movielens_collaborative_filtering_1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Collaborative Filtering\n",
    "- collaborative filtering에서는 사용자와 아이템 간 상호작용 데이터를 바탕으로 어떤 사용자와 유사한 취향을 가진 사용자가 좋아하는 아이템을 추천하거나(user-based collaborative filtering) 어떤 아이템과 유사한 아이템을 추천합니다(item-based collaborative filtering).\n",
    "- 사용자나 아이템의 세부정보와 다르게 상호작용 데이터는 시스템을 운영하면서 자연스럽게 축적되므로 데이터를 수집하고 정제하는 데 별도의 시간이 소요되지 않습니다.\n",
    "- 여기서 핵심은 유사도를 어떻게 정의하는가입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:16:23.093883Z",
     "start_time": "2022-01-11T13:16:18.550618Z"
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from datetime import datetime\n",
    "from tensorflow.keras.utils import get_file\n",
    "from itertools import combinations\n",
    "import matplotlib.pyplot as plt\n",
    "from google_drive_downloader import GoogleDriveDownloader as gdd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:16:30.848431Z",
     "start_time": "2022-01-11T13:16:23.127866Z"
    }
   },
   "outputs": [],
   "source": [
    "movies = pd.read_csv(\"./Datasets/MovieLens/movies.csv\", index_col=0)\n",
    "genres = pd.read_csv(\"./Datasets/MovieLens/genres.csv\", index_col=0)\n",
    "gdd.download_file_from_google_drive(file_id=\"15vsm-VWAC3Y-7jr7ROL_xy0ufkIfqSke\", dest_path=\"D:/ratings.csv\")\n",
    "ratings = pd.read_csv(\"D:/ratings.csv\")\n",
    "ratings = ratings.drop([\"Unnamed: 0\"], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 다음 5개 영화를 대상으로 알아보겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:10.433055Z",
     "start_time": "2022-01-11T13:18:10.396044Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>title</th>\n",
       "      <th>release_year</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>genre</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>215</td>\n",
       "      <td>Before Sunrise</td>\n",
       "      <td>1995</td>\n",
       "      <td>215</td>\n",
       "      <td>Drama</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>215</td>\n",
       "      <td>Before Sunrise</td>\n",
       "      <td>1995</td>\n",
       "      <td>215</td>\n",
       "      <td>Romance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2232</td>\n",
       "      <td>Cube</td>\n",
       "      <td>1997</td>\n",
       "      <td>2232</td>\n",
       "      <td>Horror</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2232</td>\n",
       "      <td>Cube</td>\n",
       "      <td>1997</td>\n",
       "      <td>2232</td>\n",
       "      <td>Mystery</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2232</td>\n",
       "      <td>Cube</td>\n",
       "      <td>1997</td>\n",
       "      <td>2232</td>\n",
       "      <td>Sci-Fi</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2232</td>\n",
       "      <td>Cube</td>\n",
       "      <td>1997</td>\n",
       "      <td>2232</td>\n",
       "      <td>Thriller</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6323</td>\n",
       "      <td>Identity</td>\n",
       "      <td>2003</td>\n",
       "      <td>6323</td>\n",
       "      <td>Crime</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>6323</td>\n",
       "      <td>Identity</td>\n",
       "      <td>2003</td>\n",
       "      <td>6323</td>\n",
       "      <td>Horror</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>6323</td>\n",
       "      <td>Identity</td>\n",
       "      <td>2003</td>\n",
       "      <td>6323</td>\n",
       "      <td>Mystery</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>6323</td>\n",
       "      <td>Identity</td>\n",
       "      <td>2003</td>\n",
       "      <td>6323</td>\n",
       "      <td>Thriller</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>8638</td>\n",
       "      <td>Before Sunset</td>\n",
       "      <td>2004</td>\n",
       "      <td>8638</td>\n",
       "      <td>Drama</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>8638</td>\n",
       "      <td>Before Sunset</td>\n",
       "      <td>2004</td>\n",
       "      <td>8638</td>\n",
       "      <td>Romance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>8957</td>\n",
       "      <td>Saw</td>\n",
       "      <td>2004</td>\n",
       "      <td>8957</td>\n",
       "      <td>Horror</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>8957</td>\n",
       "      <td>Saw</td>\n",
       "      <td>2004</td>\n",
       "      <td>8957</td>\n",
       "      <td>Mystery</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>8957</td>\n",
       "      <td>Saw</td>\n",
       "      <td>2004</td>\n",
       "      <td>8957</td>\n",
       "      <td>Thriller</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      id           title  release_year  movie_id     genre\n",
       "0    215  Before Sunrise          1995       215     Drama\n",
       "1    215  Before Sunrise          1995       215   Romance\n",
       "2   2232            Cube          1997      2232    Horror\n",
       "3   2232            Cube          1997      2232   Mystery\n",
       "4   2232            Cube          1997      2232    Sci-Fi\n",
       "5   2232            Cube          1997      2232  Thriller\n",
       "6   6323        Identity          2003      6323     Crime\n",
       "7   6323        Identity          2003      6323    Horror\n",
       "8   6323        Identity          2003      6323   Mystery\n",
       "9   6323        Identity          2003      6323  Thriller\n",
       "10  8638   Before Sunset          2004      8638     Drama\n",
       "11  8638   Before Sunset          2004      8638   Romance\n",
       "12  8957             Saw          2004      8957    Horror\n",
       "13  8957             Saw          2004      8957   Mystery\n",
       "14  8957             Saw          2004      8957  Thriller"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target = {8957: \"Saw\", 2232: \"Cube\", 6323: \"Identity\", 8638: \"Before Sunset\", 215: \"Before Sunrise\"}\n",
