Content_based_recommendation_system.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Content Based recommendation system"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.metrics.pairwise import cosine_similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "      <th>episodes</th>\n",
       "      <th>rating</th>\n",
       "      <th>members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32281</td>\n",
       "      <td>Kimi no Na wa.</td>\n",
       "      <td>Drama, Romance, School, Supernatural</td>\n",
       "      <td>Movie</td>\n",
       "      <td>1</td>\n",
       "      <td>9.37</td>\n",
       "      <td>200630</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5114</td>\n",
       "      <td>Fullmetal Alchemist: Brotherhood</td>\n",
       "      <td>Action, Adventure, Drama, Fantasy, Magic, Mili...</td>\n",
       "      <td>TV</td>\n",
       "      <td>64</td>\n",
       "      <td>9.26</td>\n",
       "      <td>793665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28977</td>\n",
       "      <td>Gintama°</td>\n",
       "      <td>Action, Comedy, Historical, Parody, Samurai, S...</td>\n",
       "      <td>TV</td>\n",
       "      <td>51</td>\n",
       "      <td>9.25</td>\n",
       "      <td>114262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9253</td>\n",
       "      <td>Steins;Gate</td>\n",
       "      <td>Sci-Fi, Thriller</td>\n",
       "      <td>TV</td>\n",
       "      <td>24</td>\n",
       "      <td>9.17</td>\n",
       "      <td>673572</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9969</td>\n",
       "      <td>Gintama&amp;#039;</td>\n",
       "      <td>Action, Comedy, Historical, Parody, Samurai, S...</td>\n",
       "      <td>TV</td>\n",
       "      <td>51</td>\n",
       "      <td>9.16</td>\n",
       "      <td>151266</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   anime_id                              name  \\\n",
       "0     32281                    Kimi no Na wa.   \n",
       "1      5114  Fullmetal Alchemist: Brotherhood   \n",
       "2     28977                          Gintama°   \n",
       "3      9253                       Steins;Gate   \n",
       "4      9969                     Gintama&#039;   \n",
       "\n",
       "                                               genre   type episodes  rating  \\\n",
       "0               Drama, Romance, School, Supernatural  Movie        1    9.37   \n",
       "1  Action, Adventure, Drama, Fantasy, Magic, Mili...     TV       64    9.26   \n",
       "2  Action, Comedy, Historical, Parody, Samurai, S...     TV       51    9.25   \n",
       "3                                   Sci-Fi, Thriller     TV       24    9.17   \n",
       "4  Action, Comedy, Historical, Parody, Samurai, S...     TV       51    9.16   \n",
       "\n",
       "   members  \n",
       "0   200630  \n",
       "1   793665  \n",
       "2   114262  \n",
       "3   673572  \n",
       "4   151266  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"D:/Courses/big data/project/Anime/anime.csv\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exploring the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 12294 entries, 0 to 12293\n",
      "Data columns (total 7 columns):\n",
      "anime_id    12294 non-null int64\n",
      "name        12294 non-null object\n",
      "genre       12232 non-null object\n",
      "type        12269 non-null object\n",
      "episodes    12294 non-null object\n",
      "rating      12064 non-null float64\n",
      "members     12294 non-null int64\n",
      "dtypes: float64(1), int64(2), object(4)\n",
      "memory usage: 672.4+ KB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "anime_id      0\n",
       "name          0\n",
       "genre        62\n",
       "type         25\n",
       "episodes      0\n",
       "rating      230\n",
       "members       0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### For Cotent based recommender system we will be using { Name, Genre, Type} column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# preprocessing\n",
    "# filling NaN \"Type\" with unknown\n",
    "df.loc[(df['type'].isnull()), 'type'] = 'Unknown'\n",
    "# filling NaN \"Genre\" with unknown\n",
    "df.loc[(df['genre'].isnull()), 'genre'] = 'Unknown'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "anime_id      0\n",
       "name          0\n",
       "genre         0\n",
       "type          0\n",
       "episodes      0\n",
       "rating      230\n",
       "members       0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Removing punctuation\n",
    "#df['name'] = df['name'].str.replace('[^\\w\\s]','')\n",
    "#df['name'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32281</td>\n",
       "      <td>Kimi no Na wa.</td>\n",
       "      <td>Drama, Romance, School, Supernatural</td>\n",
       "      <td>Movie</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5114</td>\n",
       "      <td>Fullmetal Alchemist: Brotherhood</td>\n",
       "      <td>Action, Adventure, Drama, Fantasy, Magic, Mili...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28977</td>\n",
       "      <td>Gintama°</td>\n",
       "      <td>Action, Comedy, Historical, Parody, Samurai, S...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9253</td>\n",
       "      <td>Steins;Gate</td>\n",
       "      <td>Sci-Fi, Thriller</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9969</td>\n",
       "      <td>Gintama&amp;#039;</td>\n",
       "      <td>Action, Comedy, Historical, Parody, Samurai, S...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   anime_id                              name  \\\n",
       "0     32281                    Kimi no Na wa.   \n",
       "1      5114  Fullmetal Alchemist: Brotherhood   \n",
       "2     28977                          Gintama°   \n",
       "3      9253                       Steins;Gate   \n",
       "4      9969                     Gintama&#039;   \n",
       "\n",
       "                                               genre   type  \n",
       "0               Drama, Romance, School, Supernatural  Movie  \n",
       "1  Action, Adventure, Drama, Fantasy, Magic, Mili...     TV  \n",
       "2  Action, Comedy, Historical, Parody, Samurai, S...     TV  \n",
       "3                                   Sci-Fi, Thriller     TV  \n",
       "4  Action, Comedy, Historical, Parody, Samurai, S...     TV  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dropping unwanted columns\n",
    "df.drop(['rating', 'members', 'episodes'], axis=1, inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Converting to lower case\n",
    "df['genre'] = df['genre'].apply(lambda x: \" \".join(x.lower() for x in x.split()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# tokenization of keywords\n",
    "#from textblob import TextBlob\n",
    "#df['genre'] = df['genre'].apply(lambda x: TextBlob(x).words)\n",
    "#df['type'] = df['type'].apply(lambda x: TextBlob(x).words)\n",
    "#df['episodes'] = df['episodes'].apply(lambda x: TextBlob(x).words)\n",
    "#df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['bag_of_words'] = \"\"\n",
    "df.loc[:,'bag_of_words'] = df.loc[:, 'genre'] +', '+ df.loc[:, 'type']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "      <th>bag_of_words</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>32281</td>\n",
       "      <td>Kimi no Na wa.</td>\n",
       "      <td>drama, romance, school, supernatural</td>\n",
       "      <td>Movie</td>\n",
       "      <td>drama, romance, school, supernatural, Movie</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5114</td>\n",
       "      <td>Fullmetal Alchemist: Brotherhood</td>\n",
       "      <td>action, adventure, drama, fantasy, magic, mili...</td>\n",
       "      <td>TV</td>\n",
       "      <td>action, adventure, drama, fantasy, magic, mili...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28977</td>\n",
       "      <td>Gintama°</td>\n",
       "      <td>action, comedy, historical, parody, samurai, s...</td>\n",
       "      <td>TV</td>\n",
       "      <td>action, comedy, historical, parody, samurai, s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9253</td>\n",
       "      <td>Steins;Gate</td>\n",
       "      <td>sci-fi, thriller</td>\n",
       "      <td>TV</td>\n",
       "      <td>sci-fi, thriller, TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9969</td>\n",
       "      <td>Gintama&amp;#039;</td>\n",
       "      <td>action, comedy, historical, parody, samurai, s...</td>\n",
       "      <td>TV</td>\n",
       "      <td>action, comedy, historical, parody, samurai, s...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   anime_id                              name  \\\n",
       "0     32281                    Kimi no Na wa.   \n",
       "1      5114  Fullmetal Alchemist: Brotherhood   \n",
       "2     28977                          Gintama°   \n",
       "3      9253                       Steins;Gate   \n",
       "4      9969                     Gintama&#039;   \n",
       "\n",
       "                                               genre   type  \\\n",
       "0               drama, romance, school, supernatural  Movie   \n",
       "1  action, adventure, drama, fantasy, magic, mili...     TV   \n",
       "2  action, comedy, historical, parody, samurai, s...     TV   \n",
       "3                                   sci-fi, thriller     TV   \n",
       "4  action, comedy, historical, parody, samurai, s...     TV   \n",
       "\n",
       "                                        bag_of_words  \n",
       "0        drama, romance, school, supernatural, Movie  \n",
       "1  action, adventure, drama, fantasy, magic, mili...  \n",
       "2  action, comedy, historical, parody, samurai, s...  \n",
       "3                               sci-fi, thriller, TV  \n",
       "4  action, comedy, historical, parody, samurai, s...  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data after preprocessing\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using sklearn TFIDF implementation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer = TfidfVectorizer()\n",
    "tfidf = vectorizer.fit_transform(df[\"bag_of_words\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "bag = vectorizer.get_feature_names()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Using Cosine to obtain similaritier between two animes\n",
    "cosine_sim = cosine_similarity(tfidf, tfidf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12294, 12294)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cosine_sim.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# recommender function\n",
    "def recommendation(name, cosine_sim):\n",
