_loss: 0.2102\n",
      "Epoch 2/5\n",
      "2000/2000 [==============================] - 1s 613us/step - loss: 0.1964 - val_loss: 0.2105\n",
      "Epoch 3/5\n",
      "2000/2000 [==============================] - 1s 623us/step - loss: 0.1924 - val_loss: 0.2100\n",
      "Epoch 4/5\n",
      "2000/2000 [==============================] - 1s 615us/step - loss: 0.1952 - val_loss: 0.2102\n",
      "Epoch 5/5\n",
      "2000/2000 [==============================] - 1s 606us/step - loss: 0.1925 - val_loss: 0.2097\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290dae95490>"
      ]
     },
     "execution_count": 336,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_6 = tf.keras.models.Sequential()\n",
    "model_maly_6.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_maly_6.add(tf.keras.layers.Dense(2, activation = \"relu\"))\n",
    "model_maly_6.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_maly_6.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_6.fit(X_train_maly_5, Y_train_maly_5, epochs = 5, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 337,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 337,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_maly_6.predict(X_test_maly_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 338,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5761708885119532"
      ]
     },
     "execution_count": 338,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_6 = model_maly_6.predict(X_test_maly_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_maly_5, pred_maly_6)\n",
    "# 0.5761708885119532"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.576"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Widać więc znaczącą poprawę w AUC."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 339,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "2000/2000 [==============================] - 2s 664us/step - loss: 0.2121 - val_loss: 0.2176\n",
      "Epoch 2/5\n",
      "2000/2000 [==============================] - 1s 607us/step - loss: 0.1957 - val_loss: 0.2168\n",
      "Epoch 3/5\n",
      "2000/2000 [==============================] - 1s 615us/step - loss: 0.1991 - val_loss: 0.2164\n",
      "Epoch 4/5\n",
      "2000/2000 [==============================] - 1s 624us/step - loss: 0.1976 - val_loss: 0.2153\n",
      "Epoch 5/5\n",
      "2000/2000 [==============================] - 1s 588us/step - loss: 0.1962 - val_loss: 0.2142\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290df6633a0>"
      ]
     },
     "execution_count": 339,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_6 = tf.keras.models.Sequential()\n",
    "model_sredni_6.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_sredni_6.add(tf.keras.layers.Dense(2, activation = \"relu\"))\n",
    "model_sredni_6.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_sredni_6.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_6.fit(X_train_sredni_5, Y_train_sredni_5, epochs = 5, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 340,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 340,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_sredni_6.predict(X_test_sredni_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5616961310938284"
      ]
     },
     "execution_count": 341,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_6 = model_sredni_6.predict(X_test_sredni_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_sredni_5, pred_sredni_6)\n",
    "# 0.5616961310938284"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.562"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Stąd w przypadku zbioru średniego, również widać wyraźną poprawę."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 7. Model przy zwiększonej liczbie warstw z funkcją ReLU z 2 do 3."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 342,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "2000/2000 [==============================] - 2s 667us/step - loss: 0.3023 - val_loss: 0.2174\n",
      "Epoch 2/5\n",
      "2000/2000 [==============================] - 1s 629us/step - loss: 0.1928 - val_loss: 0.2126\n",
      "Epoch 3/5\n",
      "2000/2000 [==============================] - 1s 635us/step - loss: 0.1936 - val_loss: 0.2117\n",
      "Epoch 4/5\n",
      "2000/2000 [==============================] - 1s 624us/step - loss: 0.1920 - val_loss: 0.2110\n",
      "Epoch 5/5\n",
      "2000/2000 [==============================] - 1s 631us/step - loss: 0.1914 - val_loss: 0.2107\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290dc871fa0>"
      ]
     },
     "execution_count": 342,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_7 = tf.keras.models.Sequential()\n",
    "model_maly_7.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_maly_7.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_maly_7.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_maly_7.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_7.fit(X_train_maly_5, Y_train_maly_5, epochs = 5, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 343,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 343,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_maly_7.predict(X_test_maly_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5889738082183392"
      ]
     },
     "execution_count": 344,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_7 = model_maly_7.predict(X_test_maly_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_maly_5, pred_maly_7)\n",
