embeddings.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Movie Rating Prediction Model\n",
    "\n",
    "## Goal\n",
    "- The purpose of this notebook is to build a predictive model for movie ratings based on various features such as keywords, genres, director, and average user ratings. \n",
    "- The model predicts how a particular user will rate a movie on a scale of 0 - 5.0, given their past rating history.\n",
    "- Utilizing embeddings to encode high cardinality categorical features like keywords(1932 unique keywords) to improve model performance, instead of traditional encoding methods like label-encoding or one-hot encoding which has its flaws in our context.\n",
    "\n",
    "## Scope\n",
    "- **Personalized Movie Recommendations**: The model can be extended to create a recommendation system that suggests movies to users based on their predicted ratings. \n",
    "- By identifying movies that a user is likely to rate highly, the system can provide tailored suggestions, enhancing user experience on platforms like streaming services.\n",
    "\n",
    "## Data\n",
    "- Contains movie rating data logged by me on *Letterboxd* \n",
    "- Gathered more features from tmdb api using the unique movie id.\n",
    "\n",
    "## Sections\n",
    "1. Data preprocessing\n",
    "2. Model Preparation:\n",
    "    - Define a neural network using embedding layers for categorical data.\n",
    "3. Model Training\n",
    "4. Model Evaluation\n",
    "5. Predicting rating by a user from a different movie sample.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "from keras import layers\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set the random seeds for reproducibility\n",
    "seed_value = 42 \n",
    "np.random.seed(seed_value)\n",
    "tf.random.set_seed(seed_value)\n",
    "random.seed(seed_value)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.5.0\n"
     ]
    }
   ],
   "source": [
    "print(keras.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "df=pd.read_pickle(\"dataframe_keywords.pkl\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>Date_logged</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>Name</th>\n",
       "      <th>Year</th>\n",
       "      <th>overview</th>\n",
       "      <th>genres</th>\n",
       "      <th>keywords</th>\n",
       "      <th>actors</th>\n",
       "      <th>director</th>\n",
       "      <th>average_rating</th>\n",
       "      <th>vote_count</th>\n",
       "      <th>popularity</th>\n",
       "      <th>rating_given</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021-10-02</td>\n",
       "      <td>278</td>\n",
       "      <td>The Shawshank Redemption</td>\n",
       "      <td>1994</td>\n",
       "      <td>Imprisoned in the 1940s for the double murder ...</td>\n",
       "      <td>Drama, Crime</td>\n",
       "      <td>prison, friendship, police brutality, corrupti...</td>\n",
       "      <td>Tim Robbins, Morgan Freeman, Bob Gunton, Willi...</td>\n",
       "      <td>Frank Darabont</td>\n",
       "      <td>8.706</td>\n",
       "      <td>26839</td>\n",
       "      <td>150.307</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021-10-02</td>\n",
       "      <td>489</td>\n",
       "      <td>Good Will Hunting</td>\n",
       "      <td>1997</td>\n",
       "      <td>When professors discover that an aimless janit...</td>\n",
       "      <td>Drama</td>\n",
       "      <td>boston, massachusetts, professor, baseball, ma...</td>\n",
       "      <td>Matt Damon, Robin Williams, Ben Affleck, Stell...</td>\n",
       "      <td>Gus Van Sant</td>\n",
       "      <td>8.158</td>\n",
       "      <td>12188</td>\n",
       "      <td>72.632</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Date_logged  movie_id                      Name  Year  \\\n",
       "0  2021-10-02       278  The Shawshank Redemption  1994   \n",
       "1  2021-10-02       489         Good Will Hunting  1997   \n",
       "\n",
       "                                            overview        genres  \\\n",
       "0  Imprisoned in the 1940s for the double murder ...  Drama, Crime   \n",
       "1  When professors discover that an aimless janit...         Drama   \n",
       "\n",
       "                                            keywords  \\\n",
       "0  prison, friendship, police brutality, corrupti...   \n",
       "1  boston, massachusetts, professor, baseball, ma...   \n",
       "\n",
       "                                              actors        director  \\\n",
       "0  Tim Robbins, Morgan Freeman, Bob Gunton, Willi...  Frank Darabont   \n",
       "1  Matt Damon, Robin Williams, Ben Affleck, Stell...    Gus Van Sant   \n",
       "\n",
       "   average_rating  vote_count  popularity  rating_given  \n",
       "0           8.706       26839     150.307           5.0  \n",
       "1           8.158       12188      72.632           5.0  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For our task, we will focus on the following features for training the model, as they are the most influential in determining movie ratings by a user:\n",
    "\n",
    "- Keywords \n",
    "- Genres \n",
    "- Average rating (audience)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "df=df[[\"movie_id\",\"Name\",\"keywords\",\"genres\",\"average_rating\",\"rating_given\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>movie_id</th>\n",
       "      <th>Name</th>\n",
