Курсовая_рекомендательная_система.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8724633e",
   "metadata": {},
   "source": [
    "# Методы машинного обучения в рекомендательных системах"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05799d28",
   "metadata": {},
   "source": [
    "### Описание\n",
    "\n",
    "**Цель:** создать рекомендательную систему фильмов, используя методы машинного обучения.\n",
    "\n",
    "**Задачи:**\n",
    "1) провести предобработку данных;\\\n",
    "2) реализовать алгоритмы рекомендательных систем, основанных на следующих подходах: demographic filtering, content-based filtering, collaborative filtering;\\\n",
    "3) продумать логику выдачи рекомендаций разными методами.\n",
    "\n",
    "**Ожидаемый результат:** алгоритм, ранжирующий фильмы на основании различных признаков с целью выдачи рекомендации."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1a3a7b7",
   "metadata": {},
   "source": [
    "### Dataset\n",
    "from: https://www.kaggle.com/datasets/rounakbanik/the-movies-dataset/versions/5?resource=download\n",
    "\n",
    "**Содержание**\\\n",
    "Этот набор данных состоит из следующих файлов:\n",
    "\n",
    "**movies_metadata.csv**: Основной файл метаданных фильмов. Содержит информацию о 45 000 фильмах, включенных в полный набор данных MovieLens. Характеристики включают плакаты, задники, бюджет, выручку, даты выпуска, языки, страны-производители и компании.\n",
    "\n",
    "**keywords.csv**: Содержит ключевые слова для сюжета фильма для наших фильмов MovieLens. Доступен в виде строкового объекта JSON.\n",
    "\n",
    "**credits.csv**: Содержит информацию об актерском составе и съемочной группе для всех наших фильмов.\n",
    "\n",
    "**links.csv**: файл, содержащий идентификаторы TMDB и IMDB всех фильмов, представленных в полном наборе данных MovieLens.\n",
    "\n",
    "**links_small.csv**: Содержит идентификаторы TMDB и IMDB небольшого подмножества из 9000 фильмов из полного набора данных.\n",
    "\n",
    "**ratings_small.csv**: Подмножество из 100 000 оценок от 700 пользователей по 9 000 фильмам."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "77311b34",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "00723429",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 1.41 s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>adult</th>\n",
       "      <th>belongs_to_collection</th>\n",
       "      <th>budget</th>\n",
       "      <th>genres</th>\n",
       "      <th>homepage</th>\n",
       "      <th>id</th>\n",
       "      <th>imdb_id</th>\n",
       "      <th>original_language</th>\n",
       "      <th>original_title</th>\n",
       "      <th>overview</th>\n",
       "      <th>...</th>\n",
       "      <th>release_date</th>\n",
       "      <th>revenue</th>\n",
       "      <th>runtime</th>\n",
       "      <th>spoken_languages</th>\n",
       "      <th>status</th>\n",
       "      <th>tagline</th>\n",
       "      <th>title</th>\n",
       "      <th>video</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "      <td>{'id': 10194, 'name': 'Toy Story Collection', ...</td>\n",
       "      <td>30000000</td>\n",
       "      <td>[{'id': 16, 'name': 'Animation'}, {'id': 35, '...</td>\n",
       "      <td>http://toystory.disney.com/toy-story</td>\n",
       "      <td>862</td>\n",
       "      <td>tt0114709</td>\n",
       "      <td>en</td>\n",
       "      <td>Toy Story</td>\n",
       "      <td>Led by Woody, Andy's toys live happily in his ...</td>\n",
       "      <td>...</td>\n",
       "      <td>1995-10-30</td>\n",
       "      <td>373554033.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>[{'iso_639_1': 'en', 'name': 'English'}]</td>\n",
       "      <td>Released</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Toy Story</td>\n",
       "      <td>False</td>\n",
       "      <td>7.7</td>\n",
       "      <td>5415.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>65000000</td>\n",
       "      <td>[{'id': 12, 'name': 'Adventure'}, {'id': 14, '...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8844</td>\n",
       "      <td>tt0113497</td>\n",
       "      <td>en</td>\n",
       "      <td>Jumanji</td>\n",
       "      <td>When siblings Judy and Peter discover an encha...</td>\n",
       "      <td>...</td>\n",
       "      <td>1995-12-15</td>\n",
       "      <td>262797249.0</td>\n",
       "      <td>104.0</td>\n",
       "      <td>[{'iso_639_1': 'en', 'name': 'English'}, {'iso...</td>\n",
       "      <td>Released</td>\n",
       "      <td>Roll the dice and unleash the excitement!</td>\n",
       "      <td>Jumanji</td>\n",
       "      <td>False</td>\n",
       "      <td>6.9</td>\n",
       "      <td>2413.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>{'id': 119050, 'name': 'Grumpy Old Men Collect...</td>\n",
       "      <td>0</td>\n",
       "      <td>[{'id': 10749, 'name': 'Romance'}, {'id': 35, ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15602</td>\n",
       "      <td>tt0113228</td>\n",
       "      <td>en</td>\n",
       "      <td>Grumpier Old Men</td>\n",
       "      <td>A family wedding reignites the ancient feud be...</td>\n",
       "      <td>...</td>\n",
       "      <td>1995-12-22</td>\n",
       "      <td>0.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>[{'iso_639_1': 'en', 'name': 'English'}]</td>\n",
       "      <td>Released</td>\n",
       "      <td>Still Yelling. Still Fighting. Still Ready for...</td>\n",
       "      <td>Grumpier Old Men</td>\n",
       "      <td>False</td>\n",
       "      <td>6.5</td>\n",
       "      <td>92.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>16000000</td>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>31357</td>\n",
       "      <td>tt0114885</td>\n",
       "      <td>en</td>\n",
       "      <td>Waiting to Exhale</td>\n",
       "      <td>Cheated on, mistreated and stepped on, the wom...</td>\n",
       "      <td>...</td>\n",
       "      <td>1995-12-22</td>\n",
       "      <td>81452156.0</td>\n",
       "      <td>127.0</td>\n",
       "      <td>[{'iso_639_1': 'en', 'name': 'English'}]</td>\n",
       "      <td>Released</td>\n",
       "      <td>Friends are the people who let you be yourself...</td>\n",
       "      <td>Waiting to Exhale</td>\n",
       "      <td>False</td>\n",
       "      <td>6.1</td>\n",
       "      <td>34.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>{'id': 96871, 'name': 'Father of the Bride Col...</td>\n",
       "      <td>0</td>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}]</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11862</td>\n",
       "      <td>tt0113041</td>\n",
       "      <td>en</td>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>Just when George Banks has recovered from his ...</td>\n",
       "      <td>...</td>\n",
       "      <td>1995-02-10</td>\n",
       "      <td>76578911.0</td>\n",
       "      <td>106.0</td>\n",
       "      <td>[{'iso_639_1': 'en', 'name': 'English'}]</td>\n",
       "      <td>Released</td>\n",
       "      <td>Just When His World Is Back To Normal... He's ...</td>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>False</td>\n",
       "      <td>5.7</td>\n",
       "      <td>173.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   adult                              belongs_to_collection    budget   \n",
       "0  False  {'id': 10194, 'name': 'Toy Story Collection', ...  30000000  \\\n",
       "1  False                                                NaN  65000000   \n",
       "2  False  {'id': 119050, 'name': 'Grumpy Old Men Collect...         0   \n",
       "3  False                                                NaN  16000000   \n",
       "4  False  {'id': 96871, 'name': 'Father of the Bride Col...         0   \n",
       "\n",
       "                                              genres   \n",
       "0  [{'id': 16, 'name': 'Animation'}, {'id': 35, '...  \\\n",
       "1  [{'id': 12, 'name': 'Adventure'}, {'id': 14, '...   \n",
       "2  [{'id': 10749, 'name': 'Romance'}, {'id': 35, ...   \n",
       "3  [{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...   \n",
       "4                     [{'id': 35, 'name': 'Comedy'}]   \n",
       "\n",
       "                               homepage     id    imdb_id original_language   \n",
       "0  http://toystory.disney.com/toy-story    862  tt0114709                en  \\\n",
       "1                                   NaN   8844  tt0113497                en   \n",
       "2                                   NaN  15602  tt0113228                en   \n",
       "3                                   NaN  31357  tt0114885                en   \n",
       "4                                   NaN  11862  tt0113041                en   \n",
       "\n",
       "                original_title   \n",
       "0                    Toy Story  \\\n",
       "1                      Jumanji   \n",
       "2             Grumpier Old Men   \n",
       "3            Waiting to Exhale   \n",
       "4  Father of the Bride Part II   \n",
       "\n",
       "                                            overview  ... release_date   \n",
       "0  Led by Woody, Andy's toys live happily in his ...  ...   1995-10-30  \\\n",
       "1  When siblings Judy and Peter discover an encha...  ...   1995-12-15   \n",
       "2  A family wedding reignites the ancient feud be...  ...   1995-12-22   \n",
       "3  Cheated on, mistreated and stepped on, the wom...  ...   1995-12-22   \n",
       "4  Just when George Banks has recovered from his ...  ...   1995-02-10   \n",
       "\n",
       "       revenue runtime                                   spoken_languages   \n",
       "0  373554033.0    81.0           [{'iso_639_1': 'en', 'name': 'English'}]  \\\n",
       "1  262797249.0   104.0  [{'iso_639_1': 'en', 'name': 'English'}, {'iso...   \n",
       "2          0.0   101.0           [{'iso_639_1': 'en', 'name': 'English'}]   \n",
       "3   81452156.0   127.0           [{'iso_639_1': 'en', 'name': 'English'}]   \n",
       "4   76578911.0   106.0           [{'iso_639_1': 'en', 'name': 'English'}]   \n",
       "\n",
       "     status                                            tagline   \n",
       "0  Released                                                NaN  \\\n",
       "1  Released          Roll the dice and unleash the excitement!   \n",
       "2  Released  Still Yelling. Still Fighting. Still Ready for...   \n",
       "3  Released  Friends are the people who let you be yourself...   \n",
       "4  Released  Just When His World Is Back To Normal... He's ...   \n",
       "\n",
       "                         title  video vote_average vote_count  \n",
       "0                    Toy Story  False          7.7     5415.0  \n",
       "1                      Jumanji  False          6.9     2413.0  \n",
       "2             Grumpier Old Men  False          6.5       92.0  \n",
       "3            Waiting to Exhale  False          6.1       34.0  \n",
       "4  Father of the Bride Part II  False          5.7      173.0  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "df = pd.read_csv('movies_metadata.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "bd2bf358",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['adult', 'belongs_to_collection', 'budget', 'genres', 'homepage', 'id',\n",
       "       'imdb_id', 'original_language', 'original_title', 'overview',\n",
       "       'popularity', 'poster_path', 'production_companies',\n",
       "       'production_countries', 'release_date', 'revenue', 'runtime',\n",
       "       'spoken_languages', 'status', 'tagline', 'title', 'video',\n",
       "       'vote_average', 'vote_count'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "372995eb",
   "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>genres</th>\n",
       "      <th>id</th>\n",
       "      <th>overview</th>\n",
       "      <th>popularity</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>title</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[{'id': 16, 'name': 'Animation'}, {'id': 35, '...</td>\n",
       "      <td>862</td>\n",
       "      <td>Led by Woody, Andy's toys live happily in his ...</td>\n",
       "      <td>21.946943</td>\n",
