recommendation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# automatically reload edited modules\n",
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline\n",
    "\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "DATASET_DIR = './dataset/'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Data Load"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "users_cols = ['user_id', 'age', 'sex', 'occupation', 'zip_code']\n",
    "movie_cols = ['movie_id', 'movie_title', 'release_date', 'video_release_date',\n",
    "              'IMDb_URL', 'unknown', 'Action', 'Adventure', 'Animation',\n",
    "              'Children\\'s', 'Comedy', 'Crime', 'Documentary', 'Drama', 'Fantasy',\n",
    "              'Film-Noir', 'Horror', 'Musical', 'Mystery', 'Romance', 'Sci-Fi',\n",
    "              'Thriller', 'War', 'Western']\n",
    "rating_cols = ['user_id', 'movie_id', 'rating', 'timestamp']\n",
    "\n",
    "#------------------\n",
    "user_data = pd.read_csv(DATASET_DIR+'u.user', sep='|', names=users_cols)\n",
    "movie_data = pd.read_csv(DATASET_DIR+'u.item', sep='|', names=movie_cols, encoding='latin')\n",
    "\n",
    "#------------------\n",
    "# train\n",
    "rating_data = pd.read_csv(DATASET_DIR+'u1.base', sep='\\t', names=rating_cols)\n",
    "# test\n",
    "test_rating_data = pd.read_csv(DATASET_DIR+'u1.test', sep='\\t', names=rating_cols)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Movie Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_title</th>\n",
       "      <th>release_date</th>\n",
       "      <th>video_release_date</th>\n",
       "      <th>IMDb_URL</th>\n",
       "      <th>unknown</th>\n",
       "      <th>Action</th>\n",
       "      <th>Adventure</th>\n",
       "      <th>Animation</th>\n",
       "      <th>Children's</th>\n",
       "      <th>...</th>\n",
       "      <th>Fantasy</th>\n",
       "      <th>Film-Noir</th>\n",
       "      <th>Horror</th>\n",
       "      <th>Musical</th>\n",
       "      <th>Mystery</th>\n",
       "      <th>Romance</th>\n",
       "      <th>Sci-Fi</th>\n",
       "      <th>Thriller</th>\n",
       "      <th>War</th>\n",
       "      <th>Western</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Toy Story (1995)</td>\n",
       "      <td>01-Jan-1995</td>\n",
       "      <td>NaN</td>\n",
       "      <td>http://us.imdb.com/M/title-exact?Toy%20Story%2...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>GoldenEye (1995)</td>\n",
       "      <td>01-Jan-1995</td>\n",
       "      <td>NaN</td>\n",
       "      <td>http://us.imdb.com/M/title-exact?GoldenEye%20(...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Four Rooms (1995)</td>\n",
       "      <td>01-Jan-1995</td>\n",
       "      <td>NaN</td>\n",
       "      <td>http://us.imdb.com/M/title-exact?Four%20Rooms%...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Get Shorty (1995)</td>\n",
       "      <td>01-Jan-1995</td>\n",
       "      <td>NaN</td>\n",
       "      <td>http://us.imdb.com/M/title-exact?Get%20Shorty%...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Copycat (1995)</td>\n",
       "      <td>01-Jan-1995</td>\n",
       "      <td>NaN</td>\n",
       "      <td>http://us.imdb.com/M/title-exact?Copycat%20(1995)</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   movie_id        movie_title release_date  video_release_date  \\\n",
       "0         1   Toy Story (1995)  01-Jan-1995                 NaN   \n",
       "1         2   GoldenEye (1995)  01-Jan-1995                 NaN   \n",
       "2         3  Four Rooms (1995)  01-Jan-1995                 NaN   \n",
       "3         4  Get Shorty (1995)  01-Jan-1995                 NaN   \n",
       "4         5     Copycat (1995)  01-Jan-1995                 NaN   \n",
       "\n",
       "                                            IMDb_URL  unknown  Action  \\\n",
       "0  http://us.imdb.com/M/title-exact?Toy%20Story%2...        0       0   \n",
       "1  http://us.imdb.com/M/title-exact?GoldenEye%20(...        0       1   \n",
       "2  http://us.imdb.com/M/title-exact?Four%20Rooms%...        0       0   \n",
       "3  http://us.imdb.com/M/title-exact?Get%20Shorty%...        0       1   \n",
       "4  http://us.imdb.com/M/title-exact?Copycat%20(1995)        0       0   \n",
       "\n",
       "   Adventure  Animation  Children's  ...  Fantasy  Film-Noir  Horror  Musical  \\\n",
       "0          0          1           1  ...        0          0       0        0   \n",
       "1          1          0           0  ...        0          0       0        0   \n",
       "2          0          0           0  ...        0          0       0        0   \n",
       "3          0          0           0  ...        0          0       0        0   \n",