    "# 0.5889738082183392"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.589"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zatem AUC jeszcze trochę wzrosło."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "2000/2000 [==============================] - 2s 665us/step - loss: 0.2942 - val_loss: 0.2252\n",
      "Epoch 2/5\n",
      "2000/2000 [==============================] - 1s 631us/step - loss: 0.2009 - val_loss: 0.2194\n",
      "Epoch 3/5\n",
      "2000/2000 [==============================] - 1s 649us/step - loss: 0.1961 - val_loss: 0.2161\n",
      "Epoch 4/5\n",
      "2000/2000 [==============================] - 1s 672us/step - loss: 0.1935 - val_loss: 0.2142\n",
      "Epoch 5/5\n",
      "2000/2000 [==============================] - 1s 620us/step - loss: 0.1956 - val_loss: 0.2132\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290de750ac0>"
      ]
     },
     "execution_count": 345,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_7 = tf.keras.models.Sequential()\n",
    "model_sredni_7.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_sredni_7.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_sredni_7.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_sredni_7.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_7.fit(X_train_sredni_5, Y_train_sredni_5, epochs = 5, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 346,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 346,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_sredni_7.predict(X_test_sredni_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5644865906420553"
      ]
     },
     "execution_count": 348,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_7 = model_sredni_7.predict(X_test_sredni_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_sredni_5, pred_sredni_7)\n",
    "# 0.5644865906420553"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.564"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zatem dla zbioru średniego, również AUC trochę się zwiększyło."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 8. Model przy zwiększonej liczbie epok z 5 do 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "2000/2000 [==============================] - 2s 666us/step - loss: 0.2245 - val_loss: 0.2201\n",
      "Epoch 2/10\n",
      "2000/2000 [==============================] - 1s 601us/step - loss: 0.2037 - val_loss: 0.2156\n",
      "Epoch 3/10\n",
      "2000/2000 [==============================] - 1s 604us/step - loss: 0.1958 - val_loss: 0.2130\n",
      "Epoch 4/10\n",
      "2000/2000 [==============================] - 1s 600us/step - loss: 0.1958 - val_loss: 0.2126\n",
      "Epoch 5/10\n",
      "2000/2000 [==============================] - 1s 607us/step - loss: 0.1935 - val_loss: 0.2109\n",
      "Epoch 6/10\n",
      "2000/2000 [==============================] - 1s 608us/step - loss: 0.1911 - val_loss: 0.2108\n",
      "Epoch 7/10\n",
      "2000/2000 [==============================] - 1s 604us/step - loss: 0.1963 - val_loss: 0.2102\n",
      "Epoch 8/10\n",
      "2000/2000 [==============================] - 1s 608us/step - loss: 0.1932 - val_loss: 0.2097\n",
      "Epoch 9/10\n",
      "2000/2000 [==============================] - 1s 612us/step - loss: 0.1912 - val_loss: 0.2094\n",
      "Epoch 10/10\n",
      "2000/2000 [==============================] - 1s 627us/step - loss: 0.1893 - val_loss: 0.2105\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290eb07f0a0>"
      ]
     },
     "execution_count": 349,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_8 = tf.keras.models.Sequential()\n",
    "model_maly_8.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_maly_8.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_maly_8.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_maly_8.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_8.fit(X_train_maly_5, Y_train_maly_5, epochs = 10, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 350,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_maly_8.predict(X_test_maly_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 351,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5998540011758232"
      ]
     },
     "execution_count": 351,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_8 = model_maly_8.predict(X_test_maly_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_maly_5, pred_maly_8)\n",
    "# 0.5998540011758232"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.600"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Stąd zwiększenie liczby epok poprawiło rezultaty."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "2000/2000 [==============================] - 2s 661us/step - loss: 0.2182 - val_loss: 0.2211\n",
      "Epoch 2/10\n",
      "2000/2000 [==============================] - 1s 612us/step - loss: 0.1989 - val_loss: 0.2177\n",
      "Epoch 3/10\n",
      "2000/2000 [==============================] - 1s 610us/step - loss: 0.1996 - val_loss: 0.2156\n",
      "Epoch 4/10\n",
      "2000/2000 [==============================] - 1s 604us/step - loss: 0.1976 - val_loss: 0.2144\n",
      "Epoch 5/10\n",
      "2000/2000 [==============================] - 1s 605us/step - loss: 0.1946 - val_loss: 0.2143\n",
      "Epoch 6/10\n",
      "2000/2000 [==============================] - 1s 607us/step - loss: 0.1937 - val_loss: 0.2137\n",
      "Epoch 7/10\n",
      "2000/2000 [==============================] - 1s 608us/step - loss: 0.1950 - val_loss: 0.2124\n",