       "      <td>[{'name': 'Pixar Animation Studios', 'id': 3}]</td>\n",
       "      <td>Toy Story</td>\n",
       "      <td>7.7</td>\n",
       "      <td>5415.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[{'id': 12, 'name': 'Adventure'}, {'id': 14, '...</td>\n",
       "      <td>8844</td>\n",
       "      <td>When siblings Judy and Peter discover an encha...</td>\n",
       "      <td>17.015539</td>\n",
       "      <td>[{'name': 'TriStar Pictures', 'id': 559}, {'na...</td>\n",
       "      <td>Jumanji</td>\n",
       "      <td>6.9</td>\n",
       "      <td>2413.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[{'id': 10749, 'name': 'Romance'}, {'id': 35, ...</td>\n",
       "      <td>15602</td>\n",
       "      <td>A family wedding reignites the ancient feud be...</td>\n",
       "      <td>11.7129</td>\n",
       "      <td>[{'name': 'Warner Bros.', 'id': 6194}, {'name'...</td>\n",
       "      <td>Grumpier Old Men</td>\n",
       "      <td>6.5</td>\n",
       "      <td>92.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...</td>\n",
       "      <td>31357</td>\n",
       "      <td>Cheated on, mistreated and stepped on, the wom...</td>\n",
       "      <td>3.859495</td>\n",
       "      <td>[{'name': 'Twentieth Century Fox Film Corporat...</td>\n",
       "      <td>Waiting to Exhale</td>\n",
       "      <td>6.1</td>\n",
       "      <td>34.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}]</td>\n",
       "      <td>11862</td>\n",
       "      <td>Just when George Banks has recovered from his ...</td>\n",
       "      <td>8.387519</td>\n",
       "      <td>[{'name': 'Sandollar Productions', 'id': 5842}...</td>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>5.7</td>\n",
       "      <td>173.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              genres     id   \n",
       "0  [{'id': 16, 'name': 'Animation'}, {'id': 35, '...    862  \\\n",
       "1  [{'id': 12, 'name': 'Adventure'}, {'id': 14, '...   8844   \n",
       "2  [{'id': 10749, 'name': 'Romance'}, {'id': 35, ...  15602   \n",
       "3  [{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...  31357   \n",
       "4                     [{'id': 35, 'name': 'Comedy'}]  11862   \n",
       "\n",
       "                                            overview popularity   \n",
       "0  Led by Woody, Andy's toys live happily in his ...  21.946943  \\\n",
       "1  When siblings Judy and Peter discover an encha...  17.015539   \n",
       "2  A family wedding reignites the ancient feud be...    11.7129   \n",
       "3  Cheated on, mistreated and stepped on, the wom...   3.859495   \n",
       "4  Just when George Banks has recovered from his ...   8.387519   \n",
       "\n",
       "                                production_companies   \n",
       "0     [{'name': 'Pixar Animation Studios', 'id': 3}]  \\\n",
       "1  [{'name': 'TriStar Pictures', 'id': 559}, {'na...   \n",
       "2  [{'name': 'Warner Bros.', 'id': 6194}, {'name'...   \n",
       "3  [{'name': 'Twentieth Century Fox Film Corporat...   \n",
       "4  [{'name': 'Sandollar Productions', 'id': 5842}...   \n",
       "\n",
       "                         title  vote_average  vote_count  \n",
       "0                    Toy Story           7.7      5415.0  \n",
       "1                      Jumanji           6.9      2413.0  \n",
       "2             Grumpier Old Men           6.5        92.0  \n",
       "3            Waiting to Exhale           6.1        34.0  \n",
       "4  Father of the Bride Part II           5.7       173.0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2 = df[['genres','id','overview','popularity','production_companies','title','vote_average','vote_count']]\n",
    "df2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "75902e17",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "genres                    0\n",
       "id                        0\n",
       "overview                954\n",
       "popularity                5\n",
       "production_companies      3\n",
       "title                     6\n",
       "vote_average              6\n",
       "vote_count                6\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2.isna().sum() # проверим наличие пропусков"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "72c362a9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(45466, 8)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "79c409df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(44506, 8)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# удалим пропущенные значения\n",
    "df2 = df2.dropna().reset_index(drop=True)\n",
    "df2.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d04e69a5",
   "metadata": {},
   "source": [
    "**В датасете есть фильмы, информация о которых записана несколько раз** "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7092fe95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id\n",
       "141971     True\n",
       "10991      True\n",
       "14788      True\n",
       "97995      True\n",
       "15028      True\n",
       "          ...  \n",
       "27583     False\n",
       "6883      False\n",
       "26390     False\n",
       "34016     False\n",
       "461257    False\n",
       "Name: count, Length: 44476, dtype: bool"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2['id'].value_counts()>1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "1ec49786",
   "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>genres</th>\n",
       "      <th>id</th>\n",
       "      <th>overview</th>\n",
       "      <th>popularity</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>title</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5689</th>\n",
       "      <td>[{'id': 18, 'name': 'Drama'}]</td>\n",
       "      <td>109962</td>\n",
       "      <td>Two literary women compete for 20 years: one w...</td>\n",
       "      <td>12.180836</td>\n",
       "      <td>[{'name': 'Metro-Goldwyn-Mayer (MGM)', 'id': 8...</td>\n",
       "      <td>Rich and Famous</td>\n",
       "      <td>4.9</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20745</th>\n",
       "      <td>[{'id': 18, 'name': 'Drama'}]</td>\n",
       "      <td>109962</td>\n",
       "      <td>Two literary women compete for 20 years: one w...</td>\n",
       "      <td>10.396878</td>\n",
       "      <td>[{'name': 'Metro-Goldwyn-Mayer (MGM)', 'id': 8...</td>\n",
       "      <td>Rich and Famous</td>\n",
       "      <td>4.9</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              genres      id   \n",
       "5689   [{'id': 18, 'name': 'Drama'}]  109962  \\\n",
       "20745  [{'id': 18, 'name': 'Drama'}]  109962   \n",
       "\n",
       "                                                overview popularity   \n",
       "5689   Two literary women compete for 20 years: one w...  12.180836  \\\n",
       "20745  Two literary women compete for 20 years: one w...  10.396878   \n",
       "\n",
       "                                    production_companies            title   \n",
       "5689   [{'name': 'Metro-Goldwyn-Mayer (MGM)', 'id': 8...  Rich and Famous  \\\n",
       "20745  [{'name': 'Metro-Goldwyn-Mayer (MGM)', 'id': 8...  Rich and Famous   \n",
       "\n",
       "       vote_average  vote_count  \n",
       "5689            4.9         7.0  \n",
       "20745           4.9         7.0  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2[df2['id']=='109962']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6ff1ca62",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# оставим по одной записи о каждом фильме\n",
    "df2 = df2.drop_duplicates(subset='id').reset_index(drop=True)\n",
    "sum(df2['id'].value_counts()>1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c40715db",
   "metadata": {},
   "source": [
    "# Demographic Filtering"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e3c5c682",
   "metadata": {},
   "source": [
    "### Холодный старт\n",
    "\n",
    "Холодный старт – это типичная ситуация, когда ещё не накоплено достаточное количество данных для корректной работы рекомендательной системы. В этой ситуации будем рекомендовать самые высокорейтинговые и самые популярные фильмы."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "35343ba7",
   "metadata": {},
   "source": [
    "**Movie rating**\n",
    "\n",
    "https://en.wikipedia.org/wiki/Shrinkage_(statistics)\n",
    "\n",
    "https://help.imdb.com/article/imdb/track-movies-tv/ratings-faq/G67Y87TFYYP6TWAV#"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbbb636e",
   "metadata": {},
   "source": [
    "$$\\text{Weighted Rating (WR)} = \\frac{v}{v+m} \\cdot R + \\frac{m}{v+m} \\cdot C$$\n",
    "* v - количество голосов, отданных за фильм;\n",
    "* m - минимальное количество голосов, необходимое для включения в таблицу;\n",
    "* R - средний рейтинг фильма;\n",
    "* C - среднее число голосов по всему датасету"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4d591b68",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Среднее число голосов = 5.639018796654375\n",
      "У фильма должно быть хотя бы m = 165.0 оценок\n",
      "Количество фильмов = 4458\n",
      "Wall time: 16.6 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "C = df2['vote_average'].mean()\n",
    "print(f'Среднее число голосов = {C}')\n",
    "\n",
    "m = df2['vote_count'].quantile(0.9)\n",
    "print(f'У фильма должно быть хотя бы m = {m} оценок')\n",
    "\n",
    "# отберем фильмы, удовлетворяющие условию\n",
    "high_rated_films = df2.copy()[df2['vote_count'] >= m]\n",
    "print(f'Количество фильмов = {high_rated_films.shape[0]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56d8d7d7",
   "metadata": {},
   "source": [
    "Отобрано 4458 фильмов с достаточным количеством голосов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d331a757",
   "metadata": {},
   "outputs": [],
   "source": [
    "# функция, рассчитывающая взвешеный рейтинг (WR)\n",
    "def weighted_rating(x, m, C): \n",
    "    v = x['vote_count']\n",
    "    R = x['vote_average']\n",
    "    return (v/(v+m) * R) + (m/(m+v) * C)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "77f37910",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 109 ms\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>genres</th>\n",
       "      <th>id</th>\n",
       "      <th>overview</th>\n",
       "      <th>popularity</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>title</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "      <th>WR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[{'id': 16, 'name': 'Animation'}, {'id': 35, '...</td>\n",
       "      <td>862</td>\n",
       "      <td>Led by Woody, Andy's toys live happily in his ...</td>\n",
       "      <td>21.946943</td>\n",
       "      <td>[{'name': 'Pixar Animation Studios', 'id': 3}]</td>\n",
       "      <td>Toy Story</td>\n",
       "      <td>7.7</td>\n",
       "      <td>5415.0</td>\n",
       "      <td>7.639057</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[{'id': 12, 'name': 'Adventure'}, {'id': 14, '...</td>\n",
       "      <td>8844</td>\n",
       "      <td>When siblings Judy and Peter discover an encha...</td>\n",
       "      <td>17.015539</td>\n",
       "      <td>[{'name': 'TriStar Pictures', 'id': 559}, {'na...</td>\n",
       "      <td>Jumanji</td>\n",
       "      <td>6.9</td>\n",
       "      <td>2413.0</td>\n",
       "      <td>6.819293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}]</td>\n",
       "      <td>11862</td>\n",
       "      <td>Just when George Banks has recovered from his ...</td>\n",
       "      <td>8.387519</td>\n",
       "      <td>[{'name': 'Sandollar Productions', 'id': 5842}...</td>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>5.7</td>\n",
       "      <td>173.0</td>\n",
       "      <td>5.670231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>[{'id': 28, 'name': 'Action'}, {'id': 80, 'nam...</td>\n",
       "      <td>949</td>\n",
       "      <td>Obsessive master thief, Neil McCauley leads a ...</td>\n",
       "      <td>17.924927</td>\n",
       "      <td>[{'name': 'Regency Enterprises', 'id': 508}, {...</td>\n",
       "      <td>Heat</td>\n",
       "      <td>7.7</td>\n",
       "      <td>1886.0</td>\n",