       "4          0          0           0  ...        0          0       0        0   \n",
       "\n",
       "   Mystery  Romance  Sci-Fi  Thriller  War  Western  \n",
       "0        0        0       0         0    0        0  \n",
       "1        0        0       0         1    0        0  \n",
       "2        0        0       0         1    0        0  \n",
       "3        0        0       0         0    0        0  \n",
       "4        0        0       0         1    0        0  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movie_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### User Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>user_id</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>occupation</th>\n",
       "      <th>zip_code</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>technician</td>\n",
       "      <td>85711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>53</td>\n",
       "      <td>F</td>\n",
       "      <td>other</td>\n",
       "      <td>94043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>23</td>\n",
       "      <td>M</td>\n",
       "      <td>writer</td>\n",
       "      <td>32067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>technician</td>\n",
       "      <td>43537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>33</td>\n",
       "      <td>F</td>\n",
       "      <td>other</td>\n",
       "      <td>15213</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id  age sex  occupation zip_code\n",
       "0        1   24   M  technician    85711\n",
       "1        2   53   F       other    94043\n",
       "2        3   23   M      writer    32067\n",
       "3        4   24   M  technician    43537\n",
       "4        5   33   F       other    15213"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Rating Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>874965758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>876893171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>878542960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>876893119</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>889751712</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id  movie_id  rating  timestamp\n",
       "0        1         1       5  874965758\n",
       "1        1         2       3  876893171\n",
       "2        1         3       4  878542960\n",
       "3        1         4       3  876893119\n",
       "4        1         5       3  889751712"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rating_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>user_id</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>5</td>\n",
       "      <td>887431973</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>875693118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "      <td>5</td>\n",
       "      <td>878542960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>5</td>\n",
       "      <td>874965706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>17</td>\n",
       "      <td>3</td>\n",
       "      <td>875073198</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id  movie_id  rating  timestamp\n",
       "0        1         6       5  887431973\n",
       "1        1        10       3  875693118\n",
       "2        1        12       5  878542960\n",
       "3        1        14       5  874965706\n",
       "4        1        17       3  875073198"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_rating_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "# 모델에 입력할 Input Data 만들기\n",
    "   \n",
    "- data_per_user   \n",
    "유저당 사용할 Target 데이터 갯수   \n",
    "\n",
    "- wathed_data_size   \n",
    "입력에 사용할 유저가 본 영화 갯수 \n",
    "\n",
    "- negative_size   \n",
    "입력에 사용할 negative_sample 갯수"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>user_id</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>watched_movies</th>\n",
       "      <th>negative_samples</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>109</td>\n",
       "      <td>5</td>\n",
       "      <td>[57, 137, 63, 165, 94, 35, 205, 71, 75, 43]</td>\n",
       "      <td>[1524, 359, 426, 1022, 449]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>58</td>\n",
       "      <td>4</td>\n",
       "      <td>[195, 165, 79, 149, 46, 75, 237, 172, 18, 127]</td>\n",
       "      <td>[460, 678, 1329, 1445, 631]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>[147, 22, 119, 220, 116, 136, 141, 167, 168, 239]</td>\n",
       "      <td>[997, 1177, 1421, 910, 32]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>216</td>\n",
       "      <td>5</td>\n",
       "      <td>[138, 251, 181, 144, 4, 3, 162, 244, 119, 94]</td>\n",
       "      <td>[315, 1321, 1291, 1134, 1527]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>M</td>\n",
       "      <td>181</td>\n",