    "    recommended = np.array([])\n",
    "    index_ = df.loc[(df['name'] == name)].index\n",
    "    score = cosine_sim[index_]\n",
    "    ind = np.argpartition(score[0], -10)[-11:]\n",
    "    ind = ind[np.argsort(score[0][ind])]\n",
    "    print(score[0][ind])\n",
    "    #for i in ind:\n",
    "     #   recommended = np.append(recommended, df.loc[(df.index == i)]['name'])        \n",
    "    return ind"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.9394034  0.9394034  0.9394034  0.9394034  0.96155633 0.96155633\n",
      " 0.96155633 0.96155633 0.96155633 1.         1.        ]\n",
      "[ 1854   959  3592  2524 10953   290   101   268   795   255   288]\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>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1854</th>\n",
       "      <td>332</td>\n",
       "      <td>Dokidoki Densetsu: Mahoujin Guru Guru</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>959</th>\n",
       "      <td>331</td>\n",
       "      <td>Mahoujin Guru Guru</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3592</th>\n",
       "      <td>1018</td>\n",
       "      <td>Majutsushi Orphen: Revenge</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2524</th>\n",
       "      <td>1164</td>\n",
       "      <td>Mahou Senshi Louie</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10953</th>\n",
       "      <td>28833</td>\n",
       "      <td>Chain Chronicle: Haecceitas no Hikari</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>290</th>\n",
       "      <td>31741</td>\n",
       "      <td>Magi: Sinbad no Bouken (TV)</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>18115</td>\n",
       "      <td>Magi: The Kingdom of Magic</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>14513</td>\n",
       "      <td>Magi: The Labyrinth of Magic</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>795</th>\n",
       "      <td>8086</td>\n",
       "      <td>Densetsu no Yuusha no Densetsu</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>255</th>\n",
       "      <td>22043</td>\n",
       "      <td>Fairy Tail (2014)</td>\n",
       "      <td>action, adventure, comedy, fantasy, magic, sho...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>6702</td>\n",
       "      <td>Fairy Tail</td>\n",
       "      <td>action, adventure, comedy, fantasy, magic, sho...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       anime_id                                   name  \\\n",
       "1854        332  Dokidoki Densetsu: Mahoujin Guru Guru   \n",
       "959         331                     Mahoujin Guru Guru   \n",
       "3592       1018             Majutsushi Orphen: Revenge   \n",
       "2524       1164                     Mahou Senshi Louie   \n",
       "10953     28833  Chain Chronicle: Haecceitas no Hikari   \n",
       "290       31741            Magi: Sinbad no Bouken (TV)   \n",
       "101       18115             Magi: The Kingdom of Magic   \n",
       "268       14513           Magi: The Labyrinth of Magic   \n",
       "795        8086         Densetsu no Yuusha no Densetsu   \n",
       "255       22043                      Fairy Tail (2014)   \n",
       "288        6702                             Fairy Tail   \n",
       "\n",
       "                                                   genre type  \n",
       "1854          adventure, comedy, fantasy, magic, shounen   TV  \n",
       "959           adventure, comedy, fantasy, magic, shounen   TV  \n",
       "3592          adventure, comedy, fantasy, magic, shounen   TV  \n",
       "2524          adventure, comedy, fantasy, magic, shounen   TV  \n",
       "10953         action, adventure, fantasy, magic, shounen   TV  \n",
       "290           action, adventure, fantasy, magic, shounen   TV  \n",
       "101           action, adventure, fantasy, magic, shounen   TV  \n",
       "268           action, adventure, fantasy, magic, shounen   TV  \n",
       "795           action, adventure, fantasy, magic, shounen   TV  \n",
       "255    action, adventure, comedy, fantasy, magic, sho...   TV  \n",
       "288    action, adventure, comedy, fantasy, magic, sho...   TV  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predicting movies similar to \"Fairy Tail\"\n",
    "indices = recommendation('Fairy Tail', cosine_sim)\n",
    "print(indices)\n",
    "df[['anime_id','name', 'genre', 'type']].loc[indices]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TF - IDF implementation without library"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>words</th>\n",
       "      <th>tf</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>school</td>\n",
       "      <td>1220.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Movie</td>\n",
       "      <td>2348.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>drama</td>\n",
       "      <td>2016.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>supernatural</td>\n",
       "      <td>1037.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>romance</td>\n",
       "      <td>1464.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          words      tf\n",
       "0        school  1220.0\n",
       "1         Movie  2348.0\n",
       "2         drama  2016.0\n",
       "3  supernatural  1037.0\n",
       "4       romance  1464.0"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculating TF\n",
    "tf = df['bag_of_words'].apply(lambda x: pd.value_counts(x.split(\",\"))).sum(axis = 0).reset_index()\n",
    "tf.columns = ['words','tf']\n",
    "\n",
    "# removing duplicate words\n",
    "for i,word1 in enumerate(tf['words']):\n",
    "    tf.loc[i, 'words']= word1.strip()\n",
    "tf2 = {}\n",
    "for i,word1 in enumerate(tf['words']):\n",
    "    if word1 not in tf2.keys():\n",
    "        tf2[word1] = tf.loc[i, 'tf']\n",
    "    else:\n",
    "        tf2[word1] += tf.loc[i, 'tf']\n",
    "tf = pd.DataFrame.from_dict(tf2, orient='index').reset_index()\n",
    "tf.columns = ['words','tf']\n",
    "tf.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Calculating IDF\n",
    "N = df.shape[0]\n",
    "idf = {}\n",
    "for i,word in enumerate(tf['words']):\n",
    "    tf.loc[i, 'idf'] = np.log(N/(len(df[df['bag_of_words'].str.contains(word)])))\n",
    "    idf[word] = np.log(N/(len(df[df['bag_of_words'].str.contains(word)])))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>words</th>\n",
       "      <th>school</th>\n",
       "      <th>Movie</th>\n",
       "      <th>drama</th>\n",
       "      <th>supernatural</th>\n",
       "      <th>romance</th>\n",
       "      <th>adventure</th>\n",
       "      <th>TV</th>\n",
       "      <th>military</th>\n",
       "      <th>action</th>\n",
       "      <th>magic</th>\n",
       "      <th>...</th>\n",
       "      <th>Music</th>\n",
       "      <th>cars</th>\n",
       "      <th>ONA</th>\n",
       "      <th>kids</th>\n",
       "      <th>shoujo ai</th>\n",
       "      <th>unknown</th>\n",
       "      <th>Unknown</th>\n",
       "      <th>hentai</th>\n",
       "      <th>yaoi</th>\n",
       "      <th>yuri</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.462052</td>\n",
       "      <td>0.331109</td>\n",
       "      <td>0.361599</td>\n",
       "      <td>0.494556</td>\n",
       "      <td>0.425588</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.225999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.206943</td>\n",
       "      <td>0.147192</td>\n",
       "      <td>0.420303</td>\n",
       "      <td>0.182944</td>\n",
       "      <td>0.345018</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.147192</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.182944</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.392512</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.147192</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.182944</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 51 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "words    school     Movie     drama  supernatural   romance  adventure  \\\n",
       "0      0.462052  0.331109  0.361599      0.494556  0.425588   0.000000   \n",
       "1      0.000000  0.000000  0.225999      0.000000  0.000000   0.206943   \n",
       "2      0.000000  0.000000  0.000000      0.000000  0.000000   0.000000   \n",
       "3      0.000000  0.000000  0.000000      0.000000  0.000000   0.000000   \n",
       "4      0.000000  0.000000  0.000000      0.000000  0.000000   0.000000   \n",
       "\n",
       "words        TV  military    action     magic  ...  Music  cars  ONA  kids  \\\n",
       "0      0.000000  0.000000  0.000000  0.000000  ...    0.0   0.0  0.0   0.0   \n",
       "1      0.147192  0.420303  0.182944  0.345018  ...    0.0   0.0  0.0   0.0   \n",
       "2      0.147192  0.000000  0.182944  0.000000  ...    0.0   0.0  0.0   0.0   \n",
       "3      0.392512  0.000000  0.000000  0.000000  ...    0.0   0.0  0.0   0.0   \n",
       "4      0.147192  0.000000  0.182944  0.000000  ...    0.0   0.0  0.0   0.0   \n",
       "\n",
       "words  shoujo ai  unknown  Unknown  hentai  yaoi  yuri  \n",
       "0            0.0      0.0      0.0     0.0   0.0   0.0  \n",
       "1            0.0      0.0      0.0     0.0   0.0   0.0  \n",
       "2            0.0      0.0      0.0     0.0   0.0   0.0  \n",
       "3            0.0      0.0      0.0     0.0   0.0   0.0  \n",
       "4            0.0      0.0      0.0     0.0   0.0   0.0  \n",
       "\n",
       "[5 rows x 51 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating tfidf vectors\n",
    "tff = pd.DataFrame(columns=tf['words'])\n",
    "for i in range(N):\n",
    "    dict_ = dict.fromkeys(tf['words'], 0)\n",
    "    for word in df['bag_of_words'][i].split(\", \"):\n",
    "        dict_[word] = (1/len(df['bag_of_words'][i].split(\", \"))) * idf[word]\n",
    "    tff = tff.append(dict_, ignore_index=True)\n",
    "tff.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "cosine_sim_ = cosine_similarity(tff, tff)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Comparing both with and without library tfidf implementation and recommending on the basis of contents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.94857288 0.94857288 0.94857288 0.94857288 0.97758854 0.97758854\n",
      " 0.97758854 0.97758854 0.97758854 1.         1.        ]\n",
      "[ 2524  1854  3592   959   101   268   795 10953   290   288   255]\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>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2524</th>\n",
       "      <td>1164</td>\n",
       "      <td>Mahou Senshi Louie</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1854</th>\n",