      "Epoch 8/10\n",
      "2000/2000 [==============================] - 1s 613us/step - loss: 0.1968 - val_loss: 0.2119\n",
      "Epoch 9/10\n",
      "2000/2000 [==============================] - 1s 611us/step - loss: 0.1913 - val_loss: 0.2110\n",
      "Epoch 10/10\n",
      "2000/2000 [==============================] - 1s 598us/step - loss: 0.1932 - val_loss: 0.2110\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290dd567670>"
      ]
     },
     "execution_count": 356,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_8 = tf.keras.models.Sequential()\n",
    "model_sredni_8.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_sredni_8.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_sredni_8.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_sredni_8.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_8.fit(X_train_sredni_5, Y_train_sredni_5, epochs = 10, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 357,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 357,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_sredni_8.predict(X_test_sredni_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 358,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.567679172442108"
      ]
     },
     "execution_count": 358,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_8 = model_sredni_8.predict(X_test_sredni_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_sredni_5, pred_sredni_8)\n",
    "# 0.567679172442108"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.568"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zatem dla zbioru średniego także AUC trochę się poprawiło."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 9. Model ze zwiększoną liczbą epok z 10 do 15"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "2000/2000 [==============================] - 2s 658us/step - loss: 0.2268 - val_loss: 0.2184\n",
      "Epoch 2/15\n",
      "2000/2000 [==============================] - 1s 614us/step - loss: 0.1964 - val_loss: 0.2154\n",
      "Epoch 3/15\n",
      "2000/2000 [==============================] - 1s 616us/step - loss: 0.1925 - val_loss: 0.2147\n",
      "Epoch 4/15\n",
      "2000/2000 [==============================] - 1s 601us/step - loss: 0.1966 - val_loss: 0.2134\n",
      "Epoch 5/15\n",
      "2000/2000 [==============================] - 1s 608us/step - loss: 0.1942 - val_loss: 0.2123\n",
      "Epoch 6/15\n",
      "2000/2000 [==============================] - 1s 610us/step - loss: 0.1922 - val_loss: 0.2124\n",
      "Epoch 7/15\n",
      "2000/2000 [==============================] - 1s 603us/step - loss: 0.1930 - val_loss: 0.2117\n",
      "Epoch 8/15\n",
      "2000/2000 [==============================] - 1s 612us/step - loss: 0.1947 - val_loss: 0.2107\n",
      "Epoch 9/15\n",
      "2000/2000 [==============================] - 1s 599us/step - loss: 0.1942 - val_loss: 0.2113\n",
      "Epoch 10/15\n",
      "2000/2000 [==============================] - 1s 613us/step - loss: 0.1917 - val_loss: 0.2101\n",
      "Epoch 11/15\n",
      "2000/2000 [==============================] - 1s 616us/step - loss: 0.1954 - val_loss: 0.2107\n",
      "Epoch 12/15\n",
      "2000/2000 [==============================] - 1s 614us/step - loss: 0.1950 - val_loss: 0.2110\n",
      "Epoch 13/15\n",
      "2000/2000 [==============================] - 1s 609us/step - loss: 0.1948 - val_loss: 0.2095\n",
      "Epoch 14/15\n",
      "2000/2000 [==============================] - 1s 620us/step - loss: 0.1950 - val_loss: 0.2104\n",
      "Epoch 15/15\n",
      "2000/2000 [==============================] - 1s 634us/step - loss: 0.1922 - val_loss: 0.2092\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290eb1288e0>"
      ]
     },
     "execution_count": 362,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_9 = tf.keras.models.Sequential()\n",
    "model_maly_9.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_maly_9.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_maly_9.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_maly_9.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_9.fit(X_train_maly_5, Y_train_maly_5, epochs = 15, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 363,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 363,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_maly_9.predict(X_test_maly_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 364,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6121626147575743"
      ]
     },
     "execution_count": 364,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_9 = model_maly_9.predict(X_test_maly_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_maly_5, pred_maly_9)\n",
    "# 0.6121626147575743"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.612"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 365,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "2000/2000 [==============================] - 2s 691us/step - loss: 0.2141 - val_loss: 0.2229\n",
      "Epoch 2/15\n",
      "2000/2000 [==============================] - 1s 640us/step - loss: 0.1995 - val_loss: 0.2173\n",
      "Epoch 3/15\n",
      "2000/2000 [==============================] - 1s 604us/step - loss: 0.1938 - val_loss: 0.2140\n",
      "Epoch 4/15\n",
      "2000/2000 [==============================] - 1s 615us/step - loss: 0.1938 - val_loss: 0.2128\n",
      "Epoch 5/15\n",
      "2000/2000 [==============================] - 1s 600us/step - loss: 0.1917 - val_loss: 0.2119\n",