       "      <td>5</td>\n",
       "      <td>[26, 48, 18, 63, 139, 55, 40, 109, 199, 217]</td>\n",
       "      <td>[907, 1471, 25, 1276, 652]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>943</td>\n",
       "      <td>22</td>\n",
       "      <td>M</td>\n",
       "      <td>98</td>\n",
       "      <td>5</td>\n",
       "      <td>[2, 232, 443, 831, 318, 186, 68, 427, 12, 739]</td>\n",
       "      <td>[757, 307, 885, 170, 1564]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>943</td>\n",
       "      <td>22</td>\n",
       "      <td>M</td>\n",
       "      <td>427</td>\n",
       "      <td>4</td>\n",
       "      <td>[121, 182, 24, 111, 161, 151, 526, 419, 117, 403]</td>\n",
       "      <td>[18, 162, 619, 1428, 910]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>943</td>\n",
       "      <td>22</td>\n",
       "      <td>M</td>\n",
       "      <td>94</td>\n",
       "      <td>4</td>\n",
       "      <td>[96, 281, 831, 415, 763, 427, 124, 1011, 785, 22]</td>\n",
       "      <td>[299, 1328, 527, 1382, 616]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>943</td>\n",
       "      <td>22</td>\n",
       "      <td>M</td>\n",
       "      <td>928</td>\n",
       "      <td>5</td>\n",
       "      <td>[282, 194, 475, 2, 31, 174, 41, 431, 232, 53]</td>\n",
       "      <td>[664, 642, 491, 805, 1479]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>943</td>\n",
       "      <td>22</td>\n",
       "      <td>M</td>\n",
       "      <td>232</td>\n",
       "      <td>4</td>\n",
       "      <td>[785, 117, 356, 80, 195, 9, 569, 68, 121, 72]</td>\n",
       "      <td>[896, 1484, 46, 1064, 738]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17601 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    user_id  age sex  movie_id  rating  \\\n",
       "0         1   24   M       109       5   \n",
       "1         1   24   M        58       4   \n",
       "2         1   24   M         5       3   \n",
       "3         1   24   M       216       5   \n",
       "4         1   24   M       181       5   \n",
       "..      ...  ...  ..       ...     ...   \n",
       "15      943   22   M        98       5   \n",
       "16      943   22   M       427       4   \n",
       "17      943   22   M        94       4   \n",
       "18      943   22   M       928       5   \n",
       "19      943   22   M       232       4   \n",
       "\n",
       "                                       watched_movies  \\\n",
       "0         [57, 137, 63, 165, 94, 35, 205, 71, 75, 43]   \n",
       "1      [195, 165, 79, 149, 46, 75, 237, 172, 18, 127]   \n",
       "2   [147, 22, 119, 220, 116, 136, 141, 167, 168, 239]   \n",
       "3       [138, 251, 181, 144, 4, 3, 162, 244, 119, 94]   \n",
       "4        [26, 48, 18, 63, 139, 55, 40, 109, 199, 217]   \n",
       "..                                                ...   \n",
       "15     [2, 232, 443, 831, 318, 186, 68, 427, 12, 739]   \n",
       "16  [121, 182, 24, 111, 161, 151, 526, 419, 117, 403]   \n",
       "17  [96, 281, 831, 415, 763, 427, 124, 1011, 785, 22]   \n",
       "18      [282, 194, 475, 2, 31, 174, 41, 431, 232, 53]   \n",
       "19      [785, 117, 356, 80, 195, 9, 569, 68, 121, 72]   \n",
       "\n",
       "                 negative_samples  \n",
       "0     [1524, 359, 426, 1022, 449]  \n",
       "1     [460, 678, 1329, 1445, 631]  \n",
       "2      [997, 1177, 1421, 910, 32]  \n",
       "3   [315, 1321, 1291, 1134, 1527]  \n",
       "4      [907, 1471, 25, 1276, 652]  \n",
       "..                            ...  \n",
       "15     [757, 307, 885, 170, 1564]  \n",
       "16      [18, 162, 619, 1428, 910]  \n",
       "17    [299, 1328, 527, 1382, 616]  \n",
       "18     [664, 642, 491, 805, 1479]  \n",
       "19     [896, 1484, 46, 1064, 738]  \n",
       "\n",
       "[17601 rows x 7 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_per_user = 20\n",
    "wathed_data_size = 10\n",
    "negative_size = 5 # big: 2-5 / small: 5-20\n",
    "\n",
    "# -------------------------\n",
    "'''\n",
    "marge user data and rating data.\n",
    "user 데이터와 rating 데이터 병합.\n",
    "\n",
    "Since it is a system that recommends movies for users, movies with scores of less than 3 stars are excluded.\n",
    "유저에게 적합한 영화를 추천하는 시스템 이므로 3점 미만 점수를 준 영화는 제외한다.\n",
    "'''\n",
    "train_dataset = pd.merge(user_data[['user_id','age','sex']], rating_data[['user_id','movie_id','rating']][rating_data.rating >= 3], on='user_id')\n",
    "\n",
    "'''\n",
    "Select sample as many as 'data_per_user', delete user less than 'data_per_user'.\n",
    "'data_per_user' 만큼 선택, 'data_per_user' 미만 개수 user 삭제.\n",
    "'''\n",
    "train_dataset = train_dataset.groupby('user_id', group_keys=False).apply(\n",