    "\n",
    "#각 영화의 세부정보를 가져옵니다.\n",
    "df1 = movies[movies[\"id\"].isin(target)]\n",
    "df2 = genres[genres[\"movie_id\"].isin(target)]\n",
    "\n",
    "pd.merge(df1, df2, left_on=\"id\", right_on=\"movie_id\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:16.743543Z",
     "start_time": "2022-01-11T13:18:15.810039Z"
    }
   },
   "outputs": [],
   "source": [
    "#id:title의 dictionary를 만듭니다.\n",
    "id2title = dict(zip(movies[\"id\"], movies[\"title\"]))\n",
    "ratings[\"title\"] = ratings[\"movie_id\"].map(id2title)\n",
    "\n",
    "#각 영화의 평점 데이터를 가져옵니다.\n",
    "ratings_target = ratings[ratings[\"movie_id\"].isin(target.keys())]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:17.713197Z",
     "start_time": "2022-01-11T13:18:17.689193Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>rated_at</th>\n",
       "      <th>title</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1099</th>\n",
       "      <td>11</td>\n",
       "      <td>2232</td>\n",
       "      <td>4.5</td>\n",
       "      <td>1230788590</td>\n",
       "      <td>Cube</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3638</th>\n",
       "      <td>31</td>\n",
       "      <td>8957</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1424736694</td>\n",
       "      <td>Saw</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4977</th>\n",
       "      <td>50</td>\n",
       "      <td>215</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1182347856</td>\n",
       "      <td>Before Sunrise</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5058</th>\n",
       "      <td>50</td>\n",
       "      <td>8638</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1182348573</td>\n",
       "      <td>Before Sunset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6156</th>\n",
       "      <td>56</td>\n",
       "      <td>215</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1003025941</td>\n",
       "      <td>Before Sunrise</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19997430</th>\n",
       "      <td>138472</td>\n",
       "      <td>8957</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1267119268</td>\n",
       "      <td>Saw</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19997639</th>\n",
       "      <td>138473</td>\n",
       "      <td>2232</td>\n",
       "      <td>4.5</td>\n",
       "      <td>1274234130</td>\n",
       "      <td>Cube</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19998940</th>\n",
       "      <td>138483</td>\n",
       "      <td>215</td>\n",
       "      <td>4.0</td>\n",
       "      <td>833831946</td>\n",
       "      <td>Before Sunrise</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19999482</th>\n",
       "      <td>138486</td>\n",
       "      <td>8957</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1207076826</td>\n",
       "      <td>Saw</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20000025</th>\n",
       "      <td>138493</td>\n",
       "      <td>2232</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1255817892</td>\n",
       "      <td>Cube</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>25260 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          user_id  movie_id  rating    rated_at           title\n",
       "1099           11      2232     4.5  1230788590            Cube\n",
       "3638           31      8957     0.5  1424736694             Saw\n",
       "4977           50       215     5.0  1182347856  Before Sunrise\n",
       "5058           50      8638     5.0  1182348573   Before Sunset\n",
       "6156           56       215     3.0  1003025941  Before Sunrise\n",
       "...           ...       ...     ...         ...             ...\n",
       "19997430   138472      8957     3.0  1267119268             Saw\n",
       "19997639   138473      2232     4.5  1274234130            Cube\n",
       "19998940   138483       215     4.0   833831946  Before Sunrise\n",
       "19999482   138486      8957     5.0  1207076826             Saw\n",
       "20000025   138493      2232     5.0  1255817892            Cube\n",
       "\n",
       "[25260 rows x 5 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ratings_target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:28.791127Z",
     "start_time": "2022-01-11T13:18:28.448346Z"
    }
   },
   "outputs": [],
   "source": [
    "ratings_target_set = ratings_target.groupby(\"user_id\")[\"title\"].apply(set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:28.960078Z",
     "start_time": "2022-01-11T13:18:28.951078Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id\n",
       "11                                 {Cube}\n",
       "31                                  {Saw}\n",
       "50        {Before Sunrise, Before Sunset}\n",
       "56                       {Before Sunrise}\n",
       "58                             {Identity}\n",
       "                       ...               \n",
       "138472                              {Saw}\n",
       "138473                             {Cube}\n",
       "138483                   {Before Sunrise}\n",
       "138486                              {Saw}\n",
       "138493                             {Cube}\n",
       "Name: title, Length: 16922, dtype: object"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ratings_target_set"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 5개의 영화를 모두 본 사용자만 뽑겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:32.593931Z",
     "start_time": "2022-01-11T13:18:32.578886Z"
    }
   },
   "outputs": [],
   "source": [
    "all_seen = ratings_target_set[ratings_target_set.map(lambda x : len(x)==5)].index\n",
    "ratings_all_seen = ratings_target[ratings_target[\"user_id\"].isin(all_seen)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:33.079899Z",
     "start_time": "2022-01-11T13:18:33.060915Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>rated_at</th>\n",