       "      <th>keywords</th>\n",
       "      <th>genres</th>\n",
       "      <th>average_rating</th>\n",
       "      <th>rating_given</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>278</td>\n",
       "      <td>The Shawshank Redemption</td>\n",
       "      <td>prison, friendship, police brutality, corrupti...</td>\n",
       "      <td>Drama, Crime</td>\n",
       "      <td>8.706</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>489</td>\n",
       "      <td>Good Will Hunting</td>\n",
       "      <td>boston, massachusetts, professor, baseball, ma...</td>\n",
       "      <td>Drama</td>\n",
       "      <td>8.158</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   movie_id                      Name  \\\n",
       "0       278  The Shawshank Redemption   \n",
       "1       489         Good Will Hunting   \n",
       "\n",
       "                                            keywords        genres  \\\n",
       "0  prison, friendship, police brutality, corrupti...  Drama, Crime   \n",
       "1  boston, massachusetts, professor, baseball, ma...         Drama   \n",
       "\n",
       "   average_rating  rating_given  \n",
       "0           8.706           5.0  \n",
       "1           8.158           5.0  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert keywords from a string to a list (if they are stored as comma-separated strings)\n",
    "df['keywords'] = df['keywords'].apply(lambda x: x.split(', '))\n",
    "df['genres'] = df['genres'].apply(lambda x: x.split(', '))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>movie_id</th>\n",
       "      <th>Name</th>\n",
       "      <th>keywords</th>\n",
       "      <th>genres</th>\n",
       "      <th>average_rating</th>\n",
       "      <th>rating_given</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>278</td>\n",
       "      <td>The Shawshank Redemption</td>\n",
       "      <td>[prison, friendship, police brutality, corrupt...</td>\n",
       "      <td>[Drama, Crime]</td>\n",
       "      <td>8.706</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>489</td>\n",
       "      <td>Good Will Hunting</td>\n",
       "      <td>[boston, massachusetts, professor, baseball, m...</td>\n",
       "      <td>[Drama]</td>\n",
       "      <td>8.158</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   movie_id                      Name  \\\n",
       "0       278  The Shawshank Redemption   \n",
       "1       489         Good Will Hunting   \n",
       "\n",
       "                                            keywords          genres  \\\n",
       "0  [prison, friendship, police brutality, corrupt...  [Drama, Crime]   \n",
       "1  [boston, massachusetts, professor, baseball, m...         [Drama]   \n",
       "\n",
       "   average_rating  rating_given  \n",
       "0           8.706           5.0  \n",
       "1           8.158           5.0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "# Flatten the list of all keywords and genres across the dataset\n",
    "all_keywords = [keyword for keywords_list in df['keywords'] for keyword in keywords_list]\n",
    "all_genres = [genre for genre_list in df['genres'] for genre in genre_list]\n",
    "\n",
    "# Create separate LabelEncoders for keywords and genres\n",
    "le_keywords = LabelEncoder()\n",
    "le_genres = LabelEncoder()\n",
    "\n",
    "# Fit the LabelEncoders on the unique keywords and genres\n",
    "le_keywords.fit(all_keywords)\n",
    "le_genres.fit(all_genres)\n",
    "\n",
    "\n",
    "# Now encode each keyword list and genre list into integer IDs\n",
    "df['keywords_encoded'] = df['keywords'].apply(lambda keywords: le_keywords.transform(keywords))\n",
    "df['genres_encoded'] = df['genres'].apply(lambda genres: le_genres.transform(genres))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>movie_id</th>\n",
       "      <th>Name</th>\n",
       "      <th>keywords</th>\n",
       "      <th>genres</th>\n",
       "      <th>average_rating</th>\n",
       "      <th>rating_given</th>\n",
       "      <th>keywords_encoded</th>\n",
       "      <th>genres_encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>207</th>\n",
       "      <td>840430</td>\n",
       "      <td>The Holdovers</td>\n",
       "      <td>[vietnam war, winter, boarding school, holiday...</td>\n",
       "      <td>[Comedy, Drama]</td>\n",
       "      <td>7.714</td>\n",
       "      <td>4.5</td>\n",
       "      <td>[1838, 1892, 228, 833, 236, 1057, 11, 38, 52, ...</td>\n",
       "      <td>[3, 6]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>208</th>\n",
       "      <td>1581</td>\n",
       "      <td>The Holiday</td>\n",
       "      <td>[london, england, movie business, christmas pa...</td>\n",
       "      <td>[Comedy, Romance]</td>\n",
       "      <td>7.093</td>\n",
       "      <td>2.5</td>\n",
       "      <td>[987, 577, 1138, 344, 409, 1821, 1673, 991, 27...</td>\n",
       "      <td>[3, 13]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>209</th>\n",
       "      <td>6977</td>\n",
       "      <td>No Country for Old Men</td>\n",
       "      <td>[vietnam veteran, sheriff, trailer park, based...</td>\n",
       "      <td>[Crime, Drama, Thriller]</td>\n",
       "      <td>7.945</td>\n",
       "      <td>4.0</td>\n",
       "      <td>[1837, 1533, 1770, 172, 830, 1352, 540, 1730, ...</td>\n",
       "      <td>[4, 6, 15]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210</th>\n",