       "      <td>7.534197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>[{'id': 28, 'name': 'Action'}, {'id': 12, 'nam...</td>\n",
       "      <td>9091</td>\n",
       "      <td>International action superstar Jean Claude Van...</td>\n",
       "      <td>5.23158</td>\n",
       "      <td>[{'name': 'Universal Pictures', 'id': 33}, {'n...</td>\n",
       "      <td>Sudden Death</td>\n",
       "      <td>5.5</td>\n",
       "      <td>174.0</td>\n",
       "      <td>5.567664</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              genres     id   \n",
       "0  [{'id': 16, 'name': 'Animation'}, {'id': 35, '...    862  \\\n",
       "1  [{'id': 12, 'name': 'Adventure'}, {'id': 14, '...   8844   \n",
       "4                     [{'id': 35, 'name': 'Comedy'}]  11862   \n",
       "5  [{'id': 28, 'name': 'Action'}, {'id': 80, 'nam...    949   \n",
       "8  [{'id': 28, 'name': 'Action'}, {'id': 12, 'nam...   9091   \n",
       "\n",
       "                                            overview popularity   \n",
       "0  Led by Woody, Andy's toys live happily in his ...  21.946943  \\\n",
       "1  When siblings Judy and Peter discover an encha...  17.015539   \n",
       "4  Just when George Banks has recovered from his ...   8.387519   \n",
       "5  Obsessive master thief, Neil McCauley leads a ...  17.924927   \n",
       "8  International action superstar Jean Claude Van...    5.23158   \n",
       "\n",
       "                                production_companies   \n",
       "0     [{'name': 'Pixar Animation Studios', 'id': 3}]  \\\n",
       "1  [{'name': 'TriStar Pictures', 'id': 559}, {'na...   \n",
       "4  [{'name': 'Sandollar Productions', 'id': 5842}...   \n",
       "5  [{'name': 'Regency Enterprises', 'id': 508}, {...   \n",
       "8  [{'name': 'Universal Pictures', 'id': 33}, {'n...   \n",
       "\n",
       "                         title  vote_average  vote_count        WR  \n",
       "0                    Toy Story           7.7      5415.0  7.639057  \n",
       "1                      Jumanji           6.9      2413.0  6.819293  \n",
       "4  Father of the Bride Part II           5.7       173.0  5.670231  \n",
       "5                         Heat           7.7      1886.0  7.534197  \n",
       "8                 Sudden Death           5.5       174.0  5.567664  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "high_rated_films['WR'] = high_rated_films.apply(weighted_rating, axis=1, args=(m,C))\n",
    "high_rated_films.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0989b7e",
   "metadata": {},
   "source": [
    "### Рекомендация самых высокооцененных фильмов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "83a433db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 472 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# сортировка фильмов по убыванию score\n",
    "high_rated_films = high_rated_films.sort_values('WR', ascending=False)\n",
    "\n",
    "# топ-10 самых высокооцененных фильмов\n",
    "top_highest_rated = high_rated_films[['title', 'vote_count', 'vote_average', 'WR']].head(10)\n",
    "\n",
    "plt.figure(figsize=(12,5))\n",
    "\n",
    "plt.barh(top_highest_rated['title'], \n",
    "         top_highest_rated['WR'].head(10),\n",
    "         color='lightgreen'\n",
    ")\n",
    "\n",
    "plt.gca().invert_yaxis()\n",
    "plt.xlabel(\"Weighted Rating\")\n",
    "plt.title(\"Best Movies\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0387a4c",
   "metadata": {},
   "source": [
    "### Рекомендация самых популярных фильмов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "679a2bff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 370 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "df2['popularity'] = df2['popularity'].astype('float64')\n",
    "\n",
    "popular_movies = df2.sort_values('popularity', ascending=False)\n",
    "\n",
    "plt.figure(figsize=(12,5))\n",
    "\n",
    "plt.barh(popular_movies['title'].head(10), \n",
    "         popular_movies['popularity'].head(10),\n",
    "         color='skyblue'\n",
    ")\n",
    "\n",
    "plt.gca().invert_yaxis()\n",
    "plt.xlabel(\"Popularity\")\n",
    "plt.title(\"Popular Movies\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "57b72ada",
   "metadata": {},
   "source": [
    "# Content-Based Filtering\n",
    "\n",
    "**Для заданного фильма найдем похожие на него по описанию**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "c164fec2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        Led by Woody, Andy's toys live happily in his ...\n",
       "1        When siblings Judy and Peter discover an encha...\n",
       "2        A family wedding reignites the ancient feud be...\n",
       "3        Cheated on, mistreated and stepped on, the wom...\n",
       "4        Just when George Banks has recovered from his ...\n",
       "                               ...                        \n",
       "44471          Rising and falling between a man and woman.\n",
       "44472    An artist struggles to finish his work while a...\n",
       "44473    When one of her hits goes wrong, a professiona...\n",
       "44474    In a small town live two brothers, one a minis...\n",
       "44475    50 years after decriminalisation of homosexual...\n",
       "Name: overview, Length: 44476, dtype: object"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2['overview']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "763d3e49",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a5b7b67e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 4.79 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(44476, 75827)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "tfidf = TfidfVectorizer(stop_words='english')\n",
    "# создание TF-IDF matrix\n",
    "tfidf_matrix = tfidf.fit_transform(df2['overview'])\n",
    "tfidf_matrix.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "94f07db4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "title\n",
       "Toy Story                      0\n",
       "Jumanji                        1\n",
       "Grumpier Old Men               2\n",
       "Waiting to Exhale              3\n",
       "Father of the Bride Part II    4\n",
       "dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# создадим series, в которой индексы - названия фильмов\n",
    "indices = pd.Series(df2.index, index=df2['title']).drop_duplicates()\n",
    "indices.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a2d333b",
   "metadata": {},
   "source": [
    "### Создание рекомендации на основе сходства описаний фильмов"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1d8c32e",
   "metadata": {},
   "source": [
    "**1. Cosine similarity**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "413cffc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.metrics.pairwise import cosine_similarity"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1978226c",
   "metadata": {},
   "source": [
    "$$\\text{similarity}(x,y) = \\frac{xy^T}{\\Vert x \\Vert \\Vert y \\Vert}$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "1f9fe4c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_recommendations_overview_cossim(title):\n",
    "    \n",
    "    # находим индекс фильма по title\n",
    "    idx = indices[title]\n",
    "    \n",
    "    # рассчитываем косинусное расстояние для конкретного фильма и всех остальных\n",
    "    cosine_sim = cosine_similarity(tfidf_matrix[idx], tfidf_matrix)[0]\n",
    "    \n",
    "    # сортируем\n",
    "    sorted_cosine_sim = cosine_sim.argsort()\n",
    "\n",
    "    # выводим 10 наиболее похожих фильмов\n",
    "    return df2['title'].iloc[sorted_cosine_sim[-11:-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "ea7c6677",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 80.9 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "40783                           Alcatraz Island\n",
       "9196     Female Prisoner Scorpion: Jailhouse 41\n",
       "33663                   Women's Prison Massacre\n",
       "9361                                    In Hell\n",
       "17342                           Girls in Prison\n",
       "36063                            Women's Prison\n",
       "11292                               Brute Force\n",
       "38404                         Seven Times Seven\n",
       "6523                                Civil Brand\n",
       "16858                   They Made Me a Fugitive\n",
       "Name: title, dtype: object"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_overview_cossim('The Shawshank Redemption')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "327521b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 76.5 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "12458               Dead Awake\n",
       "35440             Frankenstein\n",
       "5315                   Perfect\n",
       "29873               Freerunner\n",
       "28123                    Julia\n",
       "22704                  In Fear\n",
       "23739           Make Your Move\n",
       "5181                 Delirious\n",
       "31390           Unknown Caller\n",
       "15388    9/11: The Twin Towers\n",
       "Name: title, dtype: object"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_overview_cossim('Fight Club')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea93ccea",
   "metadata": {},
   "source": [
    "**2. KMeans**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "7aac9a8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "f7fbd990",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 3 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(4458, 75827)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# отберем строки tfidf_matrix, соответсвующие фильмам с наибольшим рейтингом\n",
    "tfidf_matrix_for_KMeans = tfidf_matrix[high_rated_films.index]\n",
    "tfidf_matrix_for_KMeans.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "f3f1f02a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14.86"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "4458 / 300"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "30157432",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 13.5 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([268, 233, 117,   5,  32, 243, 218, 203, 218, 281])"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "# обучим модель, чтобы она кластерзовала фильмы по их описаниям\n",
    "kmeans = KMeans(n_clusters=300, random_state=0, n_init=\"auto\").fit(tfidf_matrix_for_KMeans)\n",
    "clustered_movies = kmeans.labels_\n",
    "\n",
    "clustered_movies[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "6cd879a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_recommendations_overview_KMeans(title):\n",
    "    \n",
    "    # находим индекс фильма по title\n",
    "    df_t = high_rated_films.reset_index(drop=True)\n",
    "    idx = (df_t[df_t['title']==title]).index[0]\n",
    "    \n",
    "    # получаем названия наиболее похожих фильмов\n",
    "    recommendation = high_rated_films['title'].iloc[clustered_movies==clustered_movies[idx]]\n",
    "\n",
    "    return recommendation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "17bb10c3",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 6.07 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "312      The Shawshank Redemption\n",
       "28212              McFarland, USA\n",
       "257             A Little Princess\n",
       "979                 Cool Runnings\n",
       "4716              The Last Castle\n",
       "11174                  Invincible\n",
       "11428             We Are Marshall\n",
       "10063            The Longest Yard\n",
       "6056         Bend It Like Beckham\n",
       "2203                 The Waterboy\n",
       "9792                 Ice Princess\n",
       "Name: title, dtype: object"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_overview_KMeans('The Shawshank Redemption')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "3e8b84cf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 5.56 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "2824                                  Fight Club\n",
       "1187                          Dead Poets Society\n",
       "12981                                   The Wave\n",