      "Epoch 6/15\n",
      "2000/2000 [==============================] - 1s 604us/step - loss: 0.1940 - val_loss: 0.2115\n",
      "Epoch 7/15\n",
      "2000/2000 [==============================] - 1s 606us/step - loss: 0.1980 - val_loss: 0.2113\n",
      "Epoch 8/15\n",
      "2000/2000 [==============================] - 1s 603us/step - loss: 0.1938 - val_loss: 0.2111\n",
      "Epoch 9/15\n",
      "2000/2000 [==============================] - 1s 611us/step - loss: 0.1923 - val_loss: 0.2105\n",
      "Epoch 10/15\n",
      "2000/2000 [==============================] - 1s 638us/step - loss: 0.1948 - val_loss: 0.2100\n",
      "Epoch 11/15\n",
      "2000/2000 [==============================] - 1s 606us/step - loss: 0.1944 - val_loss: 0.2097\n",
      "Epoch 12/15\n",
      "2000/2000 [==============================] - 1s 603us/step - loss: 0.1885 - val_loss: 0.2104\n",
      "Epoch 13/15\n",
      "2000/2000 [==============================] - 1s 604us/step - loss: 0.1891 - val_loss: 0.2099\n",
      "Epoch 14/15\n",
      "2000/2000 [==============================] - 1s 635us/step - loss: 0.1899 - val_loss: 0.2100\n",
      "Epoch 15/15\n",
      "2000/2000 [==============================] - 1s 598us/step - loss: 0.1918 - val_loss: 0.2098\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290f0cf7160>"
      ]
     },
     "execution_count": 365,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_9 = tf.keras.models.Sequential()\n",
    "model_sredni_9.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_sredni_9.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_sredni_9.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_sredni_9.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_9.fit(X_train_sredni_5, Y_train_sredni_5, epochs = 15, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 366,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 366,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_sredni_9.predict(X_test_sredni_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 367,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5845638568854904"
      ]
     },
     "execution_count": 367,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_9 = model_sredni_9.predict(X_test_sredni_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_sredni_5, pred_sredni_9)\n",
    "# 0.5845638568854904"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.585"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Stąd, w obu zbiorach, AUC wzrosło."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 10. Model ze zwiększoną liczbą warstw z funkcją ReLU z 3 do 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 368,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "2000/2000 [==============================] - 2s 657us/step - loss: 0.2399 - val_loss: 0.2213\n",
      "Epoch 2/15\n",
      "2000/2000 [==============================] - 1s 601us/step - loss: 0.1990 - val_loss: 0.2142\n",
      "Epoch 3/15\n",
      "2000/2000 [==============================] - 1s 629us/step - loss: 0.1975 - val_loss: 0.2131\n",
      "Epoch 4/15\n",
      "2000/2000 [==============================] - 1s 611us/step - loss: 0.1881 - val_loss: 0.2122\n",
      "Epoch 5/15\n",
      "2000/2000 [==============================] - 1s 608us/step - loss: 0.1925 - val_loss: 0.2120\n",
      "Epoch 6/15\n",
      "2000/2000 [==============================] - 1s 626us/step - loss: 0.1905 - val_loss: 0.2114\n",
      "Epoch 7/15\n",
      "2000/2000 [==============================] - 1s 614us/step - loss: 0.1910 - val_loss: 0.2112\n",
      "Epoch 8/15\n",
      "2000/2000 [==============================] - 1s 600us/step - loss: 0.1940 - val_loss: 0.2105\n",
      "Epoch 9/15\n",
      "2000/2000 [==============================] - 1s 603us/step - loss: 0.1911 - val_loss: 0.2111\n",
      "Epoch 10/15\n",
      "2000/2000 [==============================] - 1s 610us/step - loss: 0.1956 - val_loss: 0.2108\n",
      "Epoch 11/15\n",
      "2000/2000 [==============================] - 1s 597us/step - loss: 0.1911 - val_loss: 0.2100\n",
      "Epoch 12/15\n",
      "2000/2000 [==============================] - 1s 607us/step - loss: 0.1955 - val_loss: 0.2103\n",
      "Epoch 13/15\n",
      "2000/2000 [==============================] - 1s 608us/step - loss: 0.1949 - val_loss: 0.2097\n",
      "Epoch 14/15\n",
      "2000/2000 [==============================] - 1s 617us/step - loss: 0.1923 - val_loss: 0.2094\n",
      "Epoch 15/15\n",
      "2000/2000 [==============================] - 1s 602us/step - loss: 0.1895 - val_loss: 0.2104\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x290f0b89b50>"
      ]
     },
     "execution_count": 368,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_maly_10 = tf.keras.models.Sequential()\n",
    "model_maly_10.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_maly_10.add(tf.keras.layers.Dense(4, activation = \"relu\"))\n",
    "model_maly_10.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_maly_10.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_10.fit(X_train_maly_5, Y_train_maly_5, epochs = 15, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 370,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_maly_10.predict(X_test_maly_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 371,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6011917422713919"
      ]
     },