    "    lambda x : x.sample(n=data_per_user).reset_index(drop=True)\n",
    "    if len(x) >= data_per_user\n",
    "    else x.reset_index(drop=True))\n",
    "\n",
    "'''\n",
    "Select watched movies as many as 'wathed_data_size'.\n",
    "본 영화를 'wathed_data_size' 만큼 선택.\n",
    "'''\n",
    "train_dataset['watched_movies'] = train_dataset['user_id'].apply(\n",
    "    lambda x : rating_data[rating_data.user_id == x]['movie_id'].sample(n=wathed_data_size).tolist()\n",
    "    if len(rating_data[rating_data.user_id == x]['movie_id'])>= wathed_data_size\n",
    "    else rating_data[rating_data.user_id == x]['movie_id'].tolist())\n",
    "    #else rating_data[rating_data.user_id == x]['movie_id'].sample(n=wathed_data_size, replace=True).tolist())\n",
    "\n",
    "'''\n",
    "negative sampling\n",
    "\n",
    "!!! It's not an accurate negative sampling because it's just random, needs to be modified !!! \n",
    "!!! 단순히 랜덤으로 뽑았으므로 정확한 negative sampling 아님, 수정 필요 !!!\n",
    "'''\n",
    "train_dataset['negative_samples'] = np.random.choice(len(movie_data)+1, (len(train_dataset),negative_size)).tolist() # +1 : 마지막 영화 ID도 포함\n",
    "\n",
    "train_dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>user_id</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>watched_movies</th>\n",
       "      <th>negative_samples</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>440</td>\n",
       "      <td>30</td>\n",
       "      <td>M</td>\n",
       "      <td>283</td>\n",
       "      <td>5</td>\n",
       "      <td>[751, 213, 921, 86, 988, 329, 300, 313, 283, 258]</td>\n",
       "      <td>[968, 1302, 57, 766, 1255]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>622</td>\n",
       "      <td>25</td>\n",
       "      <td>M</td>\n",
       "      <td>386</td>\n",
       "      <td>3</td>\n",
       "      <td>[1039, 169, 1203, 250, 69, 542, 1216, 730, 373...</td>\n",
       "      <td>[244, 305, 635, 406, 501]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>590</td>\n",
       "      <td>50</td>\n",
       "      <td>M</td>\n",
       "      <td>150</td>\n",
       "      <td>5</td>\n",
       "      <td>[1129, 275, 547, 740, 1331, 1009, 13, 754, 15,...</td>\n",
       "      <td>[15, 941, 1479, 1121, 490]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>354</td>\n",
       "      <td>29</td>\n",
       "      <td>F</td>\n",
       "      <td>286</td>\n",
       "      <td>4</td>\n",
       "      <td>[89, 65, 1039, 638, 154, 896, 181, 10, 811, 1017]</td>\n",
       "      <td>[892, 981, 470, 205, 847]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>37</td>\n",
       "      <td>23</td>\n",
       "      <td>M</td>\n",
       "      <td>841</td>\n",
       "      <td>3</td>\n",
       "      <td>[121, 210, 68, 833, 96, 568, 665, 230, 127, 578]</td>\n",
       "      <td>[1212, 395, 504, 1520, 1090]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id  age sex  movie_id  rating  \\\n",
       "0      440   30   M       283       5   \n",
       "1      622   25   M       386       3   \n",
       "2      590   50   M       150       5   \n",
       "3      354   29   F       286       4   \n",
       "4       37   23   M       841       3   \n",
       "\n",
       "                                      watched_movies  \\\n",
       "0  [751, 213, 921, 86, 988, 329, 300, 313, 283, 258]   \n",
       "1  [1039, 169, 1203, 250, 69, 542, 1216, 730, 373...   \n",
       "2  [1129, 275, 547, 740, 1331, 1009, 13, 754, 15,...   \n",
       "3  [89, 65, 1039, 638, 154, 896, 181, 10, 811, 1017]   \n",
       "4   [121, 210, 68, 833, 96, 568, 665, 230, 127, 578]   \n",
       "\n",
       "               negative_samples  \n",
       "0    [968, 1302, 57, 766, 1255]  \n",
       "1     [244, 305, 635, 406, 501]  \n",
       "2    [15, 941, 1479, 1121, 490]  \n",
       "3     [892, 981, 470, 205, 847]  \n",
       "4  [1212, 395, 504, 1520, 1090]  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Shuffle\n",
    "# 데이터 뒤섞기\n",
    "train_dataset = train_dataset.sample(frac=1).reset_index(drop=True)\n",
    "train_dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(17601, 6) (17601, 10) (17601, 1) (17601, 1)\n"
     ]
    }
   ],
   "source": [
    "# Dataframe 형태에서 모델 입력에 맞게 array 형태로 변환\n",
    "temp_1 = train_dataset['movie_id'].to_numpy().reshape(-1,1)\n",
    "temp_2 = np.array(train_dataset['negative_samples'].tolist())\n",
    "sample_data = np.append(temp_1, temp_2, axis=-1) # -> [Target, Negative_samples]\n",
    "\n",
    "#watched_movies_data = np.array(train_dataset['watched_movies'].tolist())\n",
    "watched_movies_data = tf.keras.preprocessing.sequence.pad_sequences(train_dataset['watched_movies'])\n",
    "\n",
    "gender_data = train_dataset['sex'].apply(lambda x: 0 if x == 'M' else 1).to_numpy().reshape(-1,1)\n",
    "\n",