       "      <td>332</td>\n",
       "      <td>Dokidoki Densetsu: Mahoujin Guru Guru</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3592</th>\n",
       "      <td>1018</td>\n",
       "      <td>Majutsushi Orphen: Revenge</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>959</th>\n",
       "      <td>331</td>\n",
       "      <td>Mahoujin Guru Guru</td>\n",
       "      <td>adventure, comedy, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>18115</td>\n",
       "      <td>Magi: The Kingdom of Magic</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>14513</td>\n",
       "      <td>Magi: The Labyrinth of Magic</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>795</th>\n",
       "      <td>8086</td>\n",
       "      <td>Densetsu no Yuusha no Densetsu</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10953</th>\n",
       "      <td>28833</td>\n",
       "      <td>Chain Chronicle: Haecceitas no Hikari</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>290</th>\n",
       "      <td>31741</td>\n",
       "      <td>Magi: Sinbad no Bouken (TV)</td>\n",
       "      <td>action, adventure, fantasy, magic, shounen</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>6702</td>\n",
       "      <td>Fairy Tail</td>\n",
       "      <td>action, adventure, comedy, fantasy, magic, sho...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>255</th>\n",
       "      <td>22043</td>\n",
       "      <td>Fairy Tail (2014)</td>\n",
       "      <td>action, adventure, comedy, fantasy, magic, sho...</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       anime_id                                   name  \\\n",
       "2524       1164                     Mahou Senshi Louie   \n",
       "1854        332  Dokidoki Densetsu: Mahoujin Guru Guru   \n",
       "3592       1018             Majutsushi Orphen: Revenge   \n",
       "959         331                     Mahoujin Guru Guru   \n",
       "101       18115             Magi: The Kingdom of Magic   \n",
       "268       14513           Magi: The Labyrinth of Magic   \n",
       "795        8086         Densetsu no Yuusha no Densetsu   \n",
       "10953     28833  Chain Chronicle: Haecceitas no Hikari   \n",
       "290       31741            Magi: Sinbad no Bouken (TV)   \n",
       "288        6702                             Fairy Tail   \n",
       "255       22043                      Fairy Tail (2014)   \n",
       "\n",
       "                                                   genre type  \n",
       "2524          adventure, comedy, fantasy, magic, shounen   TV  \n",
       "1854          adventure, comedy, fantasy, magic, shounen   TV  \n",
       "3592          adventure, comedy, fantasy, magic, shounen   TV  \n",
       "959           adventure, comedy, fantasy, magic, shounen   TV  \n",
       "101           action, adventure, fantasy, magic, shounen   TV  \n",
       "268           action, adventure, fantasy, magic, shounen   TV  \n",
       "795           action, adventure, fantasy, magic, shounen   TV  \n",
       "10953         action, adventure, fantasy, magic, shounen   TV  \n",
       "290           action, adventure, fantasy, magic, shounen   TV  \n",
       "288    action, adventure, comedy, fantasy, magic, sho...   TV  \n",
       "255    action, adventure, comedy, fantasy, magic, sho...   TV  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predicting movies similar to \"Fairy Tail\"\n",
    "indices = recommendation('Fairy Tail', cosine_sim_)\n",
    "print(indices)\n",
    "df[['anime_id','name', 'genre', 'type']].loc[indices]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating user profile and recommending movies "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "user_ratings = pd.read_csv(\"D:/Courses/big data/project/Anime/rating.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>anime_id</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rating</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>-1</th>\n",
       "      <td>1476496</td>\n",
       "      <td>1476496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16649</td>\n",
       "      <td>16649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>23150</td>\n",
       "      <td>23150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>41453</td>\n",
       "      <td>41453</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>104291</td>\n",
       "      <td>104291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>282806</td>\n",
       "      <td>282806</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>637775</td>\n",
       "      <td>637775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1375287</td>\n",
       "      <td>1375287</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1646019</td>\n",
       "      <td>1646019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1254096</td>\n",
       "      <td>1254096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>955715</td>\n",
       "      <td>955715</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        user_id  anime_id\n",
       "rating                   \n",
       "-1      1476496   1476496\n",
       " 1        16649     16649\n",
       " 2        23150     23150\n",
       " 3        41453     41453\n",
       " 4       104291    104291\n",
       " 5       282806    282806\n",
       " 6       637775    637775\n",
       " 7      1375287   1375287\n",
       " 8      1646019   1646019\n",
       " 9      1254096   1254096\n",
       " 10      955715    955715"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_ratings.groupby(user_ratings[\"rating\"]).count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>anime_id</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rating</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16649</td>\n",
       "      <td>16649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>23150</td>\n",
       "      <td>23150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>41453</td>\n",
       "      <td>41453</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>104291</td>\n",
       "      <td>104291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>282806</td>\n",
       "      <td>282806</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>637775</td>\n",
       "      <td>637775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1375287</td>\n",
       "      <td>1375287</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1646019</td>\n",
       "      <td>1646019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1254096</td>\n",
       "      <td>1254096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>955715</td>\n",
       "      <td>955715</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        user_id  anime_id\n",
       "rating                   \n",
       "1         16649     16649\n",
       "2         23150     23150\n",
       "3         41453     41453\n",
       "4        104291    104291\n",
       "5        282806    282806\n",
       "6        637775    637775\n",
       "7       1375287   1375287\n",
       "8       1646019   1646019\n",
       "9       1254096   1254096\n",
       "10       955715    955715"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# dropping missing ratings\n",
    "user_ratings.drop(user_ratings[user_ratings['rating']==-1].index, axis=0, inplace=True)\n",
    "user_ratings.groupby(user_ratings[\"rating\"]).count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data generation\n",
    "import random\n",
    "def get_test_data():\n",
    "    # random.seed(123)\n",
    "    #user = random.randint(1, max(user_ratings['user_id']))    \n",
    "    #print(\"User selected for predictions UserId:\", (user))\n",
    "    # generated in Collaborative filtering for comparing the results\n",
    "    test_data_users = [36112, 23604, 26867, 30558, 55900, 21256, 40027, 67337, 69934, 61587, 18357, 16907, 57904, 69741, 14220, 37539, 53879, 45611, 50095, 53348, 36318, 8157, 29255, 67689, 57239, 7740, 12496, 27932, 60977, 36703, 8520, 62520, 46291, 41439, 27950, 71103, 70437, 27763, 64931, 49575, 12316, 14039, 13887, 11891, 1133, 58817, 58038, 23077, 6422, 36364]\n",
    "    test_data_animes = pd.read_csv(\"D:/Courses/big data/project/BigDataProject/data/test_data_animes.csv\")\n",
    "    return test_data_users, test_data_animes['test_animes'].to_list()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "training_data = user_ratings.copy(deep=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[-0.09289936320646325,\n",
       " 0.0,\n",
       " 0.9757438666884701,\n",
       " -0.31400373450634067,\n",
       " 0.9389952707601639,\n",
       " 0.26278530738396266,\n",
       " 0.2703974287149612,\n",
       " 0.0,\n",
       " -0.10531352637211089,\n",
       " -0.3650978959376411,\n",
       " -0.6858629391126686,\n",
       " 0.12387383763795945,\n",
       " 0.0,\n",
       " -0.5190062618972052,\n",
       " 0.7635268848032309,\n",
       " -0.8418389226100651,\n",
       " 0.20805358008067795,\n",
       " -1.650319499604124,\n",
       " 0.0,\n",
       " 0.831702656985294,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.2566956088599612,\n",
       " 0.24445147155442398,\n",
       " 0.0,\n",
       " -1.0707696182994215,\n",
       " -0.296420744289998,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " -1.3277148713042386,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " -0.8144182117382499,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.46870773180012504,\n",
       " 0.0,\n",
       " -0.6580465856193073,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0,\n",
       " 0.0]"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = training_data.loc[training_data['user_id']==user]\n",
    "    \n",
    "# shifting the rating with center at zero\n",
    "mean = data.mean()['rating']\n",
    "#data['likeness'] = 0\n",
    "data = data.assign(likeness=0)\n",
    "data.loc[:, \"likeness\"] = data.loc[:, \"rating\"] - mean\n",
    "\n",
    "# creating user table for all the movies rated by him/her\n",
    "#user_prof = pd.DataFrame(columns=bag)\n",
    "user_prof = pd.DataFrame(columns=tf['words'])\n",
    "for id in data['anime_id']:\n",
    "    dict_ = dict.fromkeys(tf['words'], 0)\n",
    "    for word in df.loc[df['anime_id']==id]['bag_of_words'].values[0].split(', '):\n",
    "        #if word in tf['words']:\n",
    "            #dict_[word] = tf.loc[(tf.words == word)]['tfidf'][tf.loc[(tf.words == word), 'tfidf'].index[0]]\n",
    "        dict_[word] = (1/len(df.loc[df['anime_id']==id]['bag_of_words'].values[0].split(', '))) * idf[word]\n",
    "    #print(dict_)    \n",
    "    user_prof = user_prof.append(dict_, ignore_index=True)\n",