       "      <th>title</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>146330</th>\n",
       "      <td>982</td>\n",
       "      <td>215</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1195746062</td>\n",
       "      <td>Before Sunrise</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146993</th>\n",
       "      <td>982</td>\n",
       "      <td>2232</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1141652908</td>\n",
       "      <td>Cube</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147899</th>\n",
       "      <td>982</td>\n",
       "      <td>6323</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1193323976</td>\n",
       "      <td>Identity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148143</th>\n",
       "      <td>982</td>\n",
       "      <td>8638</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1195747667</td>\n",
       "      <td>Before Sunset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148181</th>\n",
       "      <td>982</td>\n",
       "      <td>8957</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1141659791</td>\n",
       "      <td>Saw</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19885812</th>\n",
       "      <td>137686</td>\n",
       "      <td>215</td>\n",
       "      <td>4.0</td>\n",
       "      <td>899415564</td>\n",
       "      <td>Before Sunrise</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19886569</th>\n",
       "      <td>137686</td>\n",
       "      <td>2232</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1087503954</td>\n",
       "      <td>Cube</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19887442</th>\n",
       "      <td>137686</td>\n",
       "      <td>6323</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1079896026</td>\n",
       "      <td>Identity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19887645</th>\n",
       "      <td>137686</td>\n",
       "      <td>8638</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1151393086</td>\n",
       "      <td>Before Sunset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19887671</th>\n",
       "      <td>137686</td>\n",
       "      <td>8957</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1110399064</td>\n",
       "      <td>Saw</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>905 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          user_id  movie_id  rating    rated_at           title\n",
       "146330        982       215     3.5  1195746062  Before Sunrise\n",
       "146993        982      2232     3.5  1141652908            Cube\n",
       "147899        982      6323     3.0  1193323976        Identity\n",
       "148143        982      8638     4.0  1195747667   Before Sunset\n",
       "148181        982      8957     3.0  1141659791             Saw\n",
       "...           ...       ...     ...         ...             ...\n",
       "19885812   137686       215     4.0   899415564  Before Sunrise\n",
       "19886569   137686      2232     4.0  1087503954            Cube\n",
       "19887442   137686      6323     3.5  1079896026        Identity\n",
       "19887645   137686      8638     4.0  1151393086   Before Sunset\n",
       "19887671   137686      8957     3.5  1110399064             Saw\n",
       "\n",
       "[905 rows x 5 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ratings_all_seen"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## User-Item Matrix\n",
    "- 이 데이터를 바탕으로 user-item matrix를 만듭니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:40.241566Z",
     "start_time": "2022-01-11T13:18:40.148318Z"
    }
   },
   "outputs": [],
   "source": [
    "ui = pd.pivot_table(ratings_all_seen, index=\"user_id\", columns=\"title\", values=\"rating\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:40.791161Z",
     "start_time": "2022-01-11T13:18:40.770135Z"
    }
   },
   "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>title</th>\n",
       "      <th>Before Sunrise</th>\n",
       "      <th>Before Sunset</th>\n",
       "      <th>Cube</th>\n",
       "      <th>Identity</th>\n",
       "      <th>Saw</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>982</th>\n",
       "      <td>3.5</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1165</th>\n",
       "      <td>4.5</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.5</td>\n",
       "      <td>2.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1288</th>\n",
       "      <td>4.0</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2650</th>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2669</th>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133329</th>\n",
       "      <td>5.0</td>\n",
       "      <td>4.5</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134290</th>\n",
       "      <td>4.0</td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135071</th>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>137277</th>\n",
       "      <td>3.5</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>137686</th>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>181 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "title    Before Sunrise  Before Sunset  Cube  Identity  Saw\n",
       "user_id                                                    \n",
       "982                 3.5            4.0   3.5       3.0  3.0\n",
       "1165                4.5            4.0   4.0       2.5  2.5\n",
       "1288                4.0            3.5   3.5       3.0  4.0\n",
       "2650                4.0            4.0   3.0       4.0  4.5\n",
       "2669                4.0            4.0   3.5       4.0  3.5\n",
       "...                 ...            ...   ...       ...  ...\n",
       "133329              5.0            4.5   5.0       3.0  4.0\n",
       "134290              4.0            4.5   4.0       3.0  3.5\n",
       "135071              3.5            3.5   3.5       3.5  3.5\n",