       "9425                                  The Chorus\n",
       "11772                            This Is England\n",
       "7320                                 Le Samouraï\n",
       "31127                 I Can Quit Whenever I Want\n",
       "22155                            Fuck You Goethe\n",
       "16883                                Melancholia\n",
       "18659                                 Detachment\n",
       "804                                      Matilda\n",
       "6689                              School of Rock\n",
       "27374                                   The DUFF\n",
       "20503                               In the House\n",
       "3312                Guess Who's Coming to Dinner\n",
       "15739                                     Easy A\n",
       "4751                        The Devil's Backbone\n",
       "15973                Undisputed III : Redemption\n",
       "8221                               Bad Education\n",
       "11275                               The Guardian\n",
       "11446                         Notes on a Scandal\n",
       "2466                                    Election\n",
       "13901                             The Ugly Truth\n",
       "23570                            Wish I Was Here\n",
       "17915                        What's Your Number?\n",
       "14146                       World's Greatest Dad\n",
       "4513                   See No Evil, Hear No Evil\n",
       "3302                                    Red Dawn\n",
       "11156                                   Accepted\n",
       "41092                                 Fist Fight\n",
       "10419                              Elizabethtown\n",
       "27195                                   Expelled\n",
       "13461                                    Knowing\n",
       "3111             Sister Act 2: Back in the Habit\n",
       "2449                           Never Been Kissed\n",
       "37783                    A Hologram for the King\n",
       "6170                    The Lizzie McGuire Movie\n",
       "2353                              She's All That\n",
       "32283                                    Cooties\n",
       "35828                            Serial Teachers\n",
       "19684                                   Red Dawn\n",
       "6361       Legally Blonde 2: Red, White & Blonde\n",
       "21108                               Mean Girls 2\n",
       "6088                                Dreamcatcher\n",
       "31173                                The Gallows\n",
       "6308     Dumb and Dumberer: When Harry Met Lloyd\n",
       "Name: title, dtype: object"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_overview_KMeans('Fight Club')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cd3cb09",
   "metadata": {},
   "source": [
    "### Используя жанры, ключевые слова, состав актеров и режиссеров"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "35b073a0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 5.2 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((46419, 2), (45476, 3))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "df_keywords = pd.read_csv('keywords.csv', dtype={'id': 'object'})\n",
    "df_credits = pd.read_csv('credits.csv', dtype={'id': 'object'})\n",
    "\n",
    "df_keywords.shape, df_credits.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "44b7648a",
   "metadata": {},
   "source": [
    "**Дополним датасет дополнительной информацей о фильме**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "daa80d74",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 136 ms\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>genres</th>\n",
       "      <th>id</th>\n",
       "      <th>overview</th>\n",
       "      <th>popularity</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>title</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "      <th>keywords</th>\n",
       "      <th>cast</th>\n",
       "      <th>crew</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[{'id': 16, 'name': 'Animation'}, {'id': 35, '...</td>\n",
       "      <td>862</td>\n",
       "      <td>Led by Woody, Andy's toys live happily in his ...</td>\n",
       "      <td>21.946943</td>\n",
       "      <td>[{'name': 'Pixar Animation Studios', 'id': 3}]</td>\n",
       "      <td>Toy Story</td>\n",
       "      <td>7.7</td>\n",
       "      <td>5415.0</td>\n",
       "      <td>[{'id': 931, 'name': 'jealousy'}, {'id': 4290,...</td>\n",
       "      <td>[{'cast_id': 14, 'character': 'Woody (voice)',...</td>\n",
       "      <td>[{'credit_id': '52fe4284c3a36847f8024f49', 'de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[{'id': 12, 'name': 'Adventure'}, {'id': 14, '...</td>\n",
       "      <td>8844</td>\n",
       "      <td>When siblings Judy and Peter discover an encha...</td>\n",
       "      <td>17.015539</td>\n",
       "      <td>[{'name': 'TriStar Pictures', 'id': 559}, {'na...</td>\n",
       "      <td>Jumanji</td>\n",
       "      <td>6.9</td>\n",
       "      <td>2413.0</td>\n",
       "      <td>[{'id': 10090, 'name': 'board game'}, {'id': 1...</td>\n",
       "      <td>[{'cast_id': 1, 'character': 'Alan Parrish', '...</td>\n",
       "      <td>[{'credit_id': '52fe44bfc3a36847f80a7cd1', 'de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[{'id': 10749, 'name': 'Romance'}, {'id': 35, ...</td>\n",
       "      <td>15602</td>\n",
       "      <td>A family wedding reignites the ancient feud be...</td>\n",
       "      <td>11.712900</td>\n",
       "      <td>[{'name': 'Warner Bros.', 'id': 6194}, {'name'...</td>\n",
       "      <td>Grumpier Old Men</td>\n",
       "      <td>6.5</td>\n",
       "      <td>92.0</td>\n",
       "      <td>[{'id': 1495, 'name': 'fishing'}, {'id': 12392...</td>\n",
       "      <td>[{'cast_id': 2, 'character': 'Max Goldman', 'c...</td>\n",
       "      <td>[{'credit_id': '52fe466a9251416c75077a89', 'de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...</td>\n",
       "      <td>31357</td>\n",
       "      <td>Cheated on, mistreated and stepped on, the wom...</td>\n",
       "      <td>3.859495</td>\n",
       "      <td>[{'name': 'Twentieth Century Fox Film Corporat...</td>\n",
       "      <td>Waiting to Exhale</td>\n",
       "      <td>6.1</td>\n",
       "      <td>34.0</td>\n",
       "      <td>[{'id': 818, 'name': 'based on novel'}, {'id':...</td>\n",
       "      <td>[{'cast_id': 1, 'character': \"Savannah 'Vannah...</td>\n",
       "      <td>[{'credit_id': '52fe44779251416c91011acb', 'de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}]</td>\n",
       "      <td>11862</td>\n",
       "      <td>Just when George Banks has recovered from his ...</td>\n",
       "      <td>8.387519</td>\n",
       "      <td>[{'name': 'Sandollar Productions', 'id': 5842}...</td>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>5.7</td>\n",
       "      <td>173.0</td>\n",
       "      <td>[{'id': 1009, 'name': 'baby'}, {'id': 1599, 'n...</td>\n",
       "      <td>[{'cast_id': 1, 'character': 'George Banks', '...</td>\n",
       "      <td>[{'credit_id': '52fe44959251416c75039ed7', 'de...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              genres     id   \n",
       "0  [{'id': 16, 'name': 'Animation'}, {'id': 35, '...    862  \\\n",
       "1  [{'id': 12, 'name': 'Adventure'}, {'id': 14, '...   8844   \n",
       "2  [{'id': 10749, 'name': 'Romance'}, {'id': 35, ...  15602   \n",
       "3  [{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...  31357   \n",
       "4                     [{'id': 35, 'name': 'Comedy'}]  11862   \n",
       "\n",
       "                                            overview  popularity   \n",
       "0  Led by Woody, Andy's toys live happily in his ...   21.946943  \\\n",
       "1  When siblings Judy and Peter discover an encha...   17.015539   \n",
       "2  A family wedding reignites the ancient feud be...   11.712900   \n",
       "3  Cheated on, mistreated and stepped on, the wom...    3.859495   \n",
       "4  Just when George Banks has recovered from his ...    8.387519   \n",
       "\n",
       "                                production_companies   \n",
       "0     [{'name': 'Pixar Animation Studios', 'id': 3}]  \\\n",
       "1  [{'name': 'TriStar Pictures', 'id': 559}, {'na...   \n",
       "2  [{'name': 'Warner Bros.', 'id': 6194}, {'name'...   \n",
       "3  [{'name': 'Twentieth Century Fox Film Corporat...   \n",
       "4  [{'name': 'Sandollar Productions', 'id': 5842}...   \n",
       "\n",
       "                         title  vote_average  vote_count   \n",
       "0                    Toy Story           7.7      5415.0  \\\n",
       "1                      Jumanji           6.9      2413.0   \n",
       "2             Grumpier Old Men           6.5        92.0   \n",
       "3            Waiting to Exhale           6.1        34.0   \n",
       "4  Father of the Bride Part II           5.7       173.0   \n",
       "\n",
       "                                            keywords   \n",
       "0  [{'id': 931, 'name': 'jealousy'}, {'id': 4290,...  \\\n",
       "1  [{'id': 10090, 'name': 'board game'}, {'id': 1...   \n",
       "2  [{'id': 1495, 'name': 'fishing'}, {'id': 12392...   \n",
       "3  [{'id': 818, 'name': 'based on novel'}, {'id':...   \n",
       "4  [{'id': 1009, 'name': 'baby'}, {'id': 1599, 'n...   \n",
       "\n",
       "                                                cast   \n",
       "0  [{'cast_id': 14, 'character': 'Woody (voice)',...  \\\n",
       "1  [{'cast_id': 1, 'character': 'Alan Parrish', '...   \n",
       "2  [{'cast_id': 2, 'character': 'Max Goldman', 'c...   \n",
       "3  [{'cast_id': 1, 'character': \"Savannah 'Vannah...   \n",
       "4  [{'cast_id': 1, 'character': 'George Banks', '...   \n",
       "\n",
       "                                                crew  \n",
       "0  [{'credit_id': '52fe4284c3a36847f8024f49', 'de...  \n",
       "1  [{'credit_id': '52fe44bfc3a36847f80a7cd1', 'de...  \n",
       "2  [{'credit_id': '52fe466a9251416c75077a89', 'de...  \n",
       "3  [{'credit_id': '52fe44779251416c91011acb', 'de...  \n",
       "4  [{'credit_id': '52fe44959251416c75039ed7', 'de...  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "df2 = df2.merge(df_keywords, on='id')\n",
    "df2 = df2.merge(df_credits, on='id')\n",
    "df2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "3193a0d5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# проверим наличие пропусков\n",
    "df2.isna().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "9f44be2b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "957"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(df2['id'].value_counts()>1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "0e07d6b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2 = df2.drop_duplicates(subset='id').reset_index(drop=True)\n",
    "sum(df2['id'].value_counts()>1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "ff6f9123",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(44475, 11)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a7ff4633",
   "metadata": {},
   "source": [
    "**Некоторые признаки записаны в виде строк вида '[{}, {}]', их нужно преобразовать к списку словарей**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "df44b37d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overwriting conver_strlist.py\n"
     ]
    }
   ],
   "source": [
    "%%file conver_strlist.py\n",
    "from ast import literal_eval\n",
    "def conver_strlist(dff):\n",
    "    # получим список жанров\n",
    "    dff['genres'] = dff['genres'].apply(literal_eval).apply(lambda x: [i['name'] for i in x])\n",
    "\n",
    "    # получим список компаний\n",
    "    dff['production_companies'] = dff['production_companies'].apply(literal_eval).apply(lambda x: [i['name'] for i in x])\n",
    "\n",
    "    # получим список ключевых слов\n",
    "    dff['keywords'] = dff['keywords'].apply(literal_eval).apply(lambda x: [i['name'] for i in x])\n",
    "\n",
    "    # получим список актеров\n",
    "    dff['cast'] = dff['cast'].apply(literal_eval).apply(lambda x: [i['name'] for i in x])\n",
    "\n",
    "    # получим список режиссеров\n",
    "    dff['crew'] = dff['crew'].apply(literal_eval).apply(lambda x: [i['name'] for i in x if i['job']=='Director'])\n",