    "# Creating user prfile by doing weighted sum of the attributes he/she has rated\n",
    "user_profile = []\n",
    "for col in user_prof.columns:\n",
    "    user_profile.append(user_prof[col] @ data['likeness'].values)\n",
    "user_profile    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>anime_id</th>\n",
       "      <th>rating</th>\n",
       "      <th>likeness</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1149</th>\n",
       "      <td>8</td>\n",
       "      <td>269</td>\n",
       "      <td>9</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1150</th>\n",
       "      <td>8</td>\n",
       "      <td>355</td>\n",
       "      <td>9</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1151</th>\n",
       "      <td>8</td>\n",
       "      <td>6702</td>\n",
       "      <td>10</td>\n",
       "      <td>1.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1152</th>\n",
       "      <td>8</td>\n",
       "      <td>7088</td>\n",
       "      <td>5</td>\n",
       "      <td>-3.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1153</th>\n",
       "      <td>8</td>\n",
       "      <td>7593</td>\n",
       "      <td>9</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1154</th>\n",
       "      <td>8</td>\n",
       "      <td>8630</td>\n",
       "      <td>8</td>\n",
       "      <td>-0.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1155</th>\n",
       "      <td>8</td>\n",
       "      <td>10620</td>\n",
       "      <td>6</td>\n",
       "      <td>-2.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1156</th>\n",
       "      <td>8</td>\n",
       "      <td>10793</td>\n",
       "      <td>10</td>\n",
       "      <td>1.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1157</th>\n",
       "      <td>8</td>\n",
       "      <td>11241</td>\n",
       "      <td>7</td>\n",
       "      <td>-1.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1158</th>\n",
       "      <td>8</td>\n",
       "      <td>11757</td>\n",
       "      <td>9</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1159</th>\n",
       "      <td>8</td>\n",
       "      <td>13759</td>\n",
       "      <td>9</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1160</th>\n",
       "      <td>8</td>\n",
       "      <td>15583</td>\n",
       "      <td>9</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      user_id  anime_id  rating  likeness\n",
       "1149        8       269       9  0.666667\n",
       "1150        8       355       9  0.666667\n",
       "1151        8      6702      10  1.666667\n",
       "1152        8      7088       5 -3.333333\n",
       "1153        8      7593       9  0.666667\n",
       "1154        8      8630       8 -0.333333\n",
       "1155        8     10620       6 -2.333333\n",
       "1156        8     10793      10  1.666667\n",
       "1157        8     11241       7 -1.333333\n",
       "1158        8     11757       9  0.666667\n",
       "1159        8     13759       9  0.666667\n",
       "1160        8     15583       9  0.666667"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Frequency')"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# create figure and axis\n",
    "fig, ax = plt.subplots()\n",
    "ax.hist(user_ratings['rating'])\n",
    "ax.set_title('Ratings their counts')\n",
    "ax.set_xlabel('Ratings')\n",
    "ax.set_ylabel('Frequency')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_user_profile(user):\n",
    "    data = training_data.loc[training_data['user_id']==user]\n",
    "    \n",
    "    # shifting the rating with center at zero\n",
    "    mean = data.mean()['rating']\n",
    "    #data['likeness'] = 0\n",
    "    data = data.assign(likeness=0)\n",
    "    data.loc[:, \"likeness\"] = data.loc[:, \"rating\"] - mean\n",
    "    \n",
    "    # creating user table for all the movies rated by him/her\n",
    "    #user_prof = pd.DataFrame(columns=bag)\n",
    "    user_prof = pd.DataFrame(columns=tf['words'])\n",
    "    for id in data['anime_id']:\n",
    "        dict_ = dict.fromkeys(tf['words'], 0)\n",
    "        for word in df.loc[df['anime_id']==id]['bag_of_words'].values[0].split(', '):\n",
    "            #if word in tf['words']:\n",
    "                #dict_[word] = tf.loc[(tf.words == word)]['tfidf'][tf.loc[(tf.words == word), 'tfidf'].index[0]]\n",
    "            dict_[word] = (1/len(df.loc[df['anime_id']==id]['bag_of_words'].values[0].split(', '))) * idf[word]\n",
    "        #print(dict_)    \n",
    "        user_prof = user_prof.append(dict_, ignore_index=True)\n",
    "    # Creating user prfile by doing weighted sum of the attributes he/she has rated\n",
    "    user_profile = []\n",
    "    for col in user_prof.columns:\n",
    "        user_profile.append(user_prof[col] @ data['likeness'].values)\n",
    "    return user_profile, data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluation_metrics(cosine_sim_, data):\n",
    "    predicted_likeness = {}\n",
    "    for x in data[\"anime_id\"]:\n",
    "        predicted_likeness[x] = cosine_sim_[0][df.loc[df['anime_id']==x].index[0]]\n",
    "    true_pos = 0\n",
    "    true_neg = 0\n",
    "    false_pos = 0\n",
    "    false_neg = 0\n",
    "    precision = 0\n",
    "    recall = 0\n",
    "    accuracy = 0\n",
    "    for x in predicted_likeness.keys():\n",
    "        predic = predicted_likeness[x]\n",
    "        actual = data.loc[data[\"anime_id\"]==x, 'likeness'].values[0]\n",
    "        if predic >= 0 and actual >= 0:\n",
    "            true_pos += 1\n",
    "        elif predic >= 0 and actual < 0:\n",
    "            false_pos += 1\n",
    "        elif predic < 0 and actual >= 0:\n",
    "            false_neg += 1\n",
    "        else:\n",
    "            true_neg += 1\n",
    "    print(\"true_pos : \"+ str(true_pos) +\"\\ntrue_neg : \"+str(true_neg)+\"\\nfalse_pos : \"+ str(false_pos) +\"\\nfalse_neg : \"+ str(false_neg) +\"\\n\")\n",
    "    if (true_pos + false_pos) != 0:\n",
    "        precision = true_pos/(true_pos + false_pos)\n",
    "    if (true_pos + false_neg) != 0:\n",
    "        recall = true_pos/(true_pos + false_neg)\n",
    "    accuracy = (true_pos + true_neg)/(true_pos + false_neg + true_neg + false_pos)\n",
    "    print(\"Precision : \"+str(precision*100) + \"\\nRecall : \"+str(recall*100)+\"\\nAccuracy : \"+str(accuracy*100))\n",
    "    return accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# recommender function\n",
    "def recommendation_with_user_profile(user):\n",
    "    cosine_sim_user = cosine_similarity(user, tff)\n",
    "    ind = np.argpartition(cosine_sim_user[0], -10)[-10:]\n",
    "    ind = ind[np.argsort(cosine_sim_user[0][ind])]    \n",
    "    return ind, cosine_sim_user"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def grad_descent(user, user_p):\n",
    "    # creating user vector from user dictionary\n",
    "    user_prof = []    \n",
    "    for u in user_p.keys():\n",
    "        user_prof.append(user_p[u])\n",
    "    user_prof = np.asarray(user_prof)\n",
    "    user_prof = user_prof.reshape(-1, 51)\n",
    "    print(user_prof)\n",
    "    cosine_sim_user = cosine_similarity(user_prof, tff)\n",
    "    data = training_data.loc[training_data['user_id']==user]\n",
    "    mean = data.mean()['rating']\n",
    "    data = data.assign(likeness=0)\n",
    "    data.loc[:, \"likeness\"] = data.loc[:, \"rating\"] - mean\n",
    "    predicted_likeness = {}\n",
    "    for x in data[\"anime_id\"]:\n",
    "        predicted_likeness[x] = cosine_sim_user[0][df.loc[df['anime_id']==x].index[0]]\n",
    "        \n",
    "    for x in predicted_likeness.keys():\n",
    "        predic = predicted_likeness[x]\n",
    "        actual = data.loc[data[\"anime_id\"]==x, 'likeness'].values[0]\n",
    "        if predic >= 0 and actual < 0:\n",
    "            for word in df.loc[df['anime_id']==x]['bag_of_words'].values[0].split(', '):\n",
    "                tfidf_score = (1/len(df.loc[df['anime_id']==x]['bag_of_words'].values[0].split(', '))) * idf[word]\n",
    "                user_p[word] -= 0.001*tfidf_score\n",
    "        elif predic < 0 and actual >= 0:\n",
    "            for word in df.loc[df['anime_id']==x]['bag_of_words'].values[0].split(', '):\n",
    "                tfidf_score = (1/len(df.loc[df['anime_id']==x]['bag_of_words'].values[0].split(', '))) * idf[word]\n",
    "                user_p[word] += 0.001*tfidf_score\n",
    "    return user_p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "def grad_descent2(user, user_p, anime, data):\n",
    "    # creating user vector from user dictionary\n",
    "    user_prof = []    \n",
    "    for u in user_p.keys():\n",
    "        user_prof.append(user_p[u])\n",
    "    user_prof = np.asarray(user_prof)\n",
    "    user_prof = user_prof.reshape(-1, 51)\n",
    "    print(user_prof)\n",
    "    cosine_sim_user = cosine_similarity(user_prof, tff)\n",
    "    predic = cosine_sim_user[0][df.loc[df['anime_id']==anime].index[0]]\n",
    "    actual = data.loc[data[\"anime_id\"]==anime, 'likeness'].values[0]\n",
    "    if predic >= 0 and actual < 0:\n",
    "        for word in df.loc[df['anime_id']==anime]['bag_of_words'].values[0].split(', '):\n",
    "            tfidf_score = (1/len(df.loc[df['anime_id']==anime]['bag_of_words'].values[0].split(', '))) * idf[word]\n",
    "            user_p[word] -= 0.01*tfidf_score\n",
    "    elif predic < 0 and actual >= 0:\n",
    "        for word in df.loc[df['anime_id']==anime]['bag_of_words'].values[0].split(', '):\n",
    "            tfidf_score = (1/len(df.loc[df['anime_id']==anime]['bag_of_words'].values[0].split(', '))) * idf[word]\n",
    "            user_p[word] += 0.01*tfidf_score\n",
    "    return user_p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>anime_id</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1149</th>\n",
       "      <td>8</td>\n",
       "      <td>269</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1150</th>\n",
       "      <td>8</td>\n",
       "      <td>355</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1151</th>\n",
       "      <td>8</td>\n",
       "      <td>6702</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1152</th>\n",
       "      <td>8</td>\n",
       "      <td>7088</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1153</th>\n",
       "      <td>8</td>\n",
       "      <td>7593</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1154</th>\n",
       "      <td>8</td>\n",
       "      <td>8630</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1155</th>\n",
       "      <td>8</td>\n",
       "      <td>10620</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1156</th>\n",