       "      <td>792307</td>\n",
       "      <td>Poor Things</td>\n",
       "      <td>[ship, london, england, experiment, paris, fra...</td>\n",
       "      <td>[Science Fiction, Romance, Comedy]</td>\n",
       "      <td>7.763</td>\n",
       "      <td>4.0</td>\n",
       "      <td>[1539, 987, 577, 605, 1250, 704, 172, 1205, 98...</td>\n",
       "      <td>[14, 13, 3]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>211</th>\n",
       "      <td>693134</td>\n",
       "      <td>Dune: Part Two</td>\n",
       "      <td>[epic, based on novel or book, fight, politics...</td>\n",
       "      <td>[Science Fiction, Adventure]</td>\n",
       "      <td>8.165</td>\n",
       "      <td>5.0</td>\n",
       "      <td>[581, 172, 669, 1303, 1455, 1453, 1605, 342, 4...</td>\n",
       "      <td>[14, 1]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     movie_id                    Name  \\\n",
       "207    840430           The Holdovers   \n",
       "208      1581             The Holiday   \n",
       "209      6977  No Country for Old Men   \n",
       "210    792307             Poor Things   \n",
       "211    693134          Dune: Part Two   \n",
       "\n",
       "                                              keywords  \\\n",
       "207  [vietnam war, winter, boarding school, holiday...   \n",
       "208  [london, england, movie business, christmas pa...   \n",
       "209  [vietnam veteran, sheriff, trailer park, based...   \n",
       "210  [ship, london, england, experiment, paris, fra...   \n",
       "211  [epic, based on novel or book, fight, politics...   \n",
       "\n",
       "                                 genres  average_rating  rating_given  \\\n",
       "207                     [Comedy, Drama]           7.714           4.5   \n",
       "208                   [Comedy, Romance]           7.093           2.5   \n",
       "209            [Crime, Drama, Thriller]           7.945           4.0   \n",
       "210  [Science Fiction, Romance, Comedy]           7.763           4.0   \n",
       "211        [Science Fiction, Adventure]           8.165           5.0   \n",
       "\n",
       "                                      keywords_encoded genres_encoded  \n",
       "207  [1838, 1892, 228, 833, 236, 1057, 11, 38, 52, ...         [3, 6]  \n",
       "208  [987, 577, 1138, 344, 409, 1821, 1673, 991, 27...        [3, 13]  \n",
       "209  [1837, 1533, 1770, 172, 830, 1352, 540, 1730, ...     [4, 6, 15]  \n",
       "210  [1539, 987, 577, 605, 1250, 704, 172, 1205, 98...    [14, 13, 3]  \n",
       "211  [581, 172, 669, 1303, 1455, 1453, 1605, 342, 4...        [14, 1]  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.tail(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras_preprocessing.sequence import pad_sequences\n",
    "\n",
    "MAX_KEYWORDS = 20\n",
    "MAX_GENRES = 3\n",
    "\n",
    "# Pad the sequences (post-padding will add zeros after shorter sequences)\n",
    "X_keywords_padded = np.array(pad_sequences(df['keywords_encoded'], maxlen=MAX_KEYWORDS, padding='post'))\n",
    "X_genres_padded = np.array(pad_sequences(df['genres_encoded'], maxlen=MAX_GENRES, padding='post'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 212 entries, 0 to 211\n",
      "Data columns (total 8 columns):\n",
      " #   Column            Non-Null Count  Dtype  \n",
      "---  ------            --------------  -----  \n",
      " 0   movie_id          212 non-null    int64  \n",
      " 1   Name              212 non-null    object \n",
      " 2   keywords          212 non-null    object \n",
      " 3   genres            212 non-null    object \n",
      " 4   average_rating    212 non-null    float64\n",
      " 5   rating_given      212 non-null    float64\n",
      " 6   keywords_encoded  212 non-null    object \n",
      " 7   genres_encoded    212 non-null    object \n",
      "dtypes: float64(2), int64(1), object(5)\n",
      "memory usage: 14.9+ KB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1932"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.explode(\"keywords\")[\"keywords\"].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python310\\lib\\site-packages\\keras\\src\\layers\\core\\embedding.py:90: UserWarning: Argument `input_length` is deprecated. Just remove it.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_absolute_error\n",
    "\n",
    "y_rating = np.array(df['rating_given'], dtype=np.float32)\n",
    "\n",
    "# Step 1: Define the model\n",
    "# Input for keywords\n",
    "keyword_input = keras.Input(shape=(20,), name='keywords_input')  # Shape matches padded keywords\n",
    "keywords_embedded = keras.layers.Embedding(input_dim=1932, output_dim=16, input_length=20)(keyword_input)\n",
    "\n",
    "# Input for genres\n",
    "genre_input = keras.Input(shape=(3,), name='genres_input')  # Shape matches padded genres\n",
    "genres_embedded = keras.layers.Embedding(input_dim=18, output_dim=16, input_length=3)(genre_input)\n",
    "\n",
    "\n",
    "# Flatten both embeddings\n",
    "keywords_flattened = keras.layers.Flatten()(keywords_embedded)\n",
    "genres_flattened = keras.layers.Flatten()(genres_embedded)\n",
    "\n",
    "# Input for average_rating (numerical feature)\n",
    "avg_rating_input = keras.Input(shape=(1,), name='avg_rating_input')\n",
    "\n",
    "# Concatenate embeddings from keywords and genres\n",