    "    \n",
    "    return dff"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "c9260635",
   "metadata": {},
   "outputs": [],
   "source": [
    "from ast import literal_eval\n",
    "import multiprocessing\n",
    "from conver_strlist import conver_strlist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "e532f5d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 14.7 s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>genres</th>\n",
       "      <th>id</th>\n",
       "      <th>overview</th>\n",
       "      <th>popularity</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>title</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "      <th>keywords</th>\n",
       "      <th>cast</th>\n",
       "      <th>crew</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[Animation, Comedy, Family]</td>\n",
       "      <td>862</td>\n",
       "      <td>Led by Woody, Andy's toys live happily in his ...</td>\n",
       "      <td>21.946943</td>\n",
       "      <td>[Pixar Animation Studios]</td>\n",
       "      <td>Toy Story</td>\n",
       "      <td>7.7</td>\n",
       "      <td>5415.0</td>\n",
       "      <td>[jealousy, toy, boy, friendship, friends, riva...</td>\n",
       "      <td>[Tom Hanks, Tim Allen, Don Rickles, Jim Varney...</td>\n",
       "      <td>[John Lasseter]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[Adventure, Fantasy, Family]</td>\n",
       "      <td>8844</td>\n",
       "      <td>When siblings Judy and Peter discover an encha...</td>\n",
       "      <td>17.015539</td>\n",
       "      <td>[TriStar Pictures, Teitler Film, Interscope Co...</td>\n",
       "      <td>Jumanji</td>\n",
       "      <td>6.9</td>\n",
       "      <td>2413.0</td>\n",
       "      <td>[board game, disappearance, based on children'...</td>\n",
       "      <td>[Robin Williams, Jonathan Hyde, Kirsten Dunst,...</td>\n",
       "      <td>[Joe Johnston]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[Romance, Comedy]</td>\n",
       "      <td>15602</td>\n",
       "      <td>A family wedding reignites the ancient feud be...</td>\n",
       "      <td>11.712900</td>\n",
       "      <td>[Warner Bros., Lancaster Gate]</td>\n",
       "      <td>Grumpier Old Men</td>\n",
       "      <td>6.5</td>\n",
       "      <td>92.0</td>\n",
       "      <td>[fishing, best friend, duringcreditsstinger, o...</td>\n",
       "      <td>[Walter Matthau, Jack Lemmon, Ann-Margret, Sop...</td>\n",
       "      <td>[Howard Deutch]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[Comedy, Drama, Romance]</td>\n",
       "      <td>31357</td>\n",
       "      <td>Cheated on, mistreated and stepped on, the wom...</td>\n",
       "      <td>3.859495</td>\n",
       "      <td>[Twentieth Century Fox Film Corporation]</td>\n",
       "      <td>Waiting to Exhale</td>\n",
       "      <td>6.1</td>\n",
       "      <td>34.0</td>\n",
       "      <td>[based on novel, interracial relationship, sin...</td>\n",
       "      <td>[Whitney Houston, Angela Bassett, Loretta Devi...</td>\n",
       "      <td>[Forest Whitaker]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[Comedy]</td>\n",
       "      <td>11862</td>\n",
       "      <td>Just when George Banks has recovered from his ...</td>\n",
       "      <td>8.387519</td>\n",
       "      <td>[Sandollar Productions, Touchstone Pictures]</td>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>5.7</td>\n",
       "      <td>173.0</td>\n",
       "      <td>[baby, midlife crisis, confidence, aging, daug...</td>\n",
       "      <td>[Steve Martin, Diane Keaton, Martin Short, Kim...</td>\n",
       "      <td>[Charles Shyer]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         genres     id   \n",
       "0   [Animation, Comedy, Family]    862  \\\n",
       "1  [Adventure, Fantasy, Family]   8844   \n",
       "2             [Romance, Comedy]  15602   \n",
       "3      [Comedy, Drama, Romance]  31357   \n",
       "4                      [Comedy]  11862   \n",
       "\n",
       "                                            overview  popularity   \n",
       "0  Led by Woody, Andy's toys live happily in his ...   21.946943  \\\n",
       "1  When siblings Judy and Peter discover an encha...   17.015539   \n",
       "2  A family wedding reignites the ancient feud be...   11.712900   \n",
       "3  Cheated on, mistreated and stepped on, the wom...    3.859495   \n",
       "4  Just when George Banks has recovered from his ...    8.387519   \n",
       "\n",
       "                                production_companies   \n",
       "0                          [Pixar Animation Studios]  \\\n",
       "1  [TriStar Pictures, Teitler Film, Interscope Co...   \n",
       "2                     [Warner Bros., Lancaster Gate]   \n",
       "3           [Twentieth Century Fox Film Corporation]   \n",
       "4       [Sandollar Productions, Touchstone Pictures]   \n",
       "\n",
       "                         title  vote_average  vote_count   \n",
       "0                    Toy Story           7.7      5415.0  \\\n",
       "1                      Jumanji           6.9      2413.0   \n",
       "2             Grumpier Old Men           6.5        92.0   \n",
       "3            Waiting to Exhale           6.1        34.0   \n",
       "4  Father of the Bride Part II           5.7       173.0   \n",
       "\n",
       "                                            keywords   \n",
       "0  [jealousy, toy, boy, friendship, friends, riva...  \\\n",
       "1  [board game, disappearance, based on children'...   \n",
       "2  [fishing, best friend, duringcreditsstinger, o...   \n",
       "3  [based on novel, interracial relationship, sin...   \n",
       "4  [baby, midlife crisis, confidence, aging, daug...   \n",
       "\n",
       "                                                cast               crew  \n",
       "0  [Tom Hanks, Tim Allen, Don Rickles, Jim Varney...    [John Lasseter]  \n",
       "1  [Robin Williams, Jonathan Hyde, Kirsten Dunst,...     [Joe Johnston]  \n",
       "2  [Walter Matthau, Jack Lemmon, Ann-Margret, Sop...    [Howard Deutch]  \n",
       "3  [Whitney Houston, Angela Bassett, Loretta Devi...  [Forest Whitaker]  \n",
       "4  [Steve Martin, Diane Keaton, Martin Short, Kim...    [Charles Shyer]  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "num_cores = multiprocessing.cpu_count()\n",
    "df2_split = np.array_split(df2, num_cores, axis=0)\n",
    "\n",
    "with multiprocessing.Pool(processes=num_cores) as pool:\n",
    "    df_extra_info = pd.concat(pool.map(conver_strlist, df2_split))\n",
    "    \n",
    "df_extra_info.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d57cd460",
   "metadata": {},
   "source": [
    "**Преобразуем данные:**\n",
    "* приведем слова к нижнему регистру\n",
    "* удалим пробелы, чтобы различать разных людей с одним именем, например, Tom Hardy и Tom Cruise**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "baf23920",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overwriting convert_low_nospace.py\n"
     ]
    }
   ],
   "source": [
    "%%file convert_low_nospace.py\n",
    "def convert_low_nospace(dff):\n",
    "    features = ['title', 'genres','production_companies','keywords','crew']\n",
    "    \n",
    "    for feature in features[1:]:\n",
    "        dff[feature] = dff[feature].apply(\n",
    "            lambda x: [i.replace(\" \", \"\").lower() for i in x]\n",
    "        )\n",
    "    \n",
    "    return dff[features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "88916d67",
   "metadata": {},
   "outputs": [],
   "source": [
    "from convert_low_nospace import convert_low_nospace"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "52441048",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 2.96 s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>title</th>\n",
       "      <th>genres</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>keywords</th>\n",
       "      <th>crew</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Toy Story</td>\n",
       "      <td>[animation, comedy, family]</td>\n",
       "      <td>[pixaranimationstudios]</td>\n",
       "      <td>[jealousy, toy, boy, friendship, friends, riva...</td>\n",
       "      <td>[johnlasseter]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Jumanji</td>\n",
       "      <td>[adventure, fantasy, family]</td>\n",
       "      <td>[tristarpictures, teitlerfilm, interscopecommu...</td>\n",
       "      <td>[boardgame, disappearance, basedonchildren'sbo...</td>\n",
       "      <td>[joejohnston]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Grumpier Old Men</td>\n",
       "      <td>[romance, comedy]</td>\n",
       "      <td>[warnerbros., lancastergate]</td>\n",
       "      <td>[fishing, bestfriend, duringcreditsstinger, ol...</td>\n",
       "      <td>[howarddeutch]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Waiting to Exhale</td>\n",
       "      <td>[comedy, drama, romance]</td>\n",
       "      <td>[twentiethcenturyfoxfilmcorporation]</td>\n",
       "      <td>[basedonnovel, interracialrelationship, single...</td>\n",
       "      <td>[forestwhitaker]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Father of the Bride Part II</td>\n",
       "      <td>[comedy]</td>\n",
       "      <td>[sandollarproductions, touchstonepictures]</td>\n",
       "      <td>[baby, midlifecrisis, confidence, aging, daugh...</td>\n",
       "      <td>[charlesshyer]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         title                        genres   \n",
       "0                    Toy Story   [animation, comedy, family]  \\\n",
       "1                      Jumanji  [adventure, fantasy, family]   \n",
       "2             Grumpier Old Men             [romance, comedy]   \n",
       "3            Waiting to Exhale      [comedy, drama, romance]   \n",
       "4  Father of the Bride Part II                      [comedy]   \n",
       "\n",
       "                                production_companies   \n",
       "0                            [pixaranimationstudios]  \\\n",
       "1  [tristarpictures, teitlerfilm, interscopecommu...   \n",
       "2                       [warnerbros., lancastergate]   \n",
       "3               [twentiethcenturyfoxfilmcorporation]   \n",
       "4         [sandollarproductions, touchstonepictures]   \n",
       "\n",
       "                                            keywords              crew  \n",
       "0  [jealousy, toy, boy, friendship, friends, riva...    [johnlasseter]  \n",
       "1  [boardgame, disappearance, basedonchildren'sbo...     [joejohnston]  \n",
       "2  [fishing, bestfriend, duringcreditsstinger, ol...    [howarddeutch]  \n",
       "3  [basedonnovel, interracialrelationship, single...  [forestwhitaker]  \n",
       "4  [baby, midlifecrisis, confidence, aging, daugh...    [charlesshyer]  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "num_cores = multiprocessing.cpu_count()\n",
    "df_extra_info_split = np.array_split(df_extra_info, num_cores, axis=0)\n",
    "\n",
    "with multiprocessing.Pool(processes=num_cores) as pool:\n",
    "    df_with_cast = pd.concat(pool.map(convert_low_nospace, df_extra_info_split))\n",
    "    \n",
    "df_with_cast.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c32870b4",
   "metadata": {},
   "source": [
    "**Получим текстовую строку с информацией**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "17521795",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 503 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0        animation comedy family pixaranimationstudios ...\n",
       "1        adventure fantasy family tristarpictures teitl...\n",
       "2        romance comedy warnerbros. lancastergate fishi...\n",
       "3        comedy drama romance twentiethcenturyfoxfilmco...\n",
       "4        comedy sandollarproductions touchstonepictures...\n",
       "                               ...                        \n",
       "44470              drama family tragiclove hamidnematollah\n",
       "44471           drama sineolivia artist play pinoy lavdiaz\n",
       "44472    action drama thriller americanworldpictures ma...\n",
       "44473                            yermoliev yakovprotazanov\n",