     "execution_count": 371,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_10 = model_maly_10.predict(X_test_maly_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_maly_5, pred_maly_10)\n",
    "# 0.6011917422713919"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.601"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "2000/2000 [==============================] - 2s 702us/step - loss: 0.2754 - val_loss: 0.2205\n",
      "Epoch 2/15\n",
      "2000/2000 [==============================] - 1s 632us/step - loss: 0.2005 - val_loss: 0.2150\n",
      "Epoch 3/15\n",
      "2000/2000 [==============================] - 1s 640us/step - loss: 0.1943 - val_loss: 0.2128\n",
      "Epoch 4/15\n",
      "2000/2000 [==============================] - 1s 649us/step - loss: 0.1900 - val_loss: 0.2119\n",
      "Epoch 5/15\n",
      "2000/2000 [==============================] - 1s 607us/step - loss: 0.1919 - val_loss: 0.2121\n",
      "Epoch 6/15\n",
      "2000/2000 [==============================] - 1s 642us/step - loss: 0.1942 - val_loss: 0.2113\n",
      "Epoch 7/15\n",
      "2000/2000 [==============================] - 1s 637us/step - loss: 0.1929 - val_loss: 0.2111\n",
      "Epoch 8/15\n",
      "2000/2000 [==============================] - 1s 632us/step - loss: 0.1917 - val_loss: 0.2109\n",
      "Epoch 9/15\n",
      "2000/2000 [==============================] - 1s 644us/step - loss: 0.1961 - val_loss: 0.2113\n",
      "Epoch 10/15\n",
      "2000/2000 [==============================] - 1s 691us/step - loss: 0.1950 - val_loss: 0.2103\n",
      "Epoch 11/15\n",
      "2000/2000 [==============================] - 1s 601us/step - loss: 0.1883 - val_loss: 0.2101\n",
      "Epoch 12/15\n",
      "2000/2000 [==============================] - 1s 643us/step - loss: 0.1901 - val_loss: 0.2105\n",
      "Epoch 13/15\n",
      "2000/2000 [==============================] - 1s 629us/step - loss: 0.1912 - val_loss: 0.2099\n",
      "Epoch 14/15\n",
      "2000/2000 [==============================] - 1s 663us/step - loss: 0.1899 - val_loss: 0.2103\n",
      "Epoch 15/15\n",
      "2000/2000 [==============================] - 1s 606us/step - loss: 0.1919 - val_loss: 0.2100\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x22459bb0580>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_sredni_10 = tf.keras.models.Sequential()\n",
    "model_sredni_10.add(tf.keras.Input(shape = (4, ))) # Były 4 zmienne objaśniające, zatem dałem wymiar równy (4, ).\n",
    "model_sredni_10.add(tf.keras.layers.Dense(4, activation = \"relu\"))\n",
    "model_sredni_10.add(tf.keras.layers.Dense(1, activation = \"sigmoid\"))\n",
    "model_sredni_10.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_10.fit(X_train_sredni_5, Y_train_sredni_5, epochs = 15, validation_split = 0.2, verbose = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(model_sredni_10.predict(X_test_sredni_5) > 0.5)\n",
    "# 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5814521541042691"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sredni_10 = model_sredni_10.predict(X_test_sredni_5)\n",
    "sklearn.metrics.roc_auc_score(Y_test_sredni_5, pred_sredni_10)\n",
    "# 0.586129285686636"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.586"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zatem, w przypadku małego zbioru, AUC, trochę spadło, a w przypadku średniego zbioru, bardzo niewiele wzrosło. Stąd nie warto zwiększać liczby warstw przy funkcji ReLU z 3 do 4. Optymalna liczba warstw to 3.\n",
    "#### Natomiast w przypadku liczby epok, wybrałem liczbę 15."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Później spróbowałem dodać zmienną \"AdvertiserId\" do modelu z wykorzystaniem warstwy Embedding."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### a) Mały zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 379,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_maly_11 = D100k.iloc[0:80000, :] \n",
    "test_maly_11 = D100k.iloc[80000:, :]\n",
    "\n",
    "X_train_maly_11 = train_maly_11.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\", \"AdvertiserId\"]]\n",
    "Y_train_maly_11 = train_maly_11.loc[:, \"Click\"]\n",
    "                                        \n",
    "X_test_maly_11 = test_maly_11.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\", \"AdvertiserId\"]]\n",
    "Y_test_maly_11 = test_maly_11.loc[:, \"Click\"]  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 382,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Największa liczba odkreślająca identyfikator:\n",
      "39083\n"
     ]
    }
   ],
   "source": [
    "print(\"Największa liczba odkreślająca identyfikator:\")\n",
    "print(max(x for x in D100k.AdvertiserId))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 385,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "625/625 [==============================] - 2s 2ms/step - loss: 0.3241 - val_loss: 0.1889\n",
      "Epoch 2/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.2031 - val_loss: 0.1887\n",
      "Epoch 3/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1994 - val_loss: 0.1884\n",
      "Epoch 4/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.2008 - val_loss: 0.1886\n",
      "Epoch 5/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1990 - val_loss: 0.1885\n",
      "Epoch 6/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1984 - val_loss: 0.1886\n",
      "Epoch 7/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1976 - val_loss: 0.1884\n",
      "Epoch 8/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1981 - val_loss: 0.1883\n",