    "age_data = train_dataset['age'].apply(lambda x: (x - train_dataset['age'].min())/(train_dataset['age'].max() - train_dataset['age'].min())).to_numpy().reshape(-1,1)\n",
    "\n",
    "print(sample_data.shape, watched_movies_data.shape, gender_data.shape, age_data.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "# Model 구성\n",
    "\n",
    "#### Full Model\n",
    "![full_model](./image/full_model.png)\n",
    "   \n",
    "#### Candidate Model\n",
    "![candidate_model](image/candidate_model.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " watched_movies (InputLayer)    [(None, None)]       0           []                               \n",
      "                                                                                                  \n",
      " features_embedding (Embedding)  (None, None, 16)    26928       ['watched_movies[0][0]',         \n",
      "                                                                  'samples[0][0]']                \n",
      "                                                                                                  \n",
      " features_embedding_average (Av  (None, 16)          0           ['features_embedding[0][0]']     \n",
      " g_Embedding)                                                                                     \n",
      "                                                                                                  \n",
      " age (InputLayer)               [(None, 1)]          0           []                               \n",
      "                                                                                                  \n",
      " gender (InputLayer)            [(None, 1)]          0           []                               \n",
      "                                                                                                  \n",
      " concatenate_features (Concaten  (None, 18)          0           ['features_embedding_average[0][0\n",
      " ate)                                                            ]',                              \n",
      "                                                                  'age[0][0]',                    \n",
      "                                                                  'gender[0][0]']                 \n",
      "                                                                                                  \n",
      " dense_1 (Dense)                (None, 64)           1216        ['concatenate_features[0][0]']   \n",
      "                                                                                                  \n",
      " samples (InputLayer)           [(None, None)]       0           []                               \n",
      "                                                                                                  \n",
      " dense_2 (Dense)                (None, 16)           1040        ['dense_1[0][0]']                \n",
      "                                                                                                  \n",
      " dot_product (Dot)              (None, None)         0           ['dense_2[0][0]',                \n",
      "                                                                  'features_embedding[1][0]']     \n",
      "                                                                                                  \n",
      " class_probabilities (Activatio  (None, None)        0           ['dot_product[0][0]']            \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 29,184\n",
      "Trainable params: 29,184\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "EMBEDDING_DIMS = 16 # 임베딩 크기\n",
    "DENSE_UNITS = 64    # 은닉층 크기\n",
    "MOVIE_NUM = movie_data['movie_id'].max() #1682\n",
    "LEARNING_RATE = 0.003\n",
    "BATCH_SIZE = 10\n",
    "\n",
    "#-------------------------#\n",
    "# embedding average layer\n",
    "class Avg_Embedding(tf.keras.layers.Layer):\n",
    "    def __init__(self, **kwargs):\n",
    "        super(Avg_Embedding, self).__init__(**kwargs)\n",
    "\n",
    "    def call(self, input):\n",
    "        return tf.reduce_mean(input, axis=1)\n",
    "    \n",
    "#-------------------------#\n",
    "# Input layers\n",
    "input_watched_movies = tf.keras.Input(shape=(None, ), name='watched_movies')\n",
    "input_age = tf.keras.layers.Input(shape=(1), name='age')\n",
    "input_gender = tf.keras.layers.Input(shape=(1), name='gender')\n",
    "input_samples = tf.keras.Input(shape=(None, ), name='samples')\n",
    "\n",
    "# Embedding layers (영화 기록을 임베딩 하기 위한 embedding layer, 검색 기록은 없으니 생략)\n",
    "features_embedding_layer = tf.keras.layers.Embedding(input_dim = MOVIE_NUM+1, output_dim = EMBEDDING_DIMS, mask_zero=True, name='features_embedding')\n",