       "      <td>8</td>\n",
       "      <td>10793</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1157</th>\n",
       "      <td>8</td>\n",
       "      <td>11241</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1158</th>\n",
       "      <td>8</td>\n",
       "      <td>11757</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1159</th>\n",
       "      <td>8</td>\n",
       "      <td>13759</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1160</th>\n",
       "      <td>8</td>\n",
       "      <td>15583</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      user_id  anime_id  rating\n",
       "1149        8       269       9\n",
       "1150        8       355       9\n",
       "1151        8      6702      10\n",
       "1152        8      7088       5\n",
       "1153        8      7593       9\n",
       "1154        8      8630       8\n",
       "1155        8     10620       6\n",
       "1156        8     10793      10\n",
       "1157        8     11241       7\n",
       "1158        8     11757       9\n",
       "1159        8     13759       9\n",
       "1160        8     15583       9"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_data.loc[training_data['user_id']==8]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n",
      "true_pos : 8\n",
      "true_neg : 0\n",
      "false_pos : 4\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 66.66666666666666\n",
      "Recall : 100.0\n",
      "Accuracy : 66.66666666666666\n",
      "[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n",
      "true_pos : 8\n",
      "true_neg : 0\n",
      "false_pos : 4\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 66.66666666666666\n",
      "Recall : 100.0\n",
      "Accuracy : 66.66666666666666\n",
      "[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n",
      "true_pos : 8\n",
      "true_neg : 0\n",
      "false_pos : 4\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 66.66666666666666\n",
      "Recall : 100.0\n",
      "Accuracy : 66.66666666666666\n",
      "[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n",
      "true_pos : 0\n",
      "true_neg : 4\n",
      "false_pos : 0\n",
      "false_neg : 8\n",
      "\n",
      "Precision : 0\n",
      "Recall : 0.0\n",
      "Accuracy : 33.33333333333333\n",
      "[[-0.00256696  0.          0.          0.          0.          0.\n",
      "  -0.00130837  0.         -0.00162616 -0.00306682 -0.00214972 -0.00185811\n",
      "   0.          0.          0.          0.         -0.00108147  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.         -0.003289\n",
      "   0.          0.          0.          0.         -0.00406441  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 5\n",
      "true_neg : 2\n",
      "false_pos : 2\n",
      "false_neg : 3\n",
      "\n",
      "Precision : 71.42857142857143\n",
      "Recall : 62.5\n",
      "Accuracy : 58.333333333333336\n",
      "[[ 0.00073342  0.          0.          0.          0.00303991  0.\n",
      "   0.00037382  0.         -0.00162616 -0.00306682 -0.00214972 -0.00185811\n",
      "   0.          0.          0.          0.          0.00030899  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.00444631  0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.          0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 1\n",
      "true_neg : 3\n",
      "false_pos : 1\n",
      "false_neg : 7\n",
      "\n",
      "Precision : 50.0\n",
      "Recall : 12.5\n",
      "Accuracy : 33.33333333333333\n",
      "[[-0.0038871   0.          0.          0.         -0.00121597  0.\n",
      "  -0.00198125  0.         -0.00455326 -0.00306682 -0.00214972 -0.00185811\n",
      "   0.          0.          0.          0.         -0.00163765  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.00444631  0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.          0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 1\n",
      "true_neg : 3\n",
      "false_pos : 1\n",
      "false_neg : 7\n",
      "\n",
      "Precision : 50.0\n",
      "Recall : 12.5\n",
      "Accuracy : 33.33333333333333\n",
      "[[-0.0038871   0.          0.          0.         -0.00121597  0.\n",
      "  -0.00198125  0.         -0.00455326 -0.00306682 -0.00214972 -0.00185811\n",
      "   0.          0.          0.          0.         -0.00163765  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.00444631  0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.          0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 3\n",
      "true_neg : 3\n",
      "false_pos : 1\n",
      "false_neg : 5\n",
      "\n",
      "Precision : 75.0\n",
      "Recall : 37.5\n",
      "Accuracy : 50.0\n",
      "[[-0.0038871   0.          0.00361599  0.         -0.00121597  0.\n",
      "   0.00037382  0.         -0.00162616 -0.00306682 -0.00214972 -0.00185811\n",
      "   0.          0.          0.00356313  0.         -0.00163765  0.\n",
      "   0.          0.00654966  0.          0.          0.          0.\n",
      "   0.          0.          0.00444631  0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.          0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 1\n",
      "true_neg : 4\n",
      "false_pos : 0\n",
      "false_neg : 7\n",
      "\n",
      "Precision : 100.0\n",
      "Recall : 12.5\n",
      "Accuracy : 41.66666666666667\n",
      "[[-0.0038871   0.          0.00361599  0.         -0.00121597 -0.00236507\n",
      "  -0.00130837  0.         -0.00371695 -0.00306682 -0.00214972 -0.00185811\n",
      "   0.         -0.00389255  0.00356313 -0.00631379 -0.00163765  0.\n",
      "   0.          0.00187133  0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.          0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 6\n",
      "true_neg : 4\n",
      "false_pos : 0\n",
      "false_neg : 2\n",
      "\n",
      "Precision : 100.0\n",
      "Recall : 75.0\n",
      "Accuracy : 83.33333333333334\n",
      "[[-0.0038871   0.          0.00361599  0.          0.0023306   0.00039418\n",
      "   0.00065419  0.         -0.0012777  -0.00306682 -0.00214972  0.00092905\n",
      "   0.         -0.00389255  0.00356313 -0.00631379 -0.00163765  0.\n",
      "   0.          0.00187133  0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.00703062  0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 6\n",
      "true_neg : 4\n",
      "false_pos : 0\n",
      "false_neg : 2\n",
      "\n",
      "Precision : 100.0\n",
      "Recall : 75.0\n",
      "Accuracy : 83.33333333333334\n",
      "[[-0.0038871   0.          0.00361599  0.          0.0023306   0.00039418\n",
      "   0.00065419  0.         -0.0012777  -0.00306682 -0.00214972  0.00092905\n",
      "   0.         -0.00389255  0.00356313 -0.00631379 -0.00163765  0.\n",
      "   0.          0.00187133  0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.          0.          0.          0.00093971\n",
      "   0.          0.          0.00703062  0.          0.00116126  0.\n",
      "   0.          0.          0.          0.          0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "true_pos : 6\n",
      "true_neg : 4\n",
      "false_pos : 0\n",
      "false_neg : 2\n",
      "\n",
      "Precision : 100.0\n",
      "Recall : 75.0\n",
      "Accuracy : 83.33333333333334\n",
      "--------------------------epoch : 1\n"
     ]
    }
   ],
   "source": [
    "#Gradient descent2\n",
    "#user = 59325\n",
    "user = 8\n",
    "epoch = 0\n",
    "# initializing user profile\n",
    "user_p = dict.fromkeys(tf['words'], 0)\n",
    "accuracy = 0\n",
    "data = training_data.loc[training_data['user_id']==user]\n",
    "mean = data.mean()['rating']\n",
    "data = data.assign(likeness=0)\n",
    "data.loc[:, \"likeness\"] = data.loc[:, \"rating\"] - mean\n",
    "while epoch != 2 and accuracy < 0.8:\n",
    "    for anime in data[\"anime_id\"]:\n",
    "        user_p = grad_descent2(user, user_p, anime, data)\n",
    "        user_prof = [] \n",
    "        for u in user_p.keys():#sorted(user_p.keys()):\n",
    "            user_prof.append(user_p[u])\n",
    "        user_prof = np.asarray(user_prof)\n",
    "        user_prof = user_prof.reshape(-1, 51)\n",
    "        cosine_sim_user = cosine_similarity(user_prof, tff)\n",
    "        accuracy = evaluation_metrics(cosine_sim_user, data)  \n",
    "    epoch += 1\n",
    "    print(\"--------------------------epoch : \" + str(epoch))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      "  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n",
      "epoch : 1\n",
      "true_pos : 0\n",
      "true_neg : 210\n",
      "false_pos : 0\n",
      "false_neg : 464\n",
      "\n",
      "Precision : 0\n",
      "Recall : 0.0\n",
      "Accuracy : 31.15727002967359\n",
      "[[-0.03229163 -0.02568523 -0.02055956 -0.00871725 -0.01734888 -0.03138052\n",
      "  -0.02464696 -0.00906882 -0.01637584 -0.00879981 -0.01025066 -0.01145469\n",
      "  -0.00290422 -0.01683365 -0.00333228 -0.01838304 -0.00360493 -0.00483308\n",
      "  -0.00571956 -0.01027413 -0.00454937 -0.00518185 -0.01126802 -0.01040548\n",
      "  -0.00270743 -0.00819904 -0.01355642 -0.00509486 -0.00133108 -0.0082907\n",
      "  -0.00151568 -0.01323605 -0.00451906 -0.00997141 -0.00433045 -0.01982688\n",
      "  -0.00186099  0.         -0.00210918  0.         -0.02149648  0.\n",
      "  -0.0008567  -0.00073154 -0.00067783 -0.00658804  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 2\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.0346524   0.02039997  0.02881021  0.00836314  0.03662158  0.04074508\n",
      "   0.02081364  0.02606382  0.01018514  0.0032142   0.00925064  0.01782782\n",
      "   0.01307777  0.0134918   0.00128731  0.02345701  0.00908477  0.0147291\n",
      "  -0.00080222  0.01365824  0.00023988  0.00610067  0.03174775  0.00651361\n",
      "   0.00927916  0.03024669  0.0115356   0.0004233   0.01035962  0.00804479\n",
      "   0.00387285  0.00975238  0.00955948  0.00680049  0.00389611  0.00283441\n",
      "   0.00683655  0.00087375  0.00969221  0.0020056   0.01267514  0.00470539\n",
      "   0.          0.00146307 -0.00027113  0.00109479  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 3\n",
      "true_pos : 365\n",
      "true_neg : 94\n",
      "false_pos : 116\n",
      "false_neg : 99\n",
      "\n",
      "Precision : 75.88357588357589\n",
      "Recall : 78.66379310344827\n",
      "Accuracy : 68.10089020771514\n",
      "[[ 0.00236077 -0.00528526  0.00825064 -0.0003541   0.01927269  0.00936456\n",
      "  -0.00383331  0.016995   -0.0061907  -0.00558561 -0.00100002  0.00637313\n",
      "   0.01017355 -0.00334184 -0.00204497  0.00507397  0.00547984  0.00989602\n",