       "137277              3.5            4.0   3.5       3.5  3.5\n",
       "137686              4.0            4.0   4.0       3.5  3.5\n",
       "\n",
       "[181 rows x 5 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ui"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 사용자와 아이템의 편향 제거하기\n",
    "- 예를 들어 Saw에 똑같이 4점을 준 두 사용자가 동일한 정도로 그 영화를 좋아한다고 볼 수 없습니다.\n",
    "- 왜냐하면 사용자 A는 웬만해서는 4점을 주지 않는 성향이고 사용자 B는 비교적 점수를 후하게 주는 성향일 수 있기 때문입니다.\n",
    "- 또한 어떤 사용자가 Saw와 Before Sunset에 똑같이 4점을 주었다고 해도 Saw는 전반적으로 평가가 좋은 영화인 반면 Before Sunset은 이 사용자가 유일하게 높은 점수를 주었을 수도 있습니다.\n",
    "- 이런 왜곡을 피하기 위해 사용자와 아이템의 편향을 각각 제거해주어야 합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:49.904463Z",
     "start_time": "2022-01-11T13:18:49.880461Z"
    }
   },
   "outputs": [],
   "source": [
    "#사용자별 평균 평점\n",
    "avg_rating_user = ui.mean(axis=1).sort_values()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:50.327869Z",
     "start_time": "2022-01-11T13:18:50.312847Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id\n",
       "116900    1.6\n",
       "67346     1.9\n",
       "22122     2.4\n",
       "113668    2.4\n",
       "15701     2.4\n",
       "         ... \n",
       "87383     4.6\n",
       "77247     4.6\n",
       "101887    4.7\n",
       "80285     4.8\n",
       "74241     4.8\n",
       "Length: 181, dtype: float64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "avg_rating_user"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:18:52.828740Z",
     "start_time": "2022-01-11T13:18:52.360214Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAANYElEQVR4nO3dbaykZ1kH8P9lW1ChitgDaShli0ESYiIlm36pEEUCpSD4Ek2JEhIxGxMxEDVaQmLwm2gkfjFilQaivGmQSHhRGgUJiYDb0kJrWylYYm1tF4gBokELlx/ObD1ddvfM1jM71+z8fsnkzNz77Nnrnnsm/3M/c51nq7sDANN827oLAICTEVAAjCSgABhJQAEwkoACYCQBBcBI5y9zUFXdneSrSb6R5MHuPrzKogBgqYBa+JHu/uLKKgGAPc4koJZ20UUX9aFDh1bxrQE4x9x4441f7O6dE8eXDahO8qGq6iR/1N3Xne7gQ4cO5ejRo4+gTAC2TVV94WTjywbUld19b1U9IckNVXVHd3/0hH/gSJIjSXLppZf+v4oFgKW6+Lr73sXXB5K8J8kVJznmuu4+3N2Hd3a+ZacGAGdk34CqqsdU1YXH7yd5fpJbV10YANttmVN8T0zynqo6fvzbu/uvV1oVAFtv34Dq7s8n+cGzUAsAPMSVJAAYSUABMJKAAmAkAQXASCu51BEA555D177/YY/v/u0XrfTfs4MCYCQBBcBIAgqAkQQUACNpkgDgtE5sjjhb7KAAGElAATCSgAJgJAEFwEiaJABIsr5miFOxgwJgJAEFwEgCCoCRBBQAI2mSANgS05og9mMHBcBIAgqAkQQUACMJKABG0iQBcI7YtCaI/dhBATCSgAJgJAEFwEgCCoCRNEkADHWuNT2cKTsoAEYSUACMJKAAGElAATCSJgmAA7btzQ0HxQ4KgJEEFAAjCSgARhJQAIykSQLYOpoYNsPSO6iqOq+qPlVV71tlQQCQnNkpvlcnuX1VhQDAXksFVFVdkuRFSf5kteUAwK5ld1C/n+TXk3xzdaUAwP/Zt0miql6c5IHuvrGqfvg0xx1JciRJLr300oOqD0BTw5ZaZgd1ZZKXVNXdSd6Z5LlV9WcnHtTd13X34e4+vLOzc8BlArBt9g2o7n5td1/S3YeSXJPk77r751ZeGQBbzS/qAjDSGf2ibnd/JMlHVlIJAOzhShLAymly4JFwig+AkQQUACMJKABG8hkUbACf4bCN7KAAGElAATCSgAJgJAEFwEiaJOAAaGKAg2cHBcBIAgqAkQQUACMJKABG0iTBOUnTAmw+OygARhJQAIwkoAAYSUABMJImCUbR3AAcZwcFwEgCCoCRBBQAIwkoAEbSJMFSNC8AZ5sdFAAjCSgARhJQAIwkoAAYSZPEOUpTA7Dp7KAAGElAATCSgAJgJAEFwEiaJIbT7ABsKzsoAEYSUACMJKAAGElAATCSgBrgeCPEqb4CbKN9A6qqvr2qPllVt1TVbVX1W2ejMAC22zJt5l9P8tzu/lpVXZDkY1X1we7++IprA2CL7RtQ3d1JvrZ4eMHi1qssCgCW+gyqqs6rqpuTPJDkhu7+xEqrAmDrLRVQ3f2N7n5mkkuSXFFVP3DiMVV1pKqOVtXRY8eOHXCZm2G/ZgdNEADLO6Muvu7+jyQfSXLVSf7suu4+3N2Hd3Z2DqY6ALbWMl18O1X1uMX970jyvCR3rLguALbcMl18Fyd5a1Wdl91A+/Puft9qywJg2y3TxffpJJefhVoA4CGuJHEGNDsAnD0CCoCRBBQAIwkoAEYSUACMtNUB5YoPAHNtdUABMJeAAmAkAQXASFsVUD5TAtgcWxVQAGwOAQXASAIKgJEEFAAjnVMB5RdtAc4d51RAAXDuEFAAjCSgABhJQAEwkoACYCQBBcBIAgqAkQQUACMJKABG2qiAcqUIgO2xUQEFwPYQUACMJKAAGElAATCSgAJgJAEFwEgCCoCRBBQAIwkoAEYSUACMJKAAGElAATCSgAJgJAEFwEgCCoCR9g2oqnpyVX24qm6vqtuq6tVnozAAttv5SxzzYJJf7e6bqurCJDdW1Q3d/U8rrg2ALbbvDqq77+vumxb3v5rk9iRPWnVhAGy3M/oMqqoOJbk8ySdWUg0ALCwdUFX12CTvTvKa7v7KSf78SFUdraqjx44dO8gaAdhCSwVUVV2Q3XB6W3f/5cmO6e7ruvtwdx/e2dk5yBoB2ELLdPFVkjcnub2737j6kgBguR3UlUlenuS5VXXz4nb1iusCYMvt22be3R9LUmehFgB4yEZcSeLQte9fdwkAnGUbEVAAbB8BBcBIAgqAkQQUACMJKABGElAAjCSgABhJQAEwkoACYCQBBcBIAgqAkQQUACMJKABGGh1QrmIOsL1GBxQA20tAATCSgAJgJAEFwEgCCoCRBBQAIwkoAEYSUACMJKAAGElAATCSgAJgJAEFwEgCCoCRBBQAIwkoAEYSUACMJKAAGElAATCSgAJgJAEFwEgCCoCRBBQAIwkoAEYSUACMJKAAGGnfgKqq66vqgaq69WwUBADJcjuotyS5asV1AMDD7BtQ3f3RJF8+C7UAwEN8BgXASAcWUFV1pKqOVtXRY8eOHdS3BWBLHVhAdfd13X24uw/v7Owc1LcFYEs5xQfASMu0mb8jyT8keXpV3VNVr1x9WQBsu/P3O6C7X3Y2CgGAvZziA2AkAQXASAIKgJEEFAAjCSgARhJQAIwkoAAYSUABMJKAAmAkAQXASAIKgJEEFAAjCSgARhJQAIwkoAAYSUABMJKAAmAkAQXASAIKgJEEFAAjCSgARhJQAIwkoAAYSUABMJKAAmAkAQXASAIKgJEEFAAjCSgARhJQAIwkoAAYSUABMJKAAmAkAQXASAIKgJEEFAAjCSgARhJQAIwkoAAYSUABMNJSAVVVV1XVnVV1V1Vdu+qiAGDfgKqq85L8QZIXJnlGkpdV1TNWXRgA222ZHdQVSe7q7s93938neWeSl662LAC23TIB9aQk/7rn8T2LMQBYmeru0x9Q9dNJXtDdv7B4/PIkV3T3L59w3JEkRxYPn57kzgOo76IkXzyA77Nu5jGLecxiHrOsYx5P6e6dEwfPX+Iv3pPkyXseX5Lk3hMP6u7rklz3iMs7iao62t2HD/J7roN5zGIes5jHLJPmscwpvn9M8rSquqyqHpXkmiTvXW1ZAGy7fXdQ3f1gVb0qyd8kOS/J9d1928orA2CrLXOKL939gSQfWHEtJ3OgpwzXyDxmMY9ZzGOWMfPYt0kCANbBpY4AGGlsQG3q5ZWq6slV9eGqur2qbquqVy/GX19V/1ZVNy9uV6+71v1U1d1V9ZlFvUcXY4+vqhuq6rOLr9+z7jpPp6qevuc5v7mqvlJVr9mE9aiq66vqgaq6dc/YKZ//qnrt4v1yZ1W9YD1Vf6tTzON3q+qOqvp0Vb2nqh63GD9UVf+1Z13etLbC9zjFHE75GtqwtXjXnjncXVU3L8bXvxbdPe6W3WaMzyV5apJHJbklyTPWXdeStV+c5FmL+xcm+efsXiLq9Ul+bd31neFc7k5y0Qljv5Pk2sX9a5O8Yd11nsF8zkvy70mesgnrkeQ5SZ6V5Nb9nv/Fa+yWJI9Octni/XPeuudwmnk8P8n5i/tv2DOPQ3uPm3I7xRxO+hratLU44c9/L8lvTlmLqTuojb28Unff1903Le5/NcntObeuvPHSJG9d3H9rkh9fXyln7EeTfK67v7DuQpbR3R9N8uUThk/1/L80yTu7++vd/S9J7sru+2jtTjaP7v5Qdz+4ePjx7P5+5VinWItT2ai1OK6qKsnPJHnHWS3qNKYG1DlxeaWqOpTk8iSfWAy9anFK4/rpp8YWOsmHqurGxZVCkuSJ3X1fshvGSZ6wturO3DV5+Jtv09YjOfXzv8nvmZ9P8sE9jy+rqk9V1d9X1bPXVdSSTvYa2tS1eHaS+7v7s3vG1roWUwOqTjK2Ue2GVfXYJO9O8pru/kqSP0zyfUmemeS+7G6lp7uyu5+V3SvZ/1JVPWfdBT1Si18yf0mSv1gMbeJ6nM5Gvmeq6nVJHkzytsXQfUku7e7Lk/xKkrdX1Xetq759nOo1tJFrkeRlefgPcGtfi6kBtdTllaaqqguyG05v6+6/TJLuvr+7v9Hd30zyxxmy5T+d7r538fWBJO/Jbs33V9XFSbL4+sD6KjwjL0xyU3ffn2zmeiyc6vnfuPdMVb0iyYuT/GwvPvRYnBb70uL+jdn9/Ob711flqZ3mNbSJa3F+kp9M8q7jYxPWYmpAbezllRbncd+c5PbufuOe8Yv3HPYTSW498e9OUlWPqaoLj9/P7ofat2Z3HV6xOOwVSf5qPRWesYf9dLhp67HHqZ7/9ya5pqoeXVWXJXlakk+uob6lVNVVSX4jyUu6+z/3jO/U7v9Bl6p6anbn8fn1VHl6p3kNbdRaLDwvyR3dfc/xgRFrse6uktN0m1yd3Q64zyV53brrOYO6fyi72/lPJ7l5cbs6yZ8m+cxi/L1JLl53rfvM46nZ7US6Jcltx9cgyfcm+dskn118ffy6a11iLt+Z5EtJvnvP2Pj1yG6g3pfkf7L7U/krT/f8J3nd4v1yZ5IXrrv+feZxV3Y/pzn+HnnT4tifWrzebklyU5IfW3f9p5nDKV9Dm7QWi/G3JPnFE45d+1q4kgQAI009xQfAlhNQAIwkoAAYSUABMJKAAmAkAQXASAIKgJEEFAAj/S+t3Iu/MYYywQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "\n",