      "Epoch 9/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1999 - val_loss: 0.1885\n",
      "Epoch 10/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1989 - val_loss: 0.1883\n",
      "Epoch 11/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1989 - val_loss: 0.1883\n",
      "Epoch 12/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1990 - val_loss: 0.1880\n",
      "Epoch 13/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1986 - val_loss: 0.1879\n",
      "Epoch 14/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1995 - val_loss: 0.1879\n",
      "Epoch 15/15\n",
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1981 - val_loss: 0.1880\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x2920ec3a2b0>"
      ]
     },
     "execution_count": 385,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocab_size = 39083 + 1\n",
    "\n",
    "# Gdybym nie dodał 1 do największej liczby określającej identyfikator, to miałbym następujący komunikat:\n",
    "# InvalidArgumentError:  indices[67,4] = 39083 is not in [0, 39083)\n",
    "# \t [[node sequential_43/embedding_29/embedding_lookup (defined at <ipython-input-383-ac7e3b8f2c50>:13) ]] [Op:__inference_train_function_5399745]\n",
    "\n",
    "# Errors may have originated from an input operation.\n",
    "# Input Source operations connected to node sequential_43/embedding_29/embedding_lookup:\n",
    "#  sequential_43/embedding_29/embedding_lookup/5399179 (defined at C:\\Users\\Kuboz\\anaconda3\\lib\\contextlib.py:113)\n",
    "\n",
    "# Function call stack:\n",
    "# train_function\n",
    "\n",
    "model_maly_11 = tf.keras.models.Sequential()\n",
    "\n",
    "\n",
    "model_maly_11.add(Embedding(vocab_size,1))\n",
    "\n",
    "model_maly_11.add(SimpleRNN(64, dropout=0.2, recurrent_dropout=0.2, return_sequences=False))\n",
    "model_maly_11.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_maly_11.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "model_maly_11.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_maly_11.fit(X_train_maly_11, Y_train_maly_11,\n",
    "              batch_size=128,\n",
    "              epochs=15,\n",
    "              verbose=1,\n",
    "              validation_data=(X_test_maly_11, Y_test_maly_11))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 386,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_maly_11 = model_maly_11.predict(X_test_maly_11)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 387,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.0430811 ],\n",
       "       [0.03937787],\n",
       "       [0.04359093],\n",
       "       ...,\n",
       "       [0.04219833],\n",
       "       [0.04068637],\n",
       "       [0.04407543]], dtype=float32)"
      ]
     },
     "execution_count": 387,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_11"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 388,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20000, 1)"
      ]
     },
     "execution_count": 388,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_maly_11.shape\n",
    "# (20000, 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Mamy więc macierz wymiaru 20000 x 1. Wyciągamy wszystkie elementy z kolumny."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 389,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_maly_11 = pred_maly_11[:, 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 390,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5682481198979753"
      ]
     },
     "execution_count": 390,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sklearn.metrics.roc_auc_score(Y_test_maly_11, pred_maly_11)\n",
    "# 0.5682481198979753"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.568"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### b) Średni zbiór"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_sredni_11 = D10M.iloc[0:80000, :] \n",
    "test_sredni_11 = D10M.iloc[80000:, :]\n",
    "\n",
    "X_train_sredni_11 = train_sredni_11.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\", \"AdvertiserId\"]]\n",
    "Y_train_sredni_11 = train_sredni_11.loc[:, \"Click\"]\n",
    "                                        \n",
    "X_test_sredni_11 = test_sredni_11.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\", \"AdvertiserId\"]]\n",
    "Y_test_sredni_11 = test_sredni_11.loc[:, \"Click\"]  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Największa liczba odkreślająca identyfikator:\n",
      "39164\n"
     ]
    }
   ],
   "source": [
    "print(\"Największa liczba odkreślająca identyfikator:\")\n",
    "print(max(x for x in D10M.AdvertiserId))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "620/625 [============================>.] - ETA: 0s - loss: 0.2690"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-20-a3714f3161a5>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     10\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     11\u001b[0m \u001b[0mmodel_sredni_11\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcompile\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"sgd\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mloss\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;34m\"binary_crossentropy\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 12\u001b[1;33m model_sredni_11.fit(X_train_sredni_11, Y_train_sredni_11,\n\u001b[0m\u001b[0;32m     13\u001b[0m               \u001b[0mbatch_size\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m128\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     14\u001b[0m               \u001b[0mepochs\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m15\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\keras\\engine\\training.