      "  -0.00652178  0.00338411 -0.0043095   0.00091882  0.02047973 -0.00389187\n",
      "   0.00657173  0.02204766 -0.00202083 -0.00467156  0.00902854 -0.0002459\n",
      "   0.00235717 -0.00348367  0.00504042 -0.00317093 -0.00043434 -0.01699247\n",
      "   0.00497556  0.00087375  0.00758302  0.0020056  -0.00882134  0.00470539\n",
      "  -0.0008567   0.00073154 -0.00094897 -0.00549325  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 4\n",
      "true_pos : 380\n",
      "true_neg : 75\n",
      "false_pos : 135\n",
      "false_neg : 84\n",
      "\n",
      "Precision : 73.7864077669903\n",
      "Recall : 81.89655172413794\n",
      "Accuracy : 67.50741839762611\n",
      "[[-0.0028384  -0.00705921 -0.00137987 -0.00158985  0.00593824  0.00610667\n",
      "  -0.00060833  0.01235928 -0.00424765  0.00099369 -0.00441957  0.00265685\n",
      "   0.00726933  0.003713   -0.00427233  0.00325402  0.00187491  0.00506294\n",
      "  -0.00883188  0.00150498 -0.00109185  0.00220309  0.00950049  0.00880331\n",
      "   0.0057368   0.01470906  0.00026768 -0.0046944   0.00829644 -0.00299043\n",
      "   0.00422716 -0.00798526  0.00052136 -0.00408978 -0.00086262 -0.00273115\n",
      "   0.00311457  0.00087375  0.00547384  0.00266164  0.00831593  0.00470539\n",
      "   0.          0.         -0.0012201  -0.00585389  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 5\n",
      "true_pos : 92\n",
      "true_neg : 195\n",
      "false_pos : 15\n",
      "false_neg : 372\n",
      "\n",
      "Precision : 85.98130841121495\n",
      "Recall : 19.82758620689655\n",
      "Accuracy : 42.581602373887236\n",
      "[[-0.00598149 -0.00626171  0.00072796 -0.00277041 -0.00517178 -0.00306138\n",
      "  -0.00954305  0.00674307  0.00246312 -0.00452375 -0.00516594 -0.00066132\n",
      "   0.00436511 -0.00585441 -0.00523694  0.00071447  0.00086501  0.00320044\n",
      "  -0.00620237 -0.00352552 -0.00564122  0.00094746  0.00330404 -0.00083215\n",
      "   0.00302937  0.00651002 -0.001632   -0.00360874  0.00696536  0.0034531\n",
      "   0.00353822 -0.00708905  0.00020262 -0.0013174  -0.00329087 -0.00785622\n",
      "   0.00125358  0.00174749  0.00336465  0.00344889 -0.01162591  0.00470539\n",
      "   0.0008567  -0.00073154 -0.00189793  0.00037996  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 6\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.05454578  0.03608582  0.03661309  0.01189366  0.03743567  0.05094416\n",
      "   0.03098504  0.03359775  0.02127927  0.00706995  0.01126471  0.02001085\n",
      "   0.00972893  0.01876367 -0.00061735  0.0329117   0.01082993  0.01149919\n",
      "  -0.00232496  0.01999749 -0.00085197  0.00969372  0.02799999  0.01608694\n",
      "   0.00807163  0.02247724  0.02052916  0.00190943  0.01113735  0.01025091\n",
      "   0.00644162  0.01340992  0.01168083  0.00944319  0.0049357   0.01480508\n",
      "   0.00447154  0.00174749  0.00607646  0.00545449  0.02254571  0.00470539\n",
      "   0.0017134  -0.00073154 -0.00149123  0.00716119  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 7\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.02225416  0.01040059  0.01605353  0.00317642  0.02008678  0.01956364\n",
      "   0.00633808  0.02452893  0.00490343 -0.00172985  0.00101405  0.00855616\n",
      "   0.00682471  0.00193003 -0.00394963  0.01452866  0.007225    0.00666611\n",
      "  -0.00804452  0.00972336 -0.00540135  0.00451187  0.01673197  0.00568146\n",
      "   0.0053642   0.0142782   0.00697274 -0.00318543  0.00980627  0.00196021\n",
      "   0.00492595  0.00017387  0.00716177 -0.00052822  0.00060524 -0.00502181\n",
      "   0.00261055  0.00174749  0.00396728  0.00545449  0.00104922  0.00470539\n",
      "   0.0008567  -0.00146307 -0.00216907  0.00057315  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 8\n",
      "true_pos : 34\n",
      "true_neg : 208\n",
      "false_pos : 2\n",
      "false_neg : 430\n",
      "\n",
      "Precision : 94.44444444444444\n",
      "Recall : 7.327586206896551\n",
      "Accuracy : 35.90504451038576\n",
      "[[-0.01003747 -0.01528464 -0.00450603 -0.00554083  0.0027379  -0.01181687\n",
      "  -0.01830887  0.01546011 -0.0114724  -0.01052966 -0.0092366  -0.00289853\n",
      "   0.00392049 -0.01490362 -0.00728191 -0.00385438  0.00362007  0.00183303\n",
      "  -0.01376407 -0.00055077 -0.00995072 -0.00066999  0.00546395 -0.00472402\n",
      "   0.00265677  0.00607917 -0.00658368 -0.0082803   0.00847519 -0.00633048\n",
      "   0.00341027 -0.01306219  0.00264271 -0.01049964 -0.00372521 -0.02484869\n",
      "   0.00074956  0.00174749  0.00185809  0.00545449 -0.02044726  0.00470539\n",
      "   0.         -0.00219461 -0.0028469  -0.00601489  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 9\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.05560699  0.03080056  0.04235622  0.01047528  0.04728236  0.05441823\n",
      "   0.02515154  0.04568908  0.01508857  0.00148434  0.00933837  0.02451658\n",
      "   0.01354537  0.01506552 -0.00266232  0.03501589  0.01387693  0.02139521\n",
      "  -0.00884674  0.02256289 -0.00516147  0.00925695  0.04281958  0.01219507\n",
      "   0.01087512  0.03562986  0.01850833 -0.00276213  0.01404481  0.00857855\n",
      "   0.0087988   0.00992625  0.01488056  0.00544197  0.00450136 -0.0021874\n",
      "   0.00763781  0.00262124  0.01260489  0.00746009  0.01372436  0.00618422\n",
      "   0.0008567  -0.00219461 -0.0024402   0.00166794  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 10\n",
      "true_pos : 456\n",
      "true_neg : 18\n",
      "false_pos : 192\n",
      "false_neg : 8\n",
      "\n",
      "Precision : 70.37037037037037\n",
      "Recall : 98.27586206896551\n",
      "Accuracy : 70.32640949554896\n",
      "[[ 0.02331536  0.00511533  0.02179666  0.00175803  0.02993348  0.02303772\n",
      "   0.00050459  0.03662027 -0.00128727 -0.00731546 -0.00091228  0.01306189\n",
      "   0.01064115 -0.00176813 -0.00599459  0.01663285  0.010272    0.01656213\n",
      "  -0.0145663   0.01228876 -0.00971085  0.0040751   0.03155156  0.00178959\n",
      "   0.00816769  0.02743082  0.00495191 -0.00785699  0.01271373  0.00028786\n",
      "   0.00728312 -0.0033098   0.0103615  -0.00452944  0.00017091 -0.02201428\n",
      "   0.00577682  0.00262124  0.01049571  0.00746009 -0.00777212  0.00618422\n",
      "   0.         -0.00292614 -0.00311803 -0.0049201   0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 11\n",
      "true_pos : 184\n",
      "true_neg : 174\n",
      "false_pos : 36\n",
      "false_neg : 280\n",
      "\n",
      "Precision : 83.63636363636363\n",
      "Recall : 39.6551724137931\n",
      "Accuracy : 53.11572700296736\n",
      "[[-7.69950022e-03 -1.91837463e-02  1.41789411e-03 -3.95163406e-03\n",
      "   1.25845917e-02 -5.94847439e-03 -1.96541545e-02  2.75514493e-02\n",
      "  -1.55929992e-02 -1.09595478e-02 -1.00040575e-02  1.60720272e-03\n",
      "   7.73692731e-03 -1.23069222e-02 -8.22196014e-03 -1.44207255e-04\n",
      "   6.66706794e-03  1.17290564e-02 -2.02858504e-02  2.01463429e-03\n",
      "  -1.21178774e-02  9.57866820e-05  2.02835370e-02 -4.42355309e-03\n",
      "   5.99224072e-03  1.92317872e-02 -7.51516411e-03 -1.25681457e-02\n",
      "   1.13826493e-02 -7.41547569e-03  6.45638775e-03 -1.28946404e-02\n",
      "   5.84243810e-03 -1.28734832e-02 -4.15954417e-03 -3.51809446e-02\n",
      "   3.91582957e-03  2.62123967e-03  8.38652049e-03  7.46008868e-03\n",
      "  -2.48790435e-02  6.18422316e-03 -8.56700083e-04 -3.65767885e-03\n",
      "  -3.79586381e-03 -1.15081384e-02  0.00000000e+00  0.00000000e+00\n",
      "   0.00000000e+00  0.00000000e+00  0.00000000e+00]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch : 12\n",
      "true_pos : 464\n",
      "true_neg : 1\n",
      "false_pos : 209\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.94502228826151\n",
      "Recall : 100.0\n",
      "Accuracy : 68.99109792284867\n",
      "[[ 0.03505104  0.01507237  0.0243859   0.00350096  0.02024245  0.03084493\n",
      "   0.00986887  0.03812814  0.00937432 -0.00074684  0.00131345  0.01925663\n",
      "   0.01139795  0.00939492 -0.00781156  0.02368131  0.01014441  0.01234203\n",
      "  -0.01640844  0.01764686 -0.00732863  0.00782379  0.02885167  0.01249554\n",
      "   0.00668678  0.01824897  0.00911404 -0.0078174   0.0126738   0.00368967\n",
      "   0.00935979 -0.0047861   0.00892277 -0.00367859 -0.00069292 -0.01251965\n",
      "   0.00205484  0.00349499  0.01160052  0.00890337  0.00929258  0.00658754\n",
      "   0.         -0.00365768 -0.00338916 -0.00382531  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 13\n",
      "true_pos : 238\n",
      "true_neg : 162\n",
      "false_pos : 48\n",
      "false_neg : 226\n",
      "\n",
      "Precision : 83.21678321678321\n",
      "Recall : 51.293103448275865\n",
      "Accuracy : 59.34718100890207\n",
      "[[ 0.00275941 -0.01061286  0.00382634 -0.00521629  0.00289356 -0.00053558\n",
      "  -0.01477809  0.02905932 -0.00700152 -0.00954665 -0.00852332  0.00780194\n",
      "   0.00849373 -0.00743873 -0.01114384  0.0055416   0.00653948  0.00750895\n",
      "  -0.02212799  0.00737274 -0.011878    0.00296991  0.01758365  0.00209006\n",
      "   0.00397935  0.01004994 -0.00444238 -0.01291227  0.01134272 -0.00460102\n",
      "   0.00784412 -0.01802215  0.00440371 -0.01365    -0.00502337 -0.03160651\n",
      "   0.00019385  0.00349499  0.00949133  0.00890337 -0.01220391  0.00658754\n",
      "  -0.0008567  -0.00438921 -0.004067   -0.01041335  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 14\n",
      "true_pos : 464\n",
      "true_neg : 2\n",
      "false_pos : 208\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 69.04761904761905\n",
      "Recall : 100.0\n",
      "Accuracy : 69.13946587537092\n",
      "[[ 2.80887170e-02  9.31413349e-03  2.10991566e-02  2.11213691e-03\n",
      "   1.65125822e-02  2.43263806e-02  8.47490540e-03  3.14048916e-02\n",
      "   1.29997421e-02 -6.42888331e-06 -2.52196254e-03  1.54129419e-02\n",
      "   9.39242704e-03  1.26943414e-02 -9.96176058e-03  2.08405291e-02\n",
      "   8.56576585e-03  9.77224904e-03 -1.82505802e-02  1.57288677e-02\n",
      "  -7.08875149e-03  7.82119082e-03  1.32693294e-02  1.59861077e-02\n",
      "   3.80940581e-03  1.62451048e-02  1.21645909e-02 -9.70095179e-03\n",
      "   1.26338660e-02  3.17573020e-03  1.02308156e-02 -2.16543973e-04\n",
      "   1.04642023e-02 -4.57688704e-03 -2.09007499e-03 -8.94521653e-03\n",
      "  -1.16311769e-04  3.49498623e-03  9.39089420e-03  1.09089739e-02\n",
      "   1.62369036e-02  6.58754206e-03  0.00000000e+00 -5.12075040e-03\n",