    "ax.bar(x=range(len(avg_rating_user)), height=avg_rating_user)\n",
    "\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:01.009822Z",
     "start_time": "2022-01-11T13:19:01.004817Z"
    }
   },
   "outputs": [],
   "source": [
    "#아이템별 평균 평점\n",
    "avg_rating_item = ui.mean(axis=0).sort_values()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:01.245136Z",
     "start_time": "2022-01-11T13:19:01.224135Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "title\n",
       "Saw               3.298343\n",
       "Identity          3.400552\n",
       "Cube              3.577348\n",
       "Before Sunset     3.881215\n",
       "Before Sunrise    3.950276\n",
       "dtype: float64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "avg_rating_item"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:09.802738Z",
     "start_time": "2022-01-11T13:19:09.660232Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAVNUlEQVR4nO3df7Bc5X3f8fcHIWw8OKGurosiAYpT1Y1xbYxvsQmJQ4mbAqYmbZkBjx0CnVbFxbXdxkn9o8XY08mP6TTxgBwUNXFsWmLiFkxlLMbQ2gRoASMpQiCwY9khgwoNF6gFKgRb+Ns/9ohslpV2dbXSfaR9v2Z27jnPefac73m02s+es+eem6pCkqTWHLHQBUiSNIwBJUlqkgElSWqSASVJapIBJUlq0pELteElS5bUihUrFmrzkqRGbNy48YmqmhlsX7CAWrFiBRs2bFiozUuSGpHkT4e1e4pPktQkA0qS1CQDSpLUJANKktSksQMqyaIkf5TkpiHLkuTKJNuSbElyymTLlCRNm305gvoA8NAelp0NrOweq4Cr97MuSdKUGyugkiwH3gH8zh66nAdcUz13A8cmWTqhGiVJU2jcI6hPAb8M/GAPy5cBj/TNb+/a/pIkq5JsSLJhbm5uX+qUJE2ZkQGV5Fzg8arauLduQ9pe8oemqmptVc1W1ezMzEt+aViSpBeNcwR1OvDOJA8D1wFnJvnPA322A8f3zS8HHp1IhZKkqTTyVkdV9RHgIwBJzgA+VFXvGei2DnhfkuuAtwA7quqxyZYqSe1b8eEvL3QJB83Dv/aOA7r+ed+LL8mlAFW1BlgPnANsA54FLplIdZKkqbVPAVVVtwG3ddNr+toLuGyShUmSppt3kpAkNWnB/tyGpEOP36/oYPIISpLUJANKktQkA0qS1CQDSpLUJANKktQkA0qS1CQDSpLUJANKktQkA0qS1CQDSpLUJANKktQkA0qS1CQDSpLUJANKktQkA0qS1CQDSpLUJANKktQkA0qS1CT/5LvUmZY/Z+6fMtehYuQRVJKXJ/l6kvuSbE3yiSF9zkiyI8nm7nH5gSlXkjQtxjmCeh44s6p2JlkM3Jnk5qq6e6DfHVV17uRLlCRNo5EBVVUF7OxmF3ePOpBFSZI01kUSSRYl2Qw8DtxaVfcM6XZadxrw5iQn7WE9q5JsSLJhbm5u/lVLkg57YwVUVb1QVScDy4FTk7x+oMsm4MSqeiNwFXDjHtaztqpmq2p2ZmZm/lVLkg57+3SZeVV9F7gNOGug/emq2tlNrwcWJ1kyoRolSVNonKv4ZpIc200fDbwd+MZAn+OSpJs+tVvvkxOvVpI0Nca5im8p8Lkki+gFzxeq6qYklwJU1RrgfOC9SXYBzwEXdhdXSJI0L+NcxbcFeNOQ9jV906uB1ZMtTZI0zbyTxBSYljskgHdJkA4n3otPktQkA0qS1CQDSpLUJANKktQkA0qS1KRD/io+r1CTpMOTR1CSpCYZUJKkJhlQkqQmGVCSpCYZUJKkJhlQkqQmGVCSpCYZUJKkJhlQkqQmGVCSpCYZUJKkJhlQkqQmGVCSpCYZUJKkJo0MqCQvT/L1JPcl2ZrkE0P6JMmVSbYl2ZLklANTriRpWozz96CeB86sqp1JFgN3Jrm5qu7u63M2sLJ7vAW4uvspSdK8jDyCqp6d3ezi7lED3c4Drun63g0cm2TpZEuVJE2Tsb6DSrIoyWbgceDWqrpnoMsy4JG++e1d2+B6ViXZkGTD3NzcPEuWJE2DsQKqql6oqpOB5cCpSV4/0CXDnjZkPWuraraqZmdmZva5WEnS9Ninq/iq6rvAbcBZA4u2A8f3zS8HHt2fwiRJ022cq/hmkhzbTR8NvB34xkC3dcBF3dV8bwV2VNVjky5WkjQ9xrmKbynwuSSL6AXaF6rqpiSXAlTVGmA9cA6wDXgWuOQA1StJmhIjA6qqtgBvGtK+pm+6gMsmW5okaZp5JwlJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTRgZUkuOTfC3JQ0m2JvnAkD5nJNmRZHP3uPzAlCtJmhZHjtFnF/CLVbUpySuBjUluraoHB/rdUVXnTr5ESdI0GnkEVVWPVdWmbvoZ4CFg2YEuTJI03fbpO6gkK4A3AfcMWXxakvuS3JzkpD08f1WSDUk2zM3N7Xu1kqSpMXZAJTkGuB74YFU9PbB4E3BiVb0RuAq4cdg6qmptVc1W1ezMzMw8S5YkTYOxAirJYnrhdG1V3TC4vKqerqqd3fR6YHGSJROtVJI0Vca5ii/A7wIPVdVv7KHPcV0/kpzarffJSRYqSZou41zFdzrw88D9STZ3bR8FTgCoqjXA+cB7k+wCngMurKqafLmSpGkxMqCq6k4gI/qsBlZPqihJkryThCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkjAyrJ8Um+luShJFuTfGBInyS5Msm2JFuSnHJgypUkTYsjx+izC/jFqtqU5JXAxiS3VtWDfX3OBlZ2j7cAV3c/JUmal5FHUFX1WFVt6qafAR4Clg10Ow+4pnruBo5NsnTi1UqSpsY+fQeVZAXwJuCegUXLgEf65rfz0hCTJGlsYwdUkmOA64EPVtXTg4uHPKWGrGNVkg1JNszNze1bpZKkqTJWQCVZTC+crq2qG4Z02Q4c3ze/HHh0sFNVra2q2aqanZmZmU+9kqQpMc5VfAF+F3ioqn5jD93WARd1V/O9FdhRVY9NsE5J0pQZ5yq+04GfB+5Psrlr+yhwAkBVrQHWA+cA24BngUsmXqkkaaqMDKiqupPh3zH19yngskkVJUmSd5KQJDXJgJIkNcmAkiQ1yYCSJDXJgJIkNcmAkiQ1yYCSJDXJgJIkNcmAkiQ1yYCSJDXJgJIkNcmAkiQ1yYCSJDXJgJIkNcmAkiQ1yYCSJDXJgJIkNcmAkiQ1yYCSJDXJgJIkNcmAkiQ1aWRAJflMkseTPLCH5Wck2ZFkc/e4fPJlSpKmzZFj9PkssBq4Zi997qiqcydSkSRJjHEEVVW3A08dhFokSXrRpL6DOi3JfUluTnLShNYpSZpi45ziG2UTcGJV7UxyDnAjsHJYxySrgFUAJ5xwwgQ2LUk6XO33EVRVPV1VO7vp9cDiJEv20HdtVc1W1ezMzMz+blqSdBjb74BKclySdNOndut8cn/XK0mabiNP8SX5PHAGsCTJduDjwGKAqloDnA+8N8ku4DngwqqqA1axJGkqjAyoqnrXiOWr6V2GLknSxHgnCUlSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpNGBlSSzyR5PMkDe1ieJFcm2ZZkS5JTJl+mJGnajHME9VngrL0sPxtY2T1WAVfvf1mSpGk3MqCq6nbgqb10OQ+4pnruBo5NsnRSBUqSptMkvoNaBjzSN7+9a5Mkad4mEVAZ0lZDOyarkmxIsmFubm4Cm5YkHa4mEVDbgeP75pcDjw7rWFVrq2q2qmZnZmYmsGlJ0uFqEgG1Driou5rvrcCOqnpsAuuVJE2xI0d1SPJ54AxgSZLtwMeBxQBVtQZYD5wDbAOeBS45UMVKkqbHyICqqneNWF7AZROrSJIkvJOEJKlRBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJBpQkqUkGlCSpSQaUJKlJYwVUkrOSfDPJtiQfHrL8jCQ7kmzuHpdPvlRJ0jQ5clSHJIuATwN/F9gO3JtkXVU9OND1jqo69wDUKEmaQuMcQZ0KbKuq71TV94DrgPMObFmSpGk3TkAtAx7pm9/etQ06Lcl9SW5OctKwFSVZlWRDkg1zc3PzKFeSNC3GCagMaauB+U3AiVX1RuAq4MZhK6qqtVU1W1WzMzMz+1SoJGm6jBNQ24Hj++aXA4/2d6iqp6tqZze9HlicZMnEqpQkTZ1xAupeYGWSH01yFHAhsK6/Q5LjkqSbPrVb75OTLlaSND1GXsVXVbuSvA/4CrAI+ExVbU1yabd8DXA+8N4ku4DngAuravA0oCRJYxsZUPDiabv1A21r+qZXA6snW5okaZp5JwlJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTDChJUpMMKElSkwwoSVKTxgqoJGcl+WaSbUk+PGR5klzZLd+S5JTJlypJmiYjAyrJIuDTwNnA64B3JXndQLezgZXdYxVw9YTrlCRNmXGOoE4FtlXVd6rqe8B1wHkDfc4Drqmeu4FjkyydcK2SpCly5Bh9lgGP9M1vB94yRp9lwGP9nZKsoneEBbAzyTf3qdp2LAGeONgbza8f7C3ut4M+To7RaI7ReByn0SY4RicOaxwnoDKkrebRh6paC6wdY5tNS7KhqmYXuo7WOU6jOUajOUbjORzHaZxTfNuB4/vmlwOPzqOPJEljGyeg7gVWJvnRJEcBFwLrBvqsAy7qruZ7K7Cjqh4bXJEkSeMaeYqvqnYleR/wFWAR8Jmq2prk0m75GmA9cA6wDXgWuOTAldyEQ/405UHiOI3mGI3mGI3nsBunVL3kqyJJkhacd5KQJDXJgJIkNcmAGiLJx5Js7W7btDnJ4O99HbaS7NxD+2eTnD/PdZ6c5Jy++XfuvmVWkp8bcmeSQ1qS45Jcl+TbSR5Msj7J39hL/4eTLDmYNY4jyQvd6/++JJuS/MQYz3l/koeSXDvhWl6R5Nok9yd5IMmdSY6Z5Db2su2P7qF9KsYnyY8k+a+TWNe+Guf3oKZKktOAc4FTqur57o3jqAUu61B3MjBL72Iaqmodf3El6M8BNwEPLkRhk5YkwBeBz1XVhV3bycBfA/54AUubj+eq6mSAJH8P+FXgp0c8558DZ1fVn4yzgSRHVtWuMbp+APizqvpb3fNeC3x/nG1MwEeBXxnSftiPT7f9R4F