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[0;32m   1129\u001b[0m                 \u001b[0mmodel\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1130\u001b[0m                 steps_per_execution=self._steps_per_execution)\n\u001b[1;32m-> 1131\u001b[1;33m           val_logs = self.evaluate(\n\u001b[0m\u001b[0;32m   1132\u001b[0m               \u001b[0mx\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mval_x\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1133\u001b[0m               \u001b[0my\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mval_y\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\keras\\engine\\training.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(self, x, y, batch_size, verbose, sample_weight, steps, callbacks, max_queue_size, workers, use_multiprocessing, return_dict)\u001b[0m\n\u001b[0;32m   1387\u001b[0m             \u001b[1;32mwith\u001b[0m \u001b[0mtrace\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mTrace\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'test'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mstep_num\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mstep\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0m_r\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1388\u001b[0m               \u001b[0mcallbacks\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mon_test_batch_begin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1389\u001b[1;33m               \u001b[0mtmp_logs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtest_function\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1390\u001b[0m               \u001b[1;32mif\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshould_sync\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1391\u001b[0m                 \u001b[0mcontext\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0masync_wait\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\eager\\def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, *args, **kwds)\u001b[0m\n\u001b[0;32m    826\u001b[0m     \u001b[0mtracing_count\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    827\u001b[0m     \u001b[1;32mwith\u001b[0m \u001b[0mtrace\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mTrace\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_name\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0mtm\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 828\u001b[1;33m       \u001b[0mresult\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    829\u001b[0m       \u001b[0mcompiler\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;34m\"xla\"\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_experimental_compile\u001b[0m \u001b[1;32melse\u001b[0m \u001b[1;34m\"nonXla\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    830\u001b[0m       \u001b[0mnew_tracing_count\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\eager\\def_function.py\u001b[0m in \u001b[0;36m_call\u001b[1;34m(self, *args, **kwds)\u001b[0m\n\u001b[0;32m    860\u001b[0m       \u001b[1;31m# In this case we have not created variables on the first call. So we can\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    861\u001b[0m       \u001b[1;31m# run the first trace but we should fail if variables are created.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 862\u001b[1;33m       \u001b[0mresults\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    863\u001b[0m       \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_created_variables\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    864\u001b[0m         raise ValueError(\"Creating variables on a non-first call to a function\"\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\eager\\function.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, *args, **kwargs)\u001b[0m\n\u001b[0;32m   2940\u001b[0m       (graph_function,\n\u001b[0;32m   2941\u001b[0m        filtered_flat_args) = self._maybe_define_function(args, kwargs)\n\u001b[1;32m-> 2942\u001b[1;33m     return graph_function._call_flat(\n\u001b[0m\u001b[0;32m   2943\u001b[0m         filtered_flat_args, captured_inputs=graph_function.captured_inputs)  # pylint: disable=protected-access\n\u001b[0;32m   2944\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\eager\\function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[1;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[0;32m   1916\u001b[0m         and executing_eagerly):\n\u001b[0;32m   1917\u001b[0m       \u001b[1;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1918\u001b[1;33m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0m\u001b[0;32m   1919\u001b[0m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0;32m   1920\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\eager\\function.py\u001b[0m in \u001b[0;36mcall\u001b[1;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[0;32m    553\u001b[0m       \u001b[1;32mwith\u001b[0m \u001b[0m_InterpolateFunctionError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    554\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mcancellation_manager\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 555\u001b[1;33m           outputs = execute.execute(\n\u001b[0m\u001b[0;32m    556\u001b[0m               \u001b[0mstr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msignature\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mname\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    557\u001b[0m               \u001b[0mnum_outputs\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_num_outputs\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\tensorflow\\python\\eager\\execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[1;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[0;32m     57\u001b[0m   \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 59\u001b[1;33m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0m\u001b[0;32m     60\u001b[0m                                         inputs, attrs, num_outputs)\n\u001b[0;32m     61\u001b[0m   \u001b[1;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "vocab_size = 39164 + 1\n",
    "\n",
    "model_sredni_11 = tf.keras.models.Sequential()\n",
    "\n",
    "model_sredni_11.add(Embedding(vocab_size,1))\n",
    "\n",
    "model_sredni_11.add(SimpleRNN(64, dropout=0.2, recurrent_dropout=0.2, return_sequences=False))\n",
    "model_sredni_11.add(tf.keras.layers.Dense(3, activation = \"relu\"))\n",
    "model_sredni_11.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "model_sredni_11.compile(\"sgd\", loss = \"binary_crossentropy\")\n",
    "model_sredni_11.fit(X_train_sredni_11, Y_train_sredni_11,\n",
    "              batch_size=128,\n",
    "              epochs=15,\n",
    "              verbose=1,\n",
    "              validation_data=(X_test_sredni_11, Y_test_sredni_11))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_sredni_11 = model_sredni_11.predict(X_test_sredni_11)\n",
    "pred_sredni_11 = pred_sredni_11[:, 0]\n",
    "sklearn.metrics.roc_auc_score(Y_test_sredni_11, pred_sredni_11)\n",
    "# 0.560587675159211"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC = 0.561"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Stąd dodanie tej zmiennej indykatorowej, nie poprawiło, a nawet trochę pogorszyło AUC. W dodatku, dla zbioru D10M, dopasowywanie modelu trwało bardzo długo (ponad 1 minutę zajęło wykonanie pojedynczej epoki), więc w sumie z 15 epokami, całe dopasowanie zajęło 15 minut."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zatem ostatecznie wybieram model 10, tzn. model z wszystkimi małymi zmiennymi: \"Age\", \"Gender\", Depth\", \"Position\" trenowany na zbiorze D10M, z 3 warstwami z aktywacją za pomocą funkcji ReLU i 1 warstwą z funkcją sigmoidalną, gdzie kompilowałem model z wykorzystaniem klasyfikatora stochastycznego spadku gradientu SGD z binarną funkcją straty krzyżowej entropii, z liczbą epok równą 15."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Można też zauważyć, że dodanie zmiennych Depth i Position, znacząco poprawiły wartość AUC."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Predykcje na zbiorze testowym D5M_test_x.tsv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test= test.loc[:, [\"Depth\", \"Position\", \"Gender\", \"Age\"]]\n",
    "Y_test = test.loc[:, \"Click\"]       "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_test = model_sredni_10.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.05357158],\n",
       "       [0.05663043],\n",
       "       [0.07478765],\n",
       "       [0.02977902],\n",
       "       [0.02988666],\n",
       "       [0.06446198],\n",
       "       [0.02722976],\n",
       "       [0.07082167],\n",
       "       [0.01401973],\n",
       "       [0.05357158],\n",
       "       [0.03163835],\n",
       "       [0.05663043],\n",
       "       [0.07082167],\n",
       "       [0.03764346],\n",
       "       [0.03348911],\n",
       "       [0.03348911],\n",
       "       [0.03348911],\n",
       "       [0.03544426],\n",
       "       [0.02508664],\n",
       "       [0.07478765]], dtype=float32)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_test[:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5000000, 1)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_test = pred_test[:, 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.05357158, 0.05663043, 0.07478765, ..., 0.06739086, 0.03544426,\n",
       "       0.02695978], dtype=float32)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.04908551"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_test.mean()\n",
    "# 0.047198534"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### A więc wiemy, że średnio prawdopodobieństwo wynosi 0.047, czyli jest w miarę dobrze zbliżone do prawdopodobieństwa 0.05."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zapisuję swoje otrzymane wyniki do pliku .txt."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TO JEST ZŁE!!!!!!!!!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 431,
   "metadata": {},
   "outputs": [],
   "source": [
    "plik = open(\"Predykcje_JZ.txt\", \"w\")\n",
    "plik.write(\"Jakub_Zbrzezny\" + \"\\n\")\n",
    "for i in pred_test:\n",
    "    plik.write(str(i) + \"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TO JUŻ JEST DOBRE!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "lista = []\n",
    "for i in range(len(pred_test)):\n",
    "    lista.append(float(pred_test[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"poprawione_pred_JZ.txt\", \"w\") as f:\n",
    "    for item in lista:\n",
    "        f.write(\"%s\\n\" % item)"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}