      "  -3.66029725e-03 -3.40671808e-03  0.00000000e+00  0.00000000e+00\n",
      "   0.00000000e+00  0.00000000e+00  0.00000000e+00]]\n",
      "epoch : 15\n",
      "true_pos : 113\n",
      "true_neg : 193\n",
      "false_pos : 17\n",
      "false_neg : 351\n",
      "\n",
      "Precision : 86.92307692307692\n",
      "Recall : 24.353448275862068\n",
      "Accuracy : 45.40059347181009\n",
      "[[-0.00420291 -0.0163711   0.00053959 -0.00660511 -0.0008363  -0.00675975\n",
      "  -0.01580616  0.02233608 -0.00337609 -0.00880624 -0.01194485  0.00395825\n",
      "   0.00648821 -0.00413931 -0.01218913  0.00270082  0.00496083  0.00493917\n",
      "  -0.02397014  0.00545474 -0.01163813  0.0029673   0.00200131  0.00558062\n",
      "   0.00110197  0.00804607 -0.00139183 -0.01479582  0.01130278 -0.00511497\n",
      "   0.00871514 -0.0134526   0.00594514 -0.0145483  -0.00642053 -0.02803208\n",
      "  -0.0019773   0.00349499  0.00728171  0.01090897 -0.00525958  0.00658754\n",
      "  -0.0008567  -0.00585229 -0.00433813 -0.00999476  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 16\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 5.09124740e-02  2.63202176e-02  4.01196397e-02  5.34702404e-03\n",
      "   2.94828097e-02  4.71621486e-02  1.70475919e-02  3.80068342e-02\n",
      "   2.03809262e-02  9.86166440e-04  2.53725187e-03  2.15194194e-02\n",
      "   1.40940463e-02  1.88979974e-02 -7.56953498e-03  3.52186279e-02\n",
      "   1.13316367e-02  9.18284922e-03 -1.90528035e-02  1.99628721e-02\n",
      "  -6.84887542e-03  8.78061938e-03  3.64471094e-02  1.93340743e-02\n",
      "   8.83267048e-03  1.99239633e-02  1.88722330e-02 -9.27764860e-03\n",
      "   1.25939336e-02  9.30459813e-03  9.89618580e-03  4.35300804e-03\n",
      "   1.13044196e-02 -2.05596170e-05  2.50501932e-04 -5.37078199e-03\n",
      "   5.63465905e-03  4.36873279e-03  1.17511724e-02  1.21273301e-02\n",
      "   2.63514667e-02  6.99086096e-03  0.00000000e+00 -5.85228617e-03\n",
      "  -3.93143038e-03 -2.31193155e-03  0.00000000e+00  0.00000000e+00\n",
      "   0.00000000e+00  0.00000000e+00  0.00000000e+00]]\n",
      "epoch : 17\n",
      "true_pos : 435\n",
      "true_neg : 40\n",
      "false_pos : 170\n",
      "false_neg : 29\n",
      "\n",
      "Precision : 71.900826446281\n",
      "Recall : 93.75\n",
      "Accuracy : 70.47477744807121\n",
      "[[ 0.01862085  0.00063499  0.01956008 -0.00337022  0.01213392  0.01578163\n",
      "  -0.00759936  0.02893802  0.00400509 -0.00781364 -0.00771341  0.01006473\n",
      "   0.01118983  0.00206435 -0.01090181  0.01683559  0.00772671  0.00434977\n",
      "  -0.02477236  0.00968875 -0.01139825  0.00359877  0.02517909  0.00892859\n",
      "   0.00612524  0.01172493  0.00531581 -0.01437251  0.01126285  0.0010139\n",
      "   0.00838051 -0.00888304  0.00678536 -0.00999197 -0.00407995 -0.02519767\n",
      "   0.00377367  0.00436873  0.00964199  0.01212733  0.00485499  0.00699086\n",
      "  -0.0008567  -0.00658382 -0.00460926 -0.00889997  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 18\n",
      "true_pos : 89\n",
      "true_neg : 201\n",
      "false_pos : 9\n",
      "false_neg : 375\n",
      "\n",
      "Precision : 90.81632653061224\n",
      "Recall : 19.181034482758623\n",
      "Accuracy : 43.02670623145401\n",
      "[[-0.00963445 -0.01964038  0.00129782 -0.00764508 -0.00228884 -0.00796685\n",
      "  -0.02035615  0.02119358 -0.009052   -0.00918408 -0.0135611   0.00271513\n",
      "   0.00828561 -0.00726694 -0.0124978   0.00507229  0.00496516 -0.00048331\n",
      "  -0.02708246  0.00279861 -0.01322627  0.0010094   0.01391107  0.00357103\n",
      "   0.0034178   0.00398479 -0.00322616 -0.01375583  0.00993177 -0.00685725\n",
      "   0.00755378 -0.01634354  0.00285064 -0.01200682 -0.00625487 -0.02902475\n",
      "   0.00191268  0.00436873  0.0075328   0.01212733 -0.0128166   0.00699086\n",
      "   0.         -0.00731536 -0.0048804  -0.01268664  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 19\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.04740199  0.02349098  0.04056649  0.00624246  0.03769276  0.04713212\n",
      "   0.02028519  0.04271619  0.01750897  0.00234957  0.00361455  0.02447628\n",
      "   0.01407513  0.01946145 -0.00787821  0.03511547  0.01256211  0.01907887\n",
      "  -0.02216513  0.02361988 -0.00843702  0.00903414  0.03277361  0.02006234\n",
      "   0.00917255  0.03291597  0.016471   -0.00823766  0.0139041   0.00717074\n",
      "   0.0129423   0.0066449   0.01520537  0.00218881  0.00197169 -0.00636346\n",
      "   0.00802551  0.00524248  0.01416167  0.01413293  0.02135502  0.00739418\n",
      "   0.0008567  -0.00731536 -0.0044737  -0.00500382  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 20\n",
      "true_pos : 440\n",
      "true_neg : 41\n",
      "false_pos : 169\n",
      "false_neg : 24\n",
      "\n",
      "Precision : 72.2495894909688\n",
      "Recall : 94.82758620689656\n",
      "Accuracy : 71.36498516320475\n",
      "[[ 0.01511036 -0.00219425  0.02000692 -0.00247478  0.02034387  0.0157516\n",
      "  -0.00436176  0.03364737  0.00113313 -0.00645023 -0.00663611  0.01302159\n",
      "   0.01117091  0.00262781 -0.01121049  0.01673243  0.00895718  0.01424579\n",
      "  -0.02788469  0.01334576 -0.01298639  0.00385229  0.02150559  0.00965685\n",
      "   0.00646512  0.02471693  0.00291458 -0.01333253  0.01257302 -0.00111995\n",
      "   0.01142663 -0.00659116  0.0106863  -0.00778261 -0.00235876 -0.02619034\n",
      "   0.00616452  0.00524248  0.01205248  0.01413293 -0.00014146  0.00739418\n",
      "   0.         -0.00804689 -0.00515153 -0.01159186  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 21\n",
      "true_pos : 151\n",
      "true_neg : 189\n",
      "false_pos : 21\n",
      "false_neg : 313\n",
      "\n",
      "Precision : 87.79069767441861\n",
      "Recall : 32.543103448275865\n",
      "Accuracy : 50.445103857566764\n",
      "[[-0.0084736  -0.01949653  0.00189533 -0.0087363   0.00402531 -0.00637497\n",
      "  -0.01977907  0.0251774  -0.01149023 -0.01065472 -0.01300805  0.00490353\n",
      "   0.00826669 -0.00814146 -0.01280647  0.00429128  0.00619564  0.00941272\n",
      "  -0.02816695  0.00348098 -0.01481442  0.00119733  0.01116168  0.00237423\n",
      "   0.00428967  0.0169768  -0.00518276 -0.01415477  0.01124193 -0.00882329\n",
      "   0.00991095 -0.01604322  0.00616724 -0.01600803 -0.00453368 -0.02694809\n",
      "   0.00430354  0.00524248  0.0099433   0.01413293 -0.0107584   0.00739418\n",
      "   0.0008567  -0.00877843 -0.00582936 -0.01537853  0.          0.\n",
      "   0.          0.          0.        ]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch : 22\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.04176019  0.02062874  0.0326248   0.00271917  0.02440326  0.03677854\n",
      "   0.01319897  0.04104241  0.01349352 -0.00044202  0.00120272  0.02294316\n",
      "   0.01363051  0.01472052 -0.00818688  0.03220706  0.01217718  0.01101588\n",
      "  -0.02324962  0.02056349 -0.01002517  0.00800704  0.02834928  0.017518\n",
      "   0.00951243  0.02028089  0.01284703 -0.0086366   0.01253308  0.00412787\n",
      "   0.01229765 -0.00398473  0.00942288 -0.00375767 -0.00106706 -0.0042868\n",
      "   0.00580266  0.00611623  0.01265511  0.01478897  0.02341322  0.0077975\n",
      "   0.0017134  -0.00877843 -0.00542266 -0.00769571  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 23\n",
      "true_pos : 360\n",
      "true_neg : 98\n",
      "false_pos : 112\n",
      "false_neg : 104\n",
      "\n",
      "Precision : 76.27118644067797\n",
      "Recall : 77.58620689655173\n",
      "Accuracy : 67.95252225519289\n",
      "[[ 0.00946856 -0.00505649  0.01206524 -0.00599807  0.00705438  0.00539802\n",
      "  -0.01144799  0.03197359 -0.00288232 -0.00924182 -0.00904794  0.01148847\n",
      "   0.01072629 -0.00211313 -0.01151916  0.01382402  0.00857225  0.0061828\n",
      "  -0.02896917  0.01028937 -0.01457454  0.00282519  0.01708126  0.00711252\n",
      "   0.006805    0.01208185 -0.00070939 -0.01373147  0.011202   -0.00416283\n",
      "   0.01078197 -0.01722078  0.00490382 -0.01372908 -0.00539751 -0.02411368\n",
      "   0.00394168  0.00611623  0.01054592  0.01478897  0.00191674  0.0077975\n",
      "   0.0008567  -0.00950997 -0.0061005  -0.01428374  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 24\n",
      "true_pos : 342\n",
      "true_neg : 124\n",
      "false_pos : 86\n",
      "false_neg : 122\n",
      "\n",
      "Precision : 79.90654205607477\n",
      "Recall : 73.70689655172413\n",
      "Accuracy : 69.13946587537092\n",
      "[[-8.42808133e-04 -9.01334958e-03  5.29236690e-03 -6.35086392e-03\n",
      "   8.98746484e-04  3.71060150e-03 -6.62419265e-03  2.75228329e-02\n",
      "  -5.16500674e-03 -3.14287484e-03 -9.32983400e-03  9.30175886e-03\n",
      "   8.38966720e-03  1.79853000e-03 -1.15998388e-02  1.12589805e-02\n",
      "   7.19904936e-03  7.62093969e-03 -2.62356733e-02  6.32318485e-03\n",
      "  -9.78528940e-03  3.13441094e-03  1.00762194e-02  1.21869244e-02\n",
      "   5.29453575e-03  7.95938750e-03  2.59806046e-03 -1.11596728e-02\n",
      "   9.87092093e-03 -5.95755331e-03  1.13724940e-02 -1.03846270e-02\n",
      "   5.19725546e-03 -1.22851619e-02 -3.95914262e-03 -9.39670607e-03\n",
      "   2.08068832e-03  6.11622591e-03  8.43673917e-03  1.47889698e-02\n",
      "   2.37530185e-05  7.79749876e-03  1.71340017e-03 -1.02415008e-02\n",
      "  -5.69379572e-03 -7.27711045e-03  0.00000000e+00  0.00000000e+00\n",
      "   0.00000000e+00  0.00000000e+00  0.00000000e+00]]\n",
      "epoch : 25\n",
      "true_pos : 443\n",
      "true_neg : 20\n",
      "false_pos : 190\n",
      "false_neg : 21\n",
      "\n",
      "Precision : 69.98420221169036\n",
      "Recall : 95.47413793103449\n",
      "Accuracy : 68.69436201780415\n",
      "[[ 0.0124271   0.00429715  0.01088202 -0.00470123  0.00389632  0.00724733\n",
      "  -0.00469787  0.02334386  0.00326866 -0.00206355 -0.00643006  0.01079012\n",
      "   0.00548545 -0.00093673 -0.01130169  0.01447901  0.00651353  0.00575844\n",
      "  -0.02235826  0.00025176 -0.00673307  0.00273148  0.00661633  0.00751994\n",
      "   0.00445961  0.00431495  0.00312657 -0.00794835  0.00853984  0.00203946\n",
      "   0.00985682 -0.00716492  0.00588026 -0.00594015 -0.00189856  0.0004784\n",
      "   0.00177052  0.00611623  0.00703062  0.01478897  0.00930208  0.0077975\n",
      "   0.0025701  -0.0102415  -0.0052871  -0.00027048  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 26\n",
      "true_pos : 5\n",
      "true_neg : 210\n",
      "false_pos : 0\n",
      "false_neg : 459\n",
      "\n",
      "Precision : 100.0\n",