       "44474                                         daisyasquith\n",
       "Name: genres, Length: 44475, dtype: object"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "soup = df_with_cast['genres'].copy()\n",
    "for i in features[1:]:\n",
    "    soup += df_with_cast[i].copy()\n",
    "soup = soup.apply(lambda x: ' '.join(x))\n",
    "soup"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee6935fb",
   "metadata": {},
   "source": [
    "**Создание рекомендации на основе сходства жанров, ключевых слов, состава команды**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "780a8330",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.metrics.pairwise import cosine_similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "a2df5008",
   "metadata": {},
   "outputs": [],
   "source": [
    "CntVctr = CountVectorizer(stop_words='english')\n",
    "count_matrix = CntVctr.fit_transform(soup)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "0851bc98",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_recommendations_cast_cossim(title):\n",
    "    idx = indices[title]\n",
    "    \n",
    "    # Получаем косинусное расстояние для конкретного фильма и всех остальных\n",
    "    cosine_sim = cosine_similarity(count_matrix[idx], count_matrix)\n",
    "\n",
    "    sorted_cosine_sim = cosine_sim.argsort(axis=1)[0]\n",
    "\n",
    "    recommendation = sorted_cosine_sim[-11:-1]\n",
    "\n",
    "    return df2['title'].iloc[recommendation]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "60100d3c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 22 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "14788                   Caged\n",
       "38875      Lights of New York\n",
       "11894                     McQ\n",
       "26560               Juke Girl\n",
       "25700           Lady Gangster\n",
       "43303    High School Big Shot\n",
       "27676      The Strange Affair\n",
       "28242               Unchained\n",
       "23910    Castle on the Hudson\n",
       "33928    Duffy of San Quentin\n",
       "Name: title, dtype: object"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_cast_cossim('The Shawshank Redemption')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "51540a26",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 20.3 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "4933                The Five Heartbeats\n",
       "13007    I Can Get it For You Wholesale\n",
       "2465                        Pushing Tin\n",
       "2915                  Anywhere But Here\n",
       "11321                            Flicka\n",
       "24238                        Blue Denim\n",
       "17417               Tender Is the Night\n",
       "18606                Satan Never Sleeps\n",
       "16475        Phone Call from a Stranger\n",
       "25886                   Three Brave Men\n",
       "Name: title, dtype: object"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_cast_cossim('Fight Club')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "35a7ec69",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import MiniBatchKMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "6b6cce16",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 25.8 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([721, 990,   4,  32, 627,  27, 781, 685, 163, 163])"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "kmb = MiniBatchKMeans(n_clusters=1000, random_state=0, n_init=\"auto\").fit(count_matrix)\n",
    "clustered_movies_kmb = kmb.labels_\n",
    "\n",
    "clustered_movies_kmb[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "847f7160",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_recommendations_cast_MiniBatchKMeans(title):\n",
    "    \n",
    "    # находим индекс фильма по title\n",
    "    idx = indices[title]\n",
    "    \n",
    "    particular_cluster = clustered_movies_kmb[idx]\n",
    "    # получаем названия наиболее похожих фильмов\n",
    "    recommendation = df2['title'].iloc[clustered_movies_kmb==particular_cluster]\n",
    "\n",
    "    return recommendation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "6d45a742",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 4 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "15                                        Casino\n",
       "29                                Shanghai Triad\n",
       "30                               Dangerous Minds\n",
       "40                               Dead Presidents\n",
       "79       Things to Do in Denver When You're Dead\n",
       "                          ...                   \n",
       "43847                             Secret Defense\n",
       "43977                             The Wrong Road\n",
       "43981                          One Hundred Steps\n",
       "44060                   G:MT Greenwich Mean Time\n",
       "44072                                Junction 48\n",
       "Name: title, Length: 855, dtype: object"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_cast_MiniBatchKMeans('The Shawshank Redemption')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "c1834d4d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 1.56 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "25                          Othello\n",
       "883                Sunset Boulevard\n",
       "887                   All About Eve\n",
       "889                         Rebecca\n",
       "1809        How Green Was My Valley\n",
       "                    ...            \n",
       "34762             The Man I Married\n",
       "35323                 Woman's World\n",
       "35640             The Young Messiah\n",
       "36865    The Truth About Bebe Donge\n",
       "38663                   Night Train\n",
       "Name: title, Length: 139, dtype: object"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "get_recommendations_cast_MiniBatchKMeans('Fight Club')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "124da58c",
   "metadata": {},
   "source": [
    "# Collaborative Filtering"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1fd96570",
   "metadata": {},
   "source": [
    "https://surprise.readthedocs.io/en/stable/getting_started.html#getting-started\n",
    "\n",
    "**Коллаборативная фильтрация** (англ. **collaborative filtering**) — это один из методов построения прогнозов в рекомендательных системах, использующий известные предпочтения (оценки) группы пользователей для предсказывания неизвестных предпочтений другого пользователя. Его основное допущение состоит в следующем: те, кто одинаково оценивал какие-либо предметы в прошлом, склонны давать похожие оценки другим предметам и в будущем."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "a947c4e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import Reader, Dataset, SVD, accuracy\n",
    "from surprise.model_selection import cross_validate, train_test_split, GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ccafc62c",
   "metadata": {},
   "source": [
    "## SVD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "e5dfcc05",
   "metadata": {},
   "outputs": [],
   "source": [
    "reader = Reader()\n",
    "ratings = pd.read_csv('ratings_small.csv')\n",
    "data = Dataset.load_from_df(ratings[['userId', 'movieId', 'rating']], reader)\n",
    "trainset, testset = train_test_split(data, test_size=0.25)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b270f294",
   "metadata": {},
   "source": [
    "**Подберем оптимальные гиперпараметры для модели SVD**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "0dd31863",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9013202269066788\n",
      "{'n_epochs': 40, 'lr_all': 0.005, 'reg_all': 0.4}\n",
      "Wall time: 7min 23s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "param_grid = {\"n_epochs\": [20, 30, 40], \"lr_all\": [0.005, 0.01, 1], \"reg_all\": [0.02, 0.4, 1]}\n",
    "gs = GridSearchCV(SVD, param_grid, measures=[\"rmse\", \"mae\"], cv=3)\n",
    "\n",
    "gs.fit(data)\n",
    "print(gs.best_score[\"rmse\"])\n",
    "print(gs.best_params[\"rmse\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "id": "2324cf76",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE: 0.9011\n",
      "0.9011322752622644\n",
      "MAE:  0.6997\n",
      "0.6996578484112456\n",
      "Wall time: 7.4 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "algo = SVD(n_epochs=40, lr_all=0.005, reg_all=0.4)\n",
    "algo.fit(trainset)\n",
    "predictions = algo.test(testset)\n",
    "print(accuracy.rmse(predictions))\n",
    "print(accuracy.mae(predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e72868d0",
   "metadata": {},
   "source": [
    "**Судя по MAE, при прогнозировании оценки фильма модель ошибается в среднем на 0.6997, то есть меньше, чем на балл - довально хороший результат.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "b4739eb4",
   "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>uid</th>\n",
       "      <th>iid</th>\n",
       "      <th>r_ui</th>\n",
       "      <th>est</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21774</th>\n",
       "      <td>22</td>\n",
       "      <td>8950</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.646226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19591</th>\n",
       "      <td>105</td>\n",
       "      <td>2916</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.422974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8016</th>\n",
       "      <td>384</td>\n",
       "      <td>31685</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.148068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15011</th>\n",
       "      <td>472</td>\n",
       "      <td>2285</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.440291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16441</th>\n",
       "      <td>102</td>\n",
       "      <td>58</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.921006</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       uid    iid  r_ui       est\n",
       "21774   22   8950   4.0  3.646226\n",
       "19591  105   2916   4.0  3.422974\n",
       "8016   384  31685   3.0  3.148068\n",
       "15011  472   2285   4.0  3.440291\n",
       "16441  102     58   4.0  3.921006"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred_svd = pd.DataFrame(predictions).drop('details', axis=1)\n",
    "df_pred_svd.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "013f175e",
   "metadata": {},
   "source": [
    "### Изобразим предсказание для некоторых пользователей"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "92036ec3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize_test_user(user_id):\n",
    "    user_for_vis = list(\n",
    "        filter(\n",
    "            lambda x: x[0]==user_id,\n",
    "            testset\n",
    "        )\n",
    "    )\n",
    "\n",
    "    x = list(map(lambda x: x[2], user_for_vis))\n",
    "    y = [algo.predict(*i)[3] for i in user_for_vis]\n",
    "\n",
    "    plt.figure(figsize=(10,4))\n",
    "    plt.title(f'Comparing values with predictions, user = {user_id}')\n",
    "    plt.plot(range(len(x)), x, label='true')\n",
    "    plt.plot(range(len(x)),y, c='r', label=\"predicted\")\n",
    "    plt.legend()\n",
    "    plt.grid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "3194bee0",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualize_test_user(598)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "id": "40c1d485",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualize_test_user(486)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db6e4621",
   "metadata": {},
   "source": [
    "### Пример рекомендации"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "b3d0df35",
   "metadata": {
    "scrolled": true
   },
   "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>uid</th>\n",
       "      <th>iid</th>\n",
       "      <th>r_ui</th>\n",
       "      <th>est</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>31</td>\n",
       "      <td>527</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.332079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>598</td>\n",
       "      <td>4310</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.165126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>529</td>\n",