5fTjdXx5BvdRS4Imqeh6gqp6oqkeTXJ7k3u7TydrukvpXJ9kIkOSNSSrJCd38t5O8YgH3Y790+7e6OwL4MvDqvmVvTvKHSTYm+cru21oluS3Jryf5epI/TvJT3a8mfBK4oPu0eUGSi7t1/wTwTuDfd8t+LMmmvu2s3D2+h5C/A3y/u7oVgKraDCxKctPutm7/L+573i914/b1JH+96zOT5PrudXdvktMP0j4M80PA/909k+SXupq2JPlE17YGeA2wLsm/TPKqJDd2fe5O8oau3xXd/6FbgGvG3M+lwP/ePVNV3+w+QK5I8kBfXR9KckU3/ZLXY9d+Ute2uattZdf+nr72306yKMmvAUd3bXs76jmcxufiJP8lyZeAW/rXsS9jt5exGl9V+eh7AMcAm+l92v0t4Ke79lf19flPwN/vprfSe3G+j97vjL2b3m077lrofZnn/u/sfv5D4FZ6v1rwI8B36X2KWgz8L2Cm63cBvV89ALgN+A/d9DnAf++mLwZW923jxXngs8D5fcu+BpzcTf8K8C8Wekz2cfzeD/zmkPYzgJv65lcDF3fTDwMf66Yv2t0P+H3gJ7vpE4CHDvK+vND9X/gGsAN4c9f+s/QuaQ69D7k3AW/r25cl3fRVwMe76TOBzd30FcBG4Ohx95PeUfjjwF3AvwNWdu0rgAf6+n0IuGLE6/Eq4N3d9FHA0cCPA18CFnftvwVc1P9/YorG52J6N1941eA69nXs9vfhKb4BVbUzyZuBn6L3afgP0vu+5Jkkvwy8AngVvWD6Er0369OBt9F7Qz2L3gvzjgUof5LeBny+ql4AHk3y1a79tcDrgVuTQC/A+n8p+4bu50Z6L+x99TvAJUn+Fb3wO3Ue6zgUfb7v5292028HXteNM8APJXllVT1zkGrqP4V1Gr1P86+n9wb8s8Afdf2OofeXDG4feP5PAv8IoKq+muSvJvnhbtm6qnqumx65n1W1Oclruu2+Hbi3q2n3OvZk2OvxLuBjSZYDN1TVt5L8DPDmbr3Qe+N9fMS6D9fxAbi1qp4a0n9SYzcWA2qI7k35NuC2JPcD/wx4AzBbVY90h8gv77rfQS/MTgT+G/Cv6d2H8CYOfcN+SS7A1qo6bQ/Peb77+QLze31dD3wc+CqwsaqenMc6FtJWhp+v38VfPqX+8oHlNWT6COC0vjeqBVNVd6X3fewMvdfAr1bVb4942t7u0fn/+trG2s+q2knvDfWGJD+g96n/D9j7uL7k9VhVv5/kHuAdwFeS/JOu1s9V1UdG7NOeajtsxmfI9vu3MfGx2xu/gxqQ5LW7z6t2TgZ233X9ifSujOl/A7odeA/wrar6AfAUvRfG/zwI5R5ItwMXdufhl9I7moTeWMx0n85IsjjJSSPW9QzwynGWVdWf07trydXA7+1H/Qvlq8DLkvzT3Q1J/ja9I83XJXlZ9yn5Zwaed0Hfz7u66VvonTrevZ6TD1TRoyT5m/T24Ul6/z7/uPu/QJJlSV495Gm30zvlTZIz6H23+/SQfiP3M8npSf5KN30Uvb9N96fAnwGv7o4+XkbvAqdR+/Ia4DtVdSW9i3XeAPwP4Pzd+9F9P7T7DtvfT7J4xDoPm/HZm3mM3X7xCOqljgGuSnIsvU+92+j9iZDvAvfTO4d87+7OVfVwd1i7+/D9TmB5Vb34hekh6ov0zovfT+/7uD8EqKrvpXe5+ZXdG+2RwKfoHTnsydeADyfZTO9Kp37XAf8xyfvpfRf1beBaet+B3TKxvTlIqqqS/APgU92p4T+n95r5IPAFYAvwLf7i9M9uL+s+mR4BvKtrez/w6SRb6I3z7cClB3of+hzd/ZtB71PyL3RnF25J8uPAXd1rfye9D2mDp3WuAH6vq/9Z4Bf2sJ1x9vPHgKvT2+ARwJeB67vx/iRwD/An9L4PGuUC4D1Jvg/8H+CTVfVUkn/T7dsR9K6Au4zem/xaYEuSTVX17ikYn73Z17HbL97qSM1J8iHgh6vq3y50LZIWjkdQakqSL9L7NHjmQtciaWF5BCVJapIXSUiSmmRASZKaZEBJkppkQEmSmmRASZKa9P8BpitT8bVR1GMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "\n",
    "ax.bar(x=avg_rating_item.index, height=avg_rating_item)\n",
    "\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 사용자 편향과 아이템 편향을 반영하여 user-item matrix를 조정합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:10.778753Z",
     "start_time": "2022-01-11T13:19:10.754789Z"
    }
   },
   "outputs": [],
   "source": [
    "avg_rating = ui.mean().mean()\n",
    "\n",
    "#사용자 편향을 계산합니다.\n",
    "user_bias = ui.mean(axis=1) - avg_rating\n",
    "#아이템 편향을 계산합니다.\n",
    "item_bias = ui.mean(axis=0) - avg_rating\n",
    "\n",
    "adj_ui = ui.sub(user_bias, axis=0).sub(item_bias, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:11.891829Z",
     "start_time": "2022-01-11T13:19:11.868829Z"
    }
   },
   "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>title</th>\n",
       "      <th>Before Sunrise</th>\n",
       "      <th>Before Sunset</th>\n",
       "      <th>Cube</th>\n",
       "      <th>Identity</th>\n",
       "      <th>Saw</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2650</th>\n",
       "      <td>3.392818</td>\n",
       "      <td>3.461878</td>\n",
       "      <td>2.765746</td>\n",
       "      <td>3.942541</td>\n",
       "      <td>4.544751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80501</th>\n",
       "      <td>2.492818</td>\n",
       "      <td>2.061878</td>\n",
       "      <td>4.365746</td>\n",
       "      <td>4.542541</td>\n",
       "      <td>4.644751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65304</th>\n",
       "      <td>3.592818</td>\n",
       "      <td>4.661878</td>\n",
       "      <td>3.965746</td>\n",
       "      <td>3.142541</td>\n",
       "      <td>2.744751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13431</th>\n",
       "      <td>3.792818</td>\n",
       "      <td>3.861878</td>\n",
       "      <td>3.165746</td>\n",
       "      <td>4.342541</td>\n",
       "      <td>2.944751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>130767</th>\n",
       "      <td>4.792818</td>\n",
       "      <td>4.361878</td>\n",
       "      <td>3.665746</td>\n",
       "      <td>2.842541</td>\n",
       "      <td>2.444751</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "title    Before Sunrise  Before Sunset      Cube  Identity       Saw\n",
       "user_id                                                             \n",
       "2650           3.392818       3.461878  2.765746  3.942541  4.544751\n",
       "80501          2.492818       2.061878  4.365746  4.542541  4.644751\n",
       "65304          3.592818       4.661878  3.965746  3.142541  2.744751\n",
       "13431          3.792818       3.861878  3.165746  4.342541  2.944751\n",
       "130767         4.792818       4.361878  3.665746  2.842541  2.444751"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adj_ui.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Similarity 구하기\n",
    "- 거리 기반 유사도 : 유클리드 거리 유사도 등\n",
    "- 방향 기반 유사도 : 코사인 유사도 등"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## User-Based Collaborative Filtering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:19.074221Z",
     "start_time": "2022-01-11T13:19:19.065476Z"
    }
   },
   "outputs": [],
   "source": [
    "#adj_ui에서 5개의 샘플만 랜덤으로 뽑겠습니다.\n",
    "sample5 = adj_ui.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:19.761419Z",
     "start_time": "2022-01-11T13:19:19.745414Z"
    }
   },
   "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>title</th>\n",
       "      <th>Before Sunrise</th>\n",
       "      <th>Before Sunset</th>\n",
       "      <th>Cube</th>\n",
       "      <th>Identity</th>\n",
       "      <th>Saw</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>120572</th>\n",
       "      <td>4.092818</td>\n",
       "      <td>4.161878</td>\n",
       "      <td>3.465746</td>\n",