      "Recall : 1.0775862068965518\n",
      "Accuracy : 31.899109792284868\n",
      "[[-1.87245575e-02 -2.00019219e-02 -7.05164092e-03 -8.04977093e-03\n",
      "  -1.30993081e-02 -1.90669693e-02 -1.98108576e-02  1.52608459e-02\n",
      "  -1.26471571e-02 -6.68434045e-03 -1.28256550e-02  1.07939158e-03\n",
      "   2.58122815e-03 -1.03132687e-02 -1.13823638e-02 -2.14046122e-03\n",
      "   4.68619404e-03  9.25357719e-04 -2.20165729e-02 -8.00289342e-03\n",
      "  -6.90677654e-03 -2.04431685e-03 -3.72758695e-03  1.82014492e-03\n",
      "   1.75217890e-03 -2.70624274e-03 -7.36931163e-03 -9.58066059e-03\n",
      "   7.20875770e-03 -6.25123552e-03  8.34114119e-03 -1.59531264e-02\n",
      "   1.86206213e-03 -1.43316345e-02 -6.22901335e-03 -1.83617847e-02\n",
      "  -9.04647095e-05  6.11622591e-03  4.92143118e-03  1.47889698e-02\n",
      "  -1.21944038e-02  7.79749876e-03  3.42680033e-03 -1.09730366e-02\n",
      "  -5.55822915e-03 -6.85851560e-03  0.00000000e+00  0.00000000e+00\n",
      "   0.00000000e+00  0.00000000e+00  0.00000000e+00]]\n",
      "epoch : 27\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n",
      "[[ 0.04821946  0.02608327  0.04231813  0.00903062  0.04087115  0.05276424\n",
      "   0.02516189  0.04952284  0.01391381  0.00532967  0.00626176  0.03002744\n",
      "   0.01856322  0.01965587 -0.00676277  0.03969959  0.0173759   0.02048754\n",
      "  -0.01709924  0.01592947 -0.00211753  0.00897584  0.03882613  0.01873924\n",
      "   0.01019219  0.03573949  0.0177227  -0.00406249  0.01889946  0.01008425\n",
      "   0.01372967  0.00703531  0.0159406   0.00244027  0.00199755  0.00429951\n",
      "   0.00860707  0.00698997  0.01566823  0.01679457  0.02197722  0.00820082\n",
      "   0.0042835  -0.00877843 -0.00515153  0.00082431  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 28\n",
      "true_pos : 460\n",
      "true_neg : 11\n",
      "false_pos : 199\n",
      "false_neg : 4\n",
      "\n",
      "Precision : 69.80273141122913\n",
      "Recall : 99.13793103448276\n",
      "Accuracy : 69.8813056379822\n",
      "[[ 0.01592784  0.00039804  0.02175857  0.00031337  0.02352227  0.02138373\n",
      "   0.00051493  0.04045402 -0.00246202 -0.00347014 -0.0039889   0.01857275\n",
      "   0.015659    0.00282222 -0.01009505  0.02131655  0.01377097  0.01565446\n",
      "  -0.0228188   0.00565534 -0.0066669   0.00379399  0.02755811  0.00833376\n",
      "   0.00748476  0.02754045  0.00416628 -0.00915736  0.01756838  0.00179356\n",
      "   0.01221399 -0.00620074  0.01142154 -0.00753115 -0.0023329  -0.01552737\n",
      "   0.00674608  0.00698997  0.01355905  0.01679457  0.00048073  0.00820082\n",
      "   0.0034268  -0.00950997 -0.00582936 -0.00576373  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 29\n",
      "true_pos : 152\n",
      "true_neg : 188\n",
      "false_pos : 22\n",
      "false_neg : 312\n",
      "\n",
      "Precision : 87.35632183908046\n",
      "Recall : 32.758620689655174\n",
      "Accuracy : 50.445103857566764\n",
      "[[-0.01636379 -0.02424757  0.001199   -0.00661036  0.00617338 -0.00878\n",
      "  -0.0203024   0.03138521 -0.0158477  -0.00991625 -0.01252883  0.00711806\n",
      "   0.01275478 -0.01178447 -0.01232242  0.00385817  0.01016604  0.01082138\n",
      "  -0.0279144  -0.00461879 -0.01121627  0.00025196  0.01629009 -0.00155838\n",
      "   0.00477733  0.01934141 -0.00798955 -0.01425222  0.01623729 -0.00649714\n",
      "   0.01069831 -0.01644944  0.00690248 -0.01710402 -0.00666335 -0.03165414\n",
      "   0.00488509  0.00698997  0.01144986  0.01679457 -0.02101575  0.00820082\n",
      "   0.0025701  -0.0102415  -0.0065072  -0.01099938  0.          0.\n",
      "   0.          0.          0.        ]]\n",
      "epoch : 30\n",
      "true_pos : 464\n",
      "true_neg : 0\n",
      "false_pos : 210\n",
      "false_neg : 0\n",
      "\n",
      "Precision : 68.84272997032642\n",
      "Recall : 100.0\n",
      "Accuracy : 68.84272997032642\n"
     ]
    }
   ],
   "source": [
    "#Gradient descent\n",
    "user = 59325\n",
    "epoch = 0\n",
    "# initializing user profile\n",
    "user_p = dict.fromkeys(tf['words'], 0)\n",
    "accuracy = 0\n",
    "while epoch != 30 and accuracy < 0.8:    \n",
    "    user_p = grad_descent(user, user_p)\n",
    "    #print(user_p)\n",
    "    epoch += 1\n",
    "    print(\"epoch : \" + str(epoch))\n",
    "    user_prof = [] \n",
    "    for u in user_p.keys():#sorted(user_p.keys()):\n",
    "        user_prof.append(user_p[u])\n",
    "    user_prof = np.asarray(user_prof)\n",
    "    user_prof = user_prof.reshape(-1, 51)\n",
    "    cosine_sim_user = cosine_similarity(user_prof, tff)\n",
    "    data = training_data.loc[training_data['user_id']==user]\n",
    "    mean = data.mean()['rating']\n",
    "    data = data.assign(likeness=0)\n",
    "    data.loc[:, \"likeness\"] = data.loc[:, \"rating\"] - mean\n",
    "    accuracy = evaluation_metrics(cosine_sim_user, data)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.12130752 -0.11739642  0.17498252 ...  0.01265109  0.01265109\n",
      "   0.00248652]]\n"
     ]
    }
   ],
   "source": [
    "#for user in test_user:\n",
    "user = 8\n",
    "user_profile, data = create_user_profile(user)\n",
    "# Reshaping the array\n",
    "user_prof = np.asarray(user_profile)\n",
    "user_prof = user_prof.reshape(-1, 51)\n",
    "index, cosine_sim_user = recommendation_with_user_profile(user_prof)\n",
    "print(cosine_sim_user)\n",
    "#print(\"Recommending movies for user \"+ str(user))\n",
    "#df[['anime_id', 'name', 'genre', 'type']].loc[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "true_pos : 247\n",
      "true_neg : 145\n",
      "false_pos : 65\n",
      "false_neg : 217\n",
      "\n",
      "Precision : 79.16666666666666\n",
      "Recall : 53.23275862068966\n",
      "Accuracy : 58.160237388724035\n"
     ]
    }
   ],
   "source": [
    "evaluation_metrics(cosine_sim_user, data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "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>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3473</th>\n",
       "      <td>30039</td>\n",
       "      <td>Future Card Buddyfight 100</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10603</th>\n",
       "      <td>20087</td>\n",
       "      <td>Treasure Gaust</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10766</th>\n",
       "      <td>9822</td>\n",
       "      <td>Xiang Qi Master</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5650</th>\n",
       "      <td>32785</td>\n",
       "      <td>Future Card Buddyfight DDD</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3950</th>\n",
       "      <td>21835</td>\n",
       "      <td>Majin Bone</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2745</th>\n",
       "      <td>19067</td>\n",
       "      <td>Future Card Buddyfight</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7851</th>\n",
       "      <td>33690</td>\n",
       "      <td>Nazotokine</td>\n",
       "      <td>game</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2971</th>\n",
       "      <td>5298</td>\n",
       "      <td>Yu☆Gi☆Oh! 5D&amp;#039;s: Shinkasuru Kettou! Stardu...</td>\n",
       "      <td>fantasy, game, sci-fi</td>\n",
       "      <td>Special</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4629</th>\n",
       "      <td>20533</td>\n",
       "      <td>Z/X: Ignition</td>\n",
       "      <td>adventure, fantasy, game, sci-fi</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3157</th>\n",
       "      <td>873</td>\n",
       "      <td>.hack//Roots</td>\n",
       "      <td>adventure, drama, fantasy, game, sci-fi</td>\n",
       "      <td>TV</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       anime_id                                               name  \\\n",
       "3473      30039                         Future Card Buddyfight 100   \n",
       "10603     20087                                     Treasure Gaust   \n",
       "10766      9822                                    Xiang Qi Master   \n",
       "5650      32785                         Future Card Buddyfight DDD   \n",
       "3950      21835                                         Majin Bone   \n",
       "2745      19067                             Future Card Buddyfight   \n",
       "7851      33690                                         Nazotokine   \n",
       "2971       5298  Yu☆Gi☆Oh! 5D&#039;s: Shinkasuru Kettou! Stardu...   \n",
       "4629      20533                                      Z/X: Ignition   \n",
       "3157        873                                       .hack//Roots   \n",
       "\n",
       "                                         genre     type  \n",
       "3473                                      game       TV  \n",
       "10603                                     game       TV  \n",
       "10766                                     game       TV  \n",
       "5650                                      game       TV  \n",
       "3950                                      game       TV  \n",
       "2745                                      game       TV  \n",
       "7851                                      game       TV  \n",
       "2971                     fantasy, game, sci-fi  Special  \n",
       "4629          adventure, fantasy, game, sci-fi       TV  \n",
       "3157   adventure, drama, fantasy, game, sci-fi       TV  "
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind,_ = recommendation_with_user_profile(user_prof)\n",
    "df[['anime_id', 'name', 'genre', 'type']].loc[ind]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "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>anime_id</th>\n",
       "      <th>name</th>\n",
       "      <th>genre</th>\n",
       "      <th>type</th>\n",
       "      <th>bag_of_words</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>804</th>\n",
       "      <td>11757</td>\n",
       "      <td>Sword Art Online</td>\n",
       "      <td>action, adventure, fantasy, game, romance</td>\n",
       "      <td>TV</td>\n",
       "      <td>action, adventure, fantasy, game, romance, TV</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     anime_id              name                                      genre  \\\n",
       "804     11757  Sword Art Online  action, adventure, fantasy, game, romance   \n",
       "\n",
       "    type                                   bag_of_words  \n",
       "804   TV  action, adventure, fantasy, game, romance, TV  "
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[df.anime_id == training_data.loc[training_data['user_id']==8]['anime_id'].values[9]]"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}