       "      <td>2716</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.595869</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>486</td>\n",
       "      <td>2716</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.771479</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>624</td>\n",
       "      <td>7153</td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.509867</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   uid   iid  r_ui       est\n",
       "0   31   527   5.0  4.332079\n",
       "1  598  4310   2.0  3.165126\n",
       "2  529  2716   4.0  3.595869\n",
       "3  486  2716   4.0  3.771479\n",
       "4  624  7153   4.5  3.509867"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# возьмем 3 пользователей и предскажем их оценку определенных фильмов\n",
    "example_svd = testset[:5] # (user_id, item_id, movie_rating_by_user)\n",
    "pd.DataFrame([algo.predict(*example_svd[i]) for i in range(5)]).drop('details', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebf3076a",
   "metadata": {},
   "source": [
    "**Будем рекомендовать фильм пользователю в случае, если предсказанная оценка этого фильма выше 60% оценок фильмов, которые нам известны**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "08bd5489",
   "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>userId</th>\n",
       "      <th>movieId</th>\n",
       "      <th>rating</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>46419</th>\n",
       "      <td>340</td>\n",
       "      <td>6</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1410374053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46420</th>\n",
       "      <td>340</td>\n",
       "      <td>47</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1410374238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46421</th>\n",
       "      <td>340</td>\n",
       "      <td>293</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1410374113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46422</th>\n",
       "      <td>340</td>\n",
       "      <td>296</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1410374199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46423</th>\n",
       "      <td>340</td>\n",
       "      <td>318</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1410252785</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       userId  movieId  rating   timestamp\n",
       "46419     340        6     4.0  1410374053\n",
       "46420     340       47     5.0  1410374238\n",
       "46421     340      293     4.0  1410374113\n",
       "46422     340      296     3.5  1410374199\n",
       "46423     340      318     5.0  1410252785"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "real WR = 5.0\n",
      "0.6-quantile = 4.0\n",
      "est_rating = 4.060067224790942\n",
      "YES\n"
     ]
    }
   ],
   "source": [
    "# df, содержащий оцененные пользователем 340 фильмы, кроме фильма 4993\n",
    "user_example = ratings[(ratings['userId']==340)&(ratings['movieId']!=4993)]\n",
    "display(user_example.head())\n",
    "\n",
    "real_rating_exam = ratings[(ratings['userId']==340)&(ratings['movieId']==4993)]\n",
    "print('real WR =', real_rating_exam.values[0][2])\n",
    "\n",
    "# найдем 0.6-квантиль оценок пользователя 340\n",
    "quan_exam = user_example['rating'].quantile(0.6)\n",
    "print(f'0.6-quantile = {quan_exam}')\n",
    "\n",
    "est_rating_exam = algo.predict(*real_rating_exam.values[0][:-1])[3]\n",
    "print(f'est_rating = {est_rating_exam}')\n",
    "\n",
    "# решим, рекомендовать фильм 4993 или нет\n",
    "print('YES') if est_rating_exam>=quan_exam else print('NO')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2c613c0",
   "metadata": {},
   "source": [
    "**В данном случае алогоритм верно решил рекомендовать. Пояснение:**\n",
    "1. 60% фильмов, которые этот пользователь оценивал, получили рейтинг 4 и ниже;\n",
    "2. пользователь оценил данный фильм на 5, значит, фильм понравился и его следует рекомендовать;\n",
    "3. модель предсказала оценку 4.06, что выше 0.6-квантиля, так что она правильно рекомендует этот фильм"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "25723207",
   "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>userId</th>\n",
       "      <th>movieId</th>\n",
       "      <th>rating</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90016</th>\n",
       "      <td>598</td>\n",
       "      <td>1</td>\n",
       "      <td>4.5</td>\n",
       "      <td>1057554797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90017</th>\n",
       "      <td>598</td>\n",
       "      <td>11</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1008571467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90018</th>\n",
       "      <td>598</td>\n",
       "      <td>15</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1007344372</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90019</th>\n",
       "      <td>598</td>\n",
       "      <td>17</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1009415293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90020</th>\n",
       "      <td>598</td>\n",
       "      <td>39</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1008571296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90323</th>\n",
       "      <td>598</td>\n",
       "      <td>6377</td>\n",
       "      <td>4.5</td>\n",
       "      <td>1057554688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90324</th>\n",
       "      <td>598</td>\n",
       "      <td>6461</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1057555106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90325</th>\n",
       "      <td>598</td>\n",
       "      <td>6534</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1057554710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90326</th>\n",
       "      <td>598</td>\n",
       "      <td>6537</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1057555152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90327</th>\n",
       "      <td>598</td>\n",
       "      <td>6539</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1057555166</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>311 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       userId  movieId  rating   timestamp\n",
       "90016     598        1     4.5  1057554797\n",
       "90017     598       11     4.0  1008571467\n",
       "90018     598       15     2.0  1007344372\n",
       "90019     598       17     2.0  1009415293\n",
       "90020     598       39     3.0  1008571296\n",
       "...       ...      ...     ...         ...\n",
       "90323     598     6377     4.5  1057554688\n",
       "90324     598     6461     3.5  1057555106\n",
       "90325     598     6534     2.0  1057554710\n",
       "90326     598     6537     4.0  1057555152\n",
       "90327     598     6539     5.0  1057555166\n",
       "\n",
       "[311 rows x 4 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "real WR = 2.0\n",
      "0.6-quantile = 4.0\n",
      "est_rating = 3.165126494590933\n",
      "NO\n"
     ]
    }
   ],
   "source": [
    "# df, содержащий оцененные пользователем 598 фильмы, кроме фильма 4310\n",
    "user_example = ratings[(ratings['userId']==598)&(ratings['movieId']!=4310)]\n",
    "display(user_example)\n",
    "\n",
    "real_rating_exam = ratings[(ratings['userId']==598)&(ratings['movieId']==4310)]\n",
    "print('real WR =', real_rating_exam.values[0][2])\n",
    "\n",
    "# найдем 0.6-квантиль оценок пользователя 598\n",
    "quan_exam = user_example['rating'].quantile(0.6)\n",
    "print(f'0.6-quantile = {quan_exam}')\n",
    "\n",
    "est_rating_exam = algo.predict(*real_rating_exam.values[0][:-1])[3]\n",
    "print(f'est_rating = {est_rating_exam}')\n",
    "\n",
    "# решим, рекомендовать фильм 4310 или нет\n",
    "print('YES') if est_rating_exam>=quan_exam else print('NO')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a4f9bdbe",
   "metadata": {},
   "source": [
    "**В данном случае алогоритм верно решил не рекомендовать этот фильм**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b91a666",
   "metadata": {},
   "source": [
    "### SVD recommendation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "fb1e8386",
   "metadata": {
    "scrolled": true
   },
   "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>genres</th>\n",
       "      <th>id</th>\n",
       "      <th>overview</th>\n",
       "      <th>popularity</th>\n",
       "      <th>production_companies</th>\n",
       "      <th>title</th>\n",
       "      <th>vote_average</th>\n",
       "      <th>vote_count</th>\n",
       "      <th>WR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>312</th>\n",
       "      <td>[{'id': 18, 'name': 'Drama'}, {'id': 80, 'name...</td>\n",
       "      <td>278</td>\n",
       "      <td>Framed in the 1940s for the double murder of h...</td>\n",
       "      <td>51.645403</td>\n",
       "      <td>[{'name': 'Castle Rock Entertainment', 'id': 9...</td>\n",
       "      <td>The Shawshank Redemption</td>\n",
       "      <td>8.5</td>\n",
       "      <td>8358.0</td>\n",
       "      <td>8.444613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>823</th>\n",
       "      <td>[{'id': 18, 'name': 'Drama'}, {'id': 80, 'name...</td>\n",
       "      <td>238</td>\n",
       "      <td>Spanning the years 1945 to 1955, a chronicle o...</td>\n",
       "      <td>41.109264</td>\n",
       "      <td>[{'name': 'Paramount Pictures', 'id': 4}, {'na...</td>\n",
       "      <td>The Godfather</td>\n",
       "      <td>8.5</td>\n",
       "      <td>6024.0</td>\n",
       "      <td>8.423726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10276</th>\n",
       "      <td>[{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...</td>\n",
       "      <td>19404</td>\n",
       "      <td>Raj is a rich, carefree, happy-go-lucky second...</td>\n",
       "      <td>34.457024</td>\n",
       "      <td>[{'name': 'Yash Raj Films', 'id': 1569}]</td>\n",
       "      <td>Dilwale Dulhania Le Jayenge</td>\n",
       "      <td>9.1</td>\n",
       "      <td>661.0</td>\n",
       "      <td>8.408642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12443</th>\n",
       "      <td>[{'id': 18, 'name': 'Drama'}, {'id': 28, 'name...</td>\n",
       "      <td>155</td>\n",
       "      <td>Batman raises the stakes in his war on crime. ...</td>\n",
       "      <td>123.167259</td>\n",
       "      <td>[{'name': 'DC Comics', 'id': 429}, {'name': 'L...</td>\n",
       "      <td>The Dark Knight</td>\n",
       "      <td>8.3</td>\n",
       "      <td>12269.0</td>\n",
       "      <td>8.264689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2824</th>\n",
       "      <td>[{'id': 18, 'name': 'Drama'}]</td>\n",
       "      <td>550</td>\n",
       "      <td>A ticking-time-bomb insomniac and a slippery s...</td>\n",
       "      <td>63.869599</td>\n",
       "      <td>[{'name': 'Twentieth Century Fox Film Corporat...</td>\n",
       "      <td>Fight Club</td>\n",
       "      <td>8.3</td>\n",
       "      <td>9678.0</td>\n",
       "      <td>8.255393</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  genres     id   \n",
       "312    [{'id': 18, 'name': 'Drama'}, {'id': 80, 'name...    278  \\\n",
       "823    [{'id': 18, 'name': 'Drama'}, {'id': 80, 'name...    238   \n",
       "10276  [{'id': 35, 'name': 'Comedy'}, {'id': 18, 'nam...  19404   \n",
       "12443  [{'id': 18, 'name': 'Drama'}, {'id': 28, 'name...    155   \n",
       "2824                       [{'id': 18, 'name': 'Drama'}]    550   \n",
       "\n",
       "                                                overview  popularity   \n",
       "312    Framed in the 1940s for the double murder of h...   51.645403  \\\n",
       "823    Spanning the years 1945 to 1955, a chronicle o...   41.109264   \n",
       "10276  Raj is a rich, carefree, happy-go-lucky second...   34.457024   \n",
       "12443  Batman raises the stakes in his war on crime. ...  123.167259   \n",
       "2824   A ticking-time-bomb insomniac and a slippery s...   63.869599   \n",
       "\n",
       "                                    production_companies   \n",