    "average_embedding_layer = Avg_Embedding(name='features_embedding_average')\n",
    "\n",
    "# Dense layers\n",
    "dense_1 = tf.keras.layers.Dense(DENSE_UNITS, activation='relu', name='dense_1')\n",
    "dense_2 = tf.keras.layers.Dense(EMBEDDING_DIMS, activation='relu', name='dense_2')\n",
    "\n",
    "#-------------------------#\n",
    "# Model connection\n",
    "watched_movies_embedding = features_embedding_layer(input_watched_movies)\n",
    "sample_movies_embedding = features_embedding_layer(input_samples)\n",
    "\n",
    "average_embedding = average_embedding_layer(watched_movies_embedding)\n",
    "concat_features = tf.keras.layers.concatenate([average_embedding, input_age, input_gender], axis=1, name='concatenate_features')\n",
    "dense_1_out = dense_1(concat_features)\n",
    "dense_2_out = dense_2(dense_1_out)\n",
    "# \n",
    "dot_product = tf.keras.layers.dot([dense_2_out, sample_movies_embedding], axes=(1,2), name='dot_product')\n",
    "output = tf.keras.layers.Activation('softmax', name = 'class_probabilities')(dot_product)\n",
    "\n",
    "model = tf.keras.Model(inputs=[input_watched_movies, input_age, input_gender, input_samples], outputs=[output])\n",
    "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=LEARNING_RATE), loss='sparse_categorical_crossentropy', metrics=['acc'])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "# Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-06-06 19:13:10.784867: I tensorflow/stream_executor/cuda/cuda_blas.cc:1786] TensorFloat-32 will be used for the matrix multiplication. This will only be logged once.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1761/1761 [==============================] - 9s 2ms/step - loss: 1.1830 - acc: 0.5320\n",
      "Epoch 2/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.9679 - acc: 0.6053\n",
      "Epoch 3/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.8583 - acc: 0.6466\n",
      "Epoch 4/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.7966 - acc: 0.6652\n",
      "Epoch 5/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.7437 - acc: 0.6851\n",
      "Epoch 6/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.6922 - acc: 0.7073\n",
      "Epoch 7/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.6428 - acc: 0.7262\n",
      "Epoch 8/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.5934 - acc: 0.7472\n",
      "Epoch 9/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.5473 - acc: 0.7654\n",
      "Epoch 10/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.5023 - acc: 0.7914\n",
      "Epoch 11/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.4628 - acc: 0.8063\n",
      "Epoch 12/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.4237 - acc: 0.8267\n",
      "Epoch 13/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.3907 - acc: 0.8408\n",
      "Epoch 14/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.3600 - acc: 0.8547\n",
      "Epoch 15/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.3309 - acc: 0.8664\n",
      "Epoch 16/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.3068 - acc: 0.8766\n",
      "Epoch 17/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.2831 - acc: 0.8847\n",
      "Epoch 18/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.2625 - acc: 0.8941\n",
      "Epoch 19/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.2416 - acc: 0.9035\n",
      "Epoch 20/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.2250 - acc: 0.9127\n",
      "Epoch 21/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.2077 - acc: 0.9177\n",
      "Epoch 22/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.2051 - acc: 0.9220\n",
      "Epoch 23/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1794 - acc: 0.9301\n",
      "Epoch 24/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1681 - acc: 0.9350\n",
      "Epoch 25/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1599 - acc: 0.9389\n",
      "Epoch 26/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1509 - acc: 0.9422\n",
      "Epoch 27/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1389 - acc: 0.9470\n",
      "Epoch 28/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1399 - acc: 0.9474\n",
      "Epoch 29/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1285 - acc: 0.9517\n",
      "Epoch 30/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1190 - acc: 0.9565\n",
      "Epoch 31/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1163 - acc: 0.9576\n",
      "Epoch 32/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1072 - acc: 0.9610\n",
      "Epoch 33/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1096 - acc: 0.9606\n",
      "Epoch 34/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.1114 - acc: 0.9606\n",
      "Epoch 35/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0940 - acc: 0.9665\n",