       "      <td>3.642541</td>\n",
       "      <td>2.744751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40285</th>\n",
       "      <td>2.692818</td>\n",
       "      <td>2.261878</td>\n",
       "      <td>4.065746</td>\n",
       "      <td>4.242541</td>\n",
       "      <td>4.844751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110048</th>\n",
       "      <td>3.292818</td>\n",
       "      <td>4.361878</td>\n",
       "      <td>1.665746</td>\n",
       "      <td>4.842541</td>\n",
       "      <td>3.944751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71970</th>\n",
       "      <td>3.092818</td>\n",
       "      <td>3.661878</td>\n",
       "      <td>4.465746</td>\n",
       "      <td>3.142541</td>\n",
       "      <td>3.744751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133329</th>\n",
       "      <td>3.992818</td>\n",
       "      <td>3.561878</td>\n",
       "      <td>4.365746</td>\n",
       "      <td>2.542541</td>\n",
       "      <td>3.644751</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "title    Before Sunrise  Before Sunset      Cube  Identity       Saw\n",
       "user_id                                                             \n",
       "120572         4.092818       4.161878  3.465746  3.642541  2.744751\n",
       "40285          2.692818       2.261878  4.065746  4.242541  4.844751\n",
       "110048         3.292818       4.361878  1.665746  4.842541  3.944751\n",
       "71970          3.092818       3.661878  4.465746  3.142541  3.744751\n",
       "133329         3.992818       3.561878  4.365746  2.542541  3.644751"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:26.622240Z",
     "start_time": "2022-01-11T13:19:26.613107Z"
    }
   },
   "outputs": [],
   "source": [
    "#유클리드 거리 기반 유사도를 계산하는 함수를 만듭니다.\n",
    "def euc_sim(x, y):\n",
    "    return 1/(1 + np.sqrt(np.sum((x - y)**2)))\n",
    "\n",
    "#코사인 유사도를 계산하는 함수를 만듭니다.\n",
    "def cos_sim(x, y):\n",
    "    return np.sum(x*y)/(np.linalg.norm(x)*np.linalg.norm(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-11T13:19:26.825856Z",
     "start_time": "2022-01-11T13:19:26.802479Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "120572 |  40285 | euc_sim : 0.234 | cos_sim : 0.946\n",
      "120572 | 110048 | euc_sim : 0.277 | cos_sim : 0.985\n",
      "120572 |  71970 | euc_sim : 0.348 | cos_sim : 0.973\n",
      "120572 | 133329 | euc_sim : 0.359 | cos_sim : 0.980\n",
      " 40285 | 110048 | euc_sim : 0.226 | cos_sim : 0.926\n",
      " 40285 |  71970 | euc_sim : 0.316 | cos_sim : 0.941\n",
      " 40285 | 133329 | euc_sim : 0.264 | cos_sim : 0.924\n",
      "110048 |  71970 | euc_sim : 0.229 | cos_sim : 0.887\n",
      "110048 | 133329 | euc_sim : 0.212 | cos_sim : 0.874\n",
      " 71970 | 133329 | euc_sim : 0.477 | cos_sim : 0.996\n"
     ]
    }
   ],
   "source": [
    "for user1, user2 in combinations(sample5.index, 2):\n",
    "    sim1 = euc_sim(sample5.loc[user1], sample5.loc[user2])\n",
    "    sim2 = cos_sim(sample5.loc[user1], sample5.loc[user2])\n",
    "    print(f\"{user1:>6d} | {user2:>6d} | euc_sim : {sim1:>4.3f} | cos_sim : {sim2:>4.3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Item-Based Collaborative Filtering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "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>user_id</th>\n",
       "      <th>982</th>\n",
       "      <th>1165</th>\n",
       "      <th>1288</th>\n",
       "      <th>2650</th>\n",
       "      <th>2669</th>\n",
       "      <th>3676</th>\n",
       "      <th>3907</th>\n",
       "      <th>5725</th>\n",
       "      <th>6099</th>\n",
       "      <th>6373</th>\n",
       "      <th>...</th>\n",
       "      <th>131180</th>\n",
       "      <th>131736</th>\n",
       "      <th>132509</th>\n",
       "      <th>132603</th>\n",
       "      <th>133133</th>\n",
       "      <th>133329</th>\n",
       "      <th>134290</th>\n",
       "      <th>135071</th>\n",
       "      <th>137277</th>\n",
       "      <th>137686</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>title</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Before Sunrise</th>\n",
       "      <td>3.392818</td>\n",
       "      <td>4.292818</td>\n",
       "      <td>3.692818</td>\n",
       "      <td>3.392818</td>\n",
       "      <td>3.492818</td>\n",
       "      <td>3.292818</td>\n",
       "      <td>3.892818</td>\n",
       "      <td>3.392818</td>\n",
       "      <td>3.192818</td>\n",
       "      <td>2.992818</td>\n",
       "      <td>...</td>\n",
       "      <td>4.092818</td>\n",
       "      <td>2.792818</td>\n",
       "      <td>3.492818</td>\n",
       "      <td>3.692818</td>\n",
       "      <td>3.392818</td>\n",
       "      <td>3.992818</td>\n",
       "      <td>3.492818</td>\n",
       "      <td>3.292818</td>\n",
       "      <td>3.192818</td>\n",
       "      <td>3.492818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Before Sunset</th>\n",
       "      <td>3.961878</td>\n",
       "      <td>3.861878</td>\n",
       "      <td>3.261878</td>\n",
       "      <td>3.461878</td>\n",
       "      <td>3.561878</td>\n",
       "      <td>3.361878</td>\n",
       "      <td>4.461878</td>\n",
       "      <td>3.961878</td>\n",
       "      <td>4.261878</td>\n",
       "      <td>3.061878</td>\n",
       "      <td>...</td>\n",
       "      <td>3.661878</td>\n",
       "      <td>2.861878</td>\n",
       "      <td>4.061878</td>\n",
       "      <td>3.261878</td>\n",
       "      <td>3.461878</td>\n",
       "      <td>3.561878</td>\n",
       "      <td>4.061878</td>\n",
       "      <td>3.361878</td>\n",
       "      <td>3.761878</td>\n",
       "      <td>3.561878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cube</th>\n",
       "      <td>3.765746</td>\n",
       "      <td>4.165746</td>\n",
       "      <td>3.565746</td>\n",
       "      <td>2.765746</td>\n",
       "      <td>3.365746</td>\n",