       "312    [{'name': 'Castle Rock Entertainment', 'id': 9...  \\\n",
       "823    [{'name': 'Paramount Pictures', 'id': 4}, {'na...   \n",
       "10276           [{'name': 'Yash Raj Films', 'id': 1569}]   \n",
       "12443  [{'name': 'DC Comics', 'id': 429}, {'name': 'L...   \n",
       "2824   [{'name': 'Twentieth Century Fox Film Corporat...   \n",
       "\n",
       "                             title  vote_average  vote_count        WR  \n",
       "312       The Shawshank Redemption           8.5      8358.0  8.444613  \n",
       "823                  The Godfather           8.5      6024.0  8.423726  \n",
       "10276  Dilwale Dulhania Le Jayenge           9.1       661.0  8.408642  \n",
       "12443              The Dark Knight           8.3     12269.0  8.264689  \n",
       "2824                    Fight Club           8.3      9678.0  8.255393  "
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "high_rated_films.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "41b4e4ed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4458, 9)"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "high_rated_films_svd = high_rated_films.astype({'id':'float64'})\n",
    "high_rated_films_svd.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "314731a0",
   "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>movieId</th>\n",
       "      <th>imdbId</th>\n",
       "      <th>tmdbId</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>114709</td>\n",
       "      <td>862.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>113497</td>\n",
       "      <td>8844.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>113228</td>\n",
       "      <td>15602.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>114885</td>\n",
       "      <td>31357.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>113041</td>\n",
       "      <td>11862.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   movieId  imdbId   tmdbId\n",
       "0        1  114709    862.0\n",
       "1        2  113497   8844.0\n",
       "2        3  113228  15602.0\n",
       "3        4  114885  31357.0\n",
       "4        5  113041  11862.0"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links = pd.read_csv('links_small.csv').dropna()\n",
    "links.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "1448a949",
   "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>movieId</th>\n",
       "      <th>imdbId</th>\n",
       "      <th>tmdbId</th>\n",
       "      <th>id</th>\n",
       "      <th>title</th>\n",
       "      <th>WR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>318</td>\n",
       "      <td>111161</td>\n",
       "      <td>278.0</td>\n",
       "      <td>278.0</td>\n",
       "      <td>The Shawshank Redemption</td>\n",
       "      <td>8.444613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226</th>\n",
       "      <td>858</td>\n",
       "      <td>68646</td>\n",
       "      <td>238.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>The Godfather</td>\n",
       "      <td>8.423726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2266</th>\n",
       "      <td>58559</td>\n",
       "      <td>468569</td>\n",
       "      <td>155.0</td>\n",
       "      <td>155.0</td>\n",
       "      <td>The Dark Knight</td>\n",
       "      <td>8.264689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>899</th>\n",
       "      <td>2959</td>\n",
       "      <td>137523</td>\n",
       "      <td>550.0</td>\n",
       "      <td>550.0</td>\n",
       "      <td>Fight Club</td>\n",
       "      <td>8.255393</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>296</td>\n",
       "      <td>110912</td>\n",
       "      <td>680.0</td>\n",
       "      <td>680.0</td>\n",
       "      <td>Pulp Fiction</td>\n",
       "      <td>8.250304</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3079</th>\n",
       "      <td>103596</td>\n",
       "      <td>2724064</td>\n",
       "      <td>205321.0</td>\n",
       "      <td>205321.0</td>\n",
       "      <td>Sharknado</td>\n",
       "      <td>4.267547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2321</th>\n",
       "      <td>61348</td>\n",
       "      <td>1213644</td>\n",
       "      <td>13805.0</td>\n",
       "      <td>13805.0</td>\n",
       "      <td>Disaster Movie</td>\n",
       "      <td>4.109489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1047</th>\n",
       "      <td>3593</td>\n",
       "      <td>185183</td>\n",
       "      <td>5491.0</td>\n",
       "      <td>5491.0</td>\n",
       "      <td>Battlefield Earth</td>\n",
       "      <td>4.026977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2095</th>\n",
       "      <td>50798</td>\n",
       "      <td>799949</td>\n",
       "      <td>9760.0</td>\n",
       "      <td>9760.0</td>\n",
       "      <td>Epic Movie</td>\n",
       "      <td>4.006489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2416</th>\n",
       "      <td>67867</td>\n",
       "      <td>1098327</td>\n",
       "      <td>14164.0</td>\n",
       "      <td>14164.0</td>\n",
       "      <td>Dragonball Evolution</td>\n",
       "      <td>3.606153</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3528 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      movieId   imdbId    tmdbId        id                     title        WR\n",
       "99        318   111161     278.0     278.0  The Shawshank Redemption  8.444613\n",
       "226       858    68646     238.0     238.0             The Godfather  8.423726\n",
       "2266    58559   468569     155.0     155.0           The Dark Knight  8.264689\n",
       "899      2959   137523     550.0     550.0                Fight Club  8.255393\n",
       "89        296   110912     680.0     680.0              Pulp Fiction  8.250304\n",
       "...       ...      ...       ...       ...                       ...       ...\n",
       "3079   103596  2724064  205321.0  205321.0                 Sharknado  4.267547\n",
       "2321    61348  1213644   13805.0   13805.0            Disaster Movie  4.109489\n",
       "1047     3593   185183    5491.0    5491.0         Battlefield Earth  4.026977\n",
       "2095    50798   799949    9760.0    9760.0                Epic Movie  4.006489\n",
       "2416    67867  1098327   14164.0   14164.0      Dragonball Evolution  3.606153\n",
       "\n",
       "[3528 rows x 6 columns]"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_movies_links = links.merge(high_rated_films_svd[['id', 'title', 'WR']], left_on='tmdbId', right_on='id').sort_values('WR', ascending=False)\n",
    "best_movies_links_arr = best_movies_links.values\n",
    "best_movies_links"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "63d7b2a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 9.72 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<surprise.prediction_algorithms.matrix_factorization.SVD at 0x1f288d2f5e0>"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "data_svd_rec = data.build_full_trainset()\n",
    "\n",
    "svd_complete = SVD(n_epochs=40, lr_all=0.005, reg_all=0.4)\n",
    "svd_complete.fit(data_svd_rec);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "90287102",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_recommendation_collaborative_svd(user_id, alpha=0.6):\n",
    "    df_ru = ratings[(ratings['userId']==user_id)]\n",
    "    watched_movies = df_ru['movieId'].values\n",
    "    quan = df_ru['rating'].quantile(alpha)\n",
    "    \n",
    "    c = 0\n",
    "    rec = [] # список рекомендованных фильмов\n",
    "    for i in best_movies_links_arr:\n",
    "        # не рекомендуем фильм, если пользователь его уже смотреть\n",
    "        if i[0] in watched_movies:\n",
    "            continue\n",
    "            \n",
    "        movie_id = i[0]\n",
    "        movie_WR = i[5]\n",
    "        est_r = svd_complete.predict(user_id, movie_id, movie_WR)[3]\n",
    "        if est_r >= quan:\n",
    "            c += 1\n",
    "            rec.append(i[4]) # добавляем фильм в список, если предсказанная оценка > 0.6-квантиля\n",
    "            if c >= 100:\n",
    "                #np.random.shuffle(rec)\n",
    "                return rec\n",
    "    \n",
    "    if len(rec) < 20:\n",
    "        rec = get_recommendation_collaborative_svd(user_id, alpha=alpha-0.1)\n",
    "    #np.random.shuffle(rec)\n",
    "    return rec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "ded25802",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 39.2 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['Psycho',\n",
       " \"Howl's Moving Castle\",\n",
       " 'Memento',\n",
       " 'City of God',\n",
       " 'The Departed',\n",
       " 'My Neighbor Totoro',\n",
       " 'Once Upon a Time in the West',\n",
       " 'Modern Times',\n",
       " 'There Will Be Blood',\n",
       " 'Paths of Glory',\n",
       " 'Paperman',\n",
       " 'Rush',\n",
       " \"Hachi: A Dog's Tale\",\n",
       " \"Singin' in the Rain\",\n",
       " 'Captain Phillips',\n",
       " 'Raging Bull',\n",
       " 'M',\n",
       " 'Wild Tales',\n",
       " 'The Kid',\n",
       " 'The Third Man',\n",
       " 'On the Waterfront',\n",
       " 'Yojimbo',\n",
       " 'Ran',\n",
       " 'Amores perros',\n",
       " 'A Trip to the Moon',\n",
       " \"All the President's Men\",\n",
       " 'The Fighter',\n",
       " 'The General',\n",
       " 'Rebecca',\n",
       " 'Philomena',\n",
       " 'Ikiru',\n",
       " 'Strangers on a Train',\n",
       " 'In the Name of the Father',\n",
       " 'Love Me If You Dare',\n",
       " 'A Prophet',\n",
       " 'It Happened One Night',\n",
       " 'Paris, Texas',\n",
       " 'The Conversation',\n",
       " 'The Warriors',\n",
       " 'Run Lola Run',\n",
       " 'Diabolique',\n",
       " 'Tell No One',\n",
       " 'The Philadelphia Story',\n",
       " 'Blood Simple',\n",
       " 'Shadow of a Doubt',\n",
       " 'Aguirre: The Wrath of God',\n",
       " 'The Killer',\n",
       " 'Happiness',\n",
       " 'His Girl Friday',\n",
       " 'The Sea Inside',\n",
       " 'The Producers',\n",
       " 'Drunken Master',\n",
       " 'Swingers']"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "# получаем рекомендацию\n",
    "get_recommendation_collaborative_svd(119)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5a1af8f",
   "metadata": {},
   "source": [
    "# Заключение"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9bccc58",
   "metadata": {},
   "source": [
    "В ходе курсовой работы мной были реализованы следующие подходы:\n",
    "* **Demographic Filtering**, рекомендующий лучшие фильмы по рейтингу или популярности (неперсонализированная рекомендация)\n",
    "\n",
    "* **Content-based filtering**, находящий похожие по описанию или составу актеров/режиссеров фильмы. Можно использовать как отдельную функцию: если пользователю понравился сюжет или каст фильма, он может найти похожие фильмы. Реализовано через *косинусное расстояние*, *KMeans* и *MiniBatchKMeans*.\n",
    "\n",
    "* **Collaborative filtering**. Имеются данные об оценках пользователей разных фильмов (user_id, movie_id, rating). Модель находит пользователей со схожими вкусами и рекомендуюет фильмы, которые оценил один, но не оценил другой. Например: [(user_1, \"Avatar\", 8), (user_2, \"Avatar\", 8.5), (user_1, \"Fight Club\", 9)], в данном случае пользователю user_2 будет предложен к просмотру \"Fight Club\". Реализовано с помощью алгоритма *SVD* из библиотеки surprise.\n",
    "\n",
    "При первых входах пользователя в систему, ему будет предложен список фильмов, сформированный на основе **demographic filtering**. Клиент может посмотреть некоторые из этих кинокартин и оценить их или сразу присвоить рейтинг фильмам, которые он уже смотрел. Полученная информация о предпочтениях конкретного пользователя будет использоваться для персональной рекомендации при помощи подходов **collaborative filtering** и **content-based filtering**.\n",
    "\n",
    "Также в системе будет функция поиска похожих фильмов: это идея дополнительного использования 2 подхода (**content-based filtering**). Пользователь сможет явно задать интересующий его фильм, например, \"Fight Club\", и система порекомендует похожие."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}