      "Epoch 36/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0941 - acc: 0.9657\n",
      "Epoch 37/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0875 - acc: 0.9690\n",
      "Epoch 38/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0945 - acc: 0.9664\n",
      "Epoch 39/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0845 - acc: 0.9700\n",
      "Epoch 40/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0782 - acc: 0.9731\n",
      "Epoch 41/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0848 - acc: 0.9724\n",
      "Epoch 42/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0892 - acc: 0.9701\n",
      "Epoch 43/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0808 - acc: 0.9723\n",
      "Epoch 44/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0749 - acc: 0.9755\n",
      "Epoch 45/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0708 - acc: 0.9765\n",
      "Epoch 46/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0728 - acc: 0.9760\n",
      "Epoch 47/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0642 - acc: 0.9781\n",
      "Epoch 48/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0797 - acc: 0.9743\n",
      "Epoch 49/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0741 - acc: 0.9753\n",
      "Epoch 50/50\n",
      "1761/1761 [==============================] - 4s 2ms/step - loss: 0.0656 - acc: 0.9788\n"
     ]
    }
   ],
   "source": [
    "history = model.fit([watched_movies_data, age_data, gender_data, sample_data],\n",
    "                    np.zeros(len(train_dataset),dtype=int),\n",
    "                    batch_size=BATCH_SIZE,\n",
    "                    epochs=50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "# Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1683, 16)\n"
     ]
    }
   ],
   "source": [
    "# 유저 id를 입력하면 입력 데이터를 만들어주는 함수\n",
    "def pick_user_data_input(user_id, watched_sample_num = wathed_data_size):\n",
    "    if len(rating_data[rating_data.user_id == user_id]) >= watched_sample_num:\n",
    "        watched_movies_data = rating_data[rating_data.user_id == user_id]['movie_id'].sample(n=watched_sample_num).to_numpy().reshape(-1,watched_sample_num)\n",
    "    else:\n",
    "        #watched_movies_data = rating_data[rating_data.user_id == user_id]['movie_id'].sample(n=watched_sample_num, replace=True).to_numpy().reshape(-1,watched_sample_num)\n",
    "        watched_movies_data = tf.keras.preprocessing.sequence.pad_sequences(rating_data[rating_data.user_id == user_id]['movie_id'].to_numpy().reshape(1,-1), maxlen=watched_sample_num)\n",
    "\n",
    "    age_data = user_data[user_data.user_id == user_id]['age'].apply(lambda x: (x - user_data['age'].min())/(user_data['age'].max() - user_data['age'].min())).to_numpy().reshape(-1,1)\n",
    "    gender_data = user_data[user_data.user_id == user_id]['sex'].apply(lambda x : 0 if x=='M' else 1).to_numpy().reshape(-1,1)\n",
    "\n",
    "    return [watched_movies_data, age_data, gender_data]\n",
    "\n",
    "# 유저 임베딩 vector\n",
    "get_output = tf.keras.backend.function([model.get_layer('watched_movies').input,\n",
    "                                        model.get_layer('age').input,\n",
    "                                        model.get_layer('gender').input], [model.layers[-3].output])\n",
    "\n",
    "# 영화 임베딩 vector\n",
    "embedding_table = model.get_layer('features_embedding').get_weights()[0]\n",
    "print(embedding_table.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Movie Recommendation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Movie Recommendation for user 1 (movie_ID): [53, 1064, 1111, 510, 1488, 239, 230, 1157, 281, 162]\n"
     ]
    }
   ],
   "source": [
    "pick = 1 # 유저 ID\n",
    "top_N_size = 50 # 최대 50개 추천\n",
    "\n",
    "a = pick_user_data_input(pick)\n",
    "pred = get_output(a)[0]\n",
    "result = np.dot(pred, embedding_table.transpose())[0]\n",
    "rank = sorted(range(len(result)), key=lambda k: result[k], reverse=True) # 영화 추천 순위\n",
    "\n",
    "print(f\"Movie Recommendation for user {pick} (movie_ID): {rank[:10]}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### mean average precision\n",
    "\n",
    "`model prediction VS Random pick`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "top N  = 1 -------------------\n",
      "\n",
      "model  = Rank 1 Recall    : 0.0072992700729927005 (1/137)\n",
      "model  = Rank 1 Precision : 1.0 (1/1)\n",
      "\n",
      "random = Rank 1 Recall    : 0.0 (0/137)\n",
      "random = Rank 1 Precision : 0.0 (0/1)\n",
      "\n",
      "top N  = 10 -------------------\n",
      "\n",
      "model  = Rank 10 Recall    : 0.014598540145985401 (2/137)\n",
      "model  = Rank 10 Precision : 0.2 (2/10)\n",
      "\n",