       "      <td>3.165746</td>\n",
       "      <td>3.765746</td>\n",
       "      <td>3.765746</td>\n",
       "      <td>3.065746</td>\n",
       "      <td>3.365746</td>\n",
       "      <td>...</td>\n",
       "      <td>3.465746</td>\n",
       "      <td>4.165746</td>\n",
       "      <td>3.865746</td>\n",
       "      <td>3.565746</td>\n",
       "      <td>3.765746</td>\n",
       "      <td>4.365746</td>\n",
       "      <td>3.865746</td>\n",
       "      <td>3.665746</td>\n",
       "      <td>3.565746</td>\n",
       "      <td>3.865746</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Identity</th>\n",
       "      <td>3.442541</td>\n",
       "      <td>2.842541</td>\n",
       "      <td>3.242541</td>\n",
       "      <td>3.942541</td>\n",
       "      <td>4.042541</td>\n",
       "      <td>4.342541</td>\n",
       "      <td>3.442541</td>\n",
       "      <td>2.942541</td>\n",
       "      <td>3.742541</td>\n",
       "      <td>4.042541</td>\n",
       "      <td>...</td>\n",
       "      <td>3.142541</td>\n",
       "      <td>3.842541</td>\n",
       "      <td>4.042541</td>\n",
       "      <td>3.742541</td>\n",
       "      <td>3.442541</td>\n",
       "      <td>2.542541</td>\n",
       "      <td>3.042541</td>\n",
       "      <td>3.842541</td>\n",
       "      <td>3.742541</td>\n",
       "      <td>3.542541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Saw</th>\n",
       "      <td>3.544751</td>\n",
       "      <td>2.944751</td>\n",
       "      <td>4.344751</td>\n",
       "      <td>4.544751</td>\n",
       "      <td>3.644751</td>\n",
       "      <td>3.944751</td>\n",
       "      <td>2.544751</td>\n",
       "      <td>4.044751</td>\n",
       "      <td>3.844751</td>\n",
       "      <td>4.644751</td>\n",
       "      <td>...</td>\n",
       "      <td>3.744751</td>\n",
       "      <td>4.444751</td>\n",
       "      <td>2.644751</td>\n",
       "      <td>3.844751</td>\n",
       "      <td>4.044751</td>\n",
       "      <td>3.644751</td>\n",
       "      <td>3.644751</td>\n",
       "      <td>3.944751</td>\n",
       "      <td>3.844751</td>\n",
       "      <td>3.644751</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 181 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "user_id           982       1165      1288      2650      2669      3676    \\\n",
       "title                                                                        \n",
       "Before Sunrise  3.392818  4.292818  3.692818  3.392818  3.492818  3.292818   \n",
       "Before Sunset   3.961878  3.861878  3.261878  3.461878  3.561878  3.361878   \n",
       "Cube            3.765746  4.165746  3.565746  2.765746  3.365746  3.165746   \n",
       "Identity        3.442541  2.842541  3.242541  3.942541  4.042541  4.342541   \n",
       "Saw             3.544751  2.944751  4.344751  4.544751  3.644751  3.944751   \n",
       "\n",
       "user_id           3907      5725      6099      6373    ...    131180  \\\n",
       "title                                                   ...             \n",
       "Before Sunrise  3.892818  3.392818  3.192818  2.992818  ...  4.092818   \n",
       "Before Sunset   4.461878  3.961878  4.261878  3.061878  ...  3.661878   \n",
       "Cube            3.765746  3.765746  3.065746  3.365746  ...  3.465746   \n",
       "Identity        3.442541  2.942541  3.742541  4.042541  ...  3.142541   \n",
       "Saw             2.544751  4.044751  3.844751  4.644751  ...  3.744751   \n",
       "\n",
       "user_id           131736    132509    132603    133133    133329    134290  \\\n",
       "title                                                                        \n",
       "Before Sunrise  2.792818  3.492818  3.692818  3.392818  3.992818  3.492818   \n",
       "Before Sunset   2.861878  4.061878  3.261878  3.461878  3.561878  4.061878   \n",
       "Cube            4.165746  3.865746  3.565746  3.765746  4.365746  3.865746   \n",
       "Identity        3.842541  4.042541  3.742541  3.442541  2.542541  3.042541   \n",
       "Saw             4.444751  2.644751  3.844751  4.044751  3.644751  3.644751   \n",
       "\n",
       "user_id           135071    137277    137686  \n",
       "title                                         \n",
       "Before Sunrise  3.292818  3.192818  3.492818  \n",
       "Before Sunset   3.361878  3.761878  3.561878  \n",
       "Cube            3.665746  3.565746  3.865746  \n",
       "Identity        3.842541  3.742541  3.542541  \n",
       "Saw             3.944751  3.844751  3.644751  \n",
       "\n",
       "[5 rows x 181 columns]"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adj_ui.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before Sunrise |  Before Sunset | euc_sim : 0.112 | cos_sim : 0.987\n",
      "Before Sunrise |           Cube | euc_sim : 0.063 | cos_sim : 0.954\n",
      "Before Sunrise |       Identity | euc_sim : 0.064 | cos_sim : 0.955\n",
      "Before Sunrise |            Saw | euc_sim : 0.056 | cos_sim : 0.947\n",
      " Before Sunset |           Cube | euc_sim : 0.063 | cos_sim : 0.954\n",
      " Before Sunset |       Identity | euc_sim : 0.064 | cos_sim : 0.955\n",
      " Before Sunset |            Saw | euc_sim : 0.056 | cos_sim : 0.947\n",
      "          Cube |       Identity | euc_sim : 0.072 | cos_sim : 0.965\n",
      "          Cube |            Saw | euc_sim : 0.071 | cos_sim : 0.969\n",
      "      Identity |            Saw | euc_sim : 0.072 | cos_sim : 0.972\n"
     ]
    }
   ],
   "source": [
    "for item1, item2 in combinations(adj_ui.T.index, 2):\n",
    "    sim1 = euc_sim(adj_ui.T.loc[item1], adj_ui.T.loc[item2])\n",
    "    sim2 = cos_sim(adj_ui.T.loc[item1], adj_ui.T.loc[item2])\n",
    "    print(f\"{item1:>14s} | {item2:>14s} | euc_sim : {sim1:>4.3f} | cos_sim : {sim2:>4.3f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "165px"
   },
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}