      "random = Rank 10 Recall    : 0.0 (0/137)\n",
      "random = Rank 10 Precision : 0.0 (0/10)\n",
      "\n",
      "top N  = 20 -------------------\n",
      "\n",
      "model  = Rank 20 Recall    : 0.014598540145985401 (2/137)\n",
      "model  = Rank 20 Precision : 0.1 (2/20)\n",
      "\n",
      "random = Rank 20 Recall    : 0.0072992700729927005 (1/137)\n",
      "random = Rank 20 Precision : 0.05 (1/20)\n",
      "\n",
      "top N  = 30 -------------------\n",
      "\n",
      "model  = Rank 30 Recall    : 0.029197080291970802 (4/137)\n",
      "model  = Rank 30 Precision : 0.13333333333333333 (4/30)\n",
      "\n",
      "random = Rank 30 Recall    : 0.0364963503649635 (5/137)\n",
      "random = Rank 30 Precision : 0.16666666666666666 (5/30)\n",
      "\n",
      "top N  = 40 -------------------\n",
      "\n",
      "model  = Rank 40 Recall    : 0.058394160583941604 (8/137)\n",
      "model  = Rank 40 Precision : 0.2 (8/40)\n",
      "\n",
      "random = Rank 40 Recall    : 0.014598540145985401 (2/137)\n",
      "random = Rank 40 Precision : 0.05 (2/40)\n",
      "\n",
      "top N  = 50 -------------------\n",
      "\n",
      "model  = Rank 50 Recall    : 0.06569343065693431 (9/137)\n",
      "model  = Rank 50 Precision : 0.18 (9/50)\n",
      "\n",
      "random = Rank 50 Recall    : 0.014598540145985401 (2/137)\n",
      "random = Rank 50 Precision : 0.04 (2/50)\n",
      "\n",
      "mAP@50 = 0.19529945953383454\n"
     ]
    }
   ],
   "source": [
    "# mean average precision 측정\n",
    "\n",
    "mAP_N = 0\n",
    "display_freq = 10\n",
    "\n",
    "for top_N in range(1, top_N_size+1):\n",
    "    True_positive = test_rating_data[(test_rating_data.user_id == pick) & (test_rating_data.movie_id.isin(rank[:top_N]))]\n",
    "    positive = test_rating_data[test_rating_data.user_id == pick]\n",
    "\n",
    "    random_pick = np.random.choice(len(movie_data), top_N)\n",
    "    Random_positive = test_rating_data[(test_rating_data.user_id == pick) & (test_rating_data.movie_id.isin(random_pick))]\n",
    "    \n",
    "    if top_N % display_freq == 0 or top_N == 1:\n",
    "        print(f'top N  = {top_N} -------------------\\n')\n",
    "\n",
    "        print(f'model  = Rank {top_N} Recall    : {len(True_positive)/len(positive)} ({len(True_positive)}/{len(positive)})')\n",
    "        print(f'model  = Rank {top_N} Precision : {len(True_positive)/top_N} ({len(True_positive)}/{top_N})\\n')\n",
    "\n",
    "\n",
    "        print(f'random = Rank {top_N} Recall    : {len(Random_positive)/len(positive)} ({len(Random_positive)}/{len(positive)})')\n",
    "        print(f'random = Rank {top_N} Precision : {len(Random_positive)/top_N} ({len(Random_positive)}/{top_N})\\n')\n",
    "    \n",
    "    mAP_N += len(True_positive)/top_N\n",
    "    \n",
    "print(f'mAP@{top_N_size} = {mAP_N/top_N_size}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 모든 user에 대해 MAP(mean average precision)의 평균 측정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AVG_mAP@50 = 0.14774015325555956 (50/462)\n",
      "AVG_mAP@50 = 0.14651703451085207 (100/462)\n",
      "AVG_mAP@50 = 0.14080320414689043 (150/462)\n",
      "AVG_mAP@50 = 0.1386486893241169 (200/462)\n",
      "AVG_mAP@50 = 0.13491874545034896 (250/462)\n",
      "AVG_mAP@50 = 0.13819096102581413 (300/462)\n",
      "AVG_mAP@50 = 0.13818576585674347 (350/462)\n",
      "AVG_mAP@50 = 0.13215568231662916 (400/462)\n",
      "AVG_mAP@50 = 0.12283793241821864 (450/462)\n",
      "최종 : AVG_mAP@50 = 0.12002009124594983\n"
     ]
    }
   ],
   "source": [
    "top_N_size = 50\n",
    "display_freq = 50\n",
    "AVG_mAP = 0\n",
    "\n",
    "for pick in range(1, test_rating_data.user_id.max()+1):\n",
    "    mAP_N = 0\n",
    "    pred = get_output(pick_user_data_input(pick))[0]\n",
    "    result = np.dot(pred, embedding_table.transpose())[0]\n",
    "    rank = sorted(range(len(result)), key=lambda k: result[k], reverse=True)\n",
    "    \n",
    "    for top_N in range(1, top_N_size+1):\n",
    "        True_positive = test_rating_data[(test_rating_data.user_id == pick) & (test_rating_data.movie_id.isin(rank[:top_N]))]\n",
    "        positive = test_rating_data[test_rating_data.user_id == pick]\n",
    "\n",
    "        mAP_N += len(True_positive)/top_N\n",
    "        \n",
    "    AVG_mAP += mAP_N/top_N_size\n",
    "    if pick % display_freq == 0 : print(f'AVG_mAP@{top_N_size} = {AVG_mAP/pick} ({pick}/{test_rating_data.user_id.max()})')\n",
    "        \n",
    "print(f'최종 : AVG_mAP@{top_N_size} = {AVG_mAP/test_rating_data.user_id.max()}')"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "1ec7e7a7bda2752b69488ef6b463cd212a85bea9beda62a0e3d72f3155be411c"
  },
  "kernelspec": {
   "display_name": "Python 3.8.10 ('venv': venv)",
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}