    "combined = keras.layers.Concatenate()([keywords_flattened, genres_flattened, avg_rating_input])\n",
    "\n",
    "# Add a Dense layer after concatenation\n",
    "out = keras.layers.Dense(32, activation='relu')(combined)\n",
    "out = keras.layers.Dense(16, activation='relu')(out)\n",
    "\n",
    "# Add hidden layers (using loop for flexibility)\n",
    "hidden_units = (32, 16)\n",
    "for n_hidden in hidden_units:\n",
    "    out = keras.layers.Dense(n_hidden, activation='relu')(out)\n",
    "\n",
    "# Output layer: predicted rating\n",
    "output = keras.layers.Dense(1, activation='linear', name='prediction')(out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = keras.Model(inputs=[keyword_input,genre_input,avg_rating_input], outputs=output)\n",
    "model.compile(optimizer='adam', loss='mse', metrics=['mean_absolute_error'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"functional\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"functional\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)        </span>┃<span style=\"font-weight: bold\"> Output Shape      </span>┃<span style=\"font-weight: bold\">    Param # </span>┃<span style=\"font-weight: bold\"> Connected to      </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩\n",
       "│ keywords_input      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">20</span>)        │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ genres_input        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ embedding           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">20</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>)    │     <span style=\"color: #00af00; text-decoration-color: #00af00\">30,912</span> │ keywords_input[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>… │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Embedding</span>)         │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ embedding_1         │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>)     │        <span style=\"color: #00af00; text-decoration-color: #00af00\">288</span> │ genres_input[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">…</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Embedding</span>)         │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ flatten (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>)   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">320</span>)       │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ embedding[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ flatten_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)        │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ embedding_1[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>] │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ avg_rating_input    │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ concatenate         │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">369</span>)       │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ flatten[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],    │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Concatenate</span>)       │                   │            │ flatten_1[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],  │\n",
       "│                     │                   │            │ avg_rating_input… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)       │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)        │     <span style=\"color: #00af00; text-decoration-color: #00af00\">11,840</span> │ concatenate[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>] │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>)        │        <span style=\"color: #00af00; text-decoration-color: #00af00\">528</span> │ dense[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]       │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)        │        <span style=\"color: #00af00; text-decoration-color: #00af00\">544</span> │ dense_1[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>)        │        <span style=\"color: #00af00; text-decoration-color: #00af00\">528</span> │ dense_2[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ prediction (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)         │         <span style=\"color: #00af00; text-decoration-color: #00af00\">17</span> │ dense_3[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]     │\n",
       "└─────────────────────┴───────────────────┴────────────┴───────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)       \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape     \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m   Param #\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConnected to     \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩\n",
       "│ keywords_input      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m20\u001b[0m)        │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ genres_input        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m3\u001b[0m)         │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ embedding           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m20\u001b[0m, \u001b[38;5;34m16\u001b[0m)    │     \u001b[38;5;34m30,912\u001b[0m │ keywords_input[\u001b[38;5;34m0\u001b[0m… │\n",
       "│ (\u001b[38;5;33mEmbedding\u001b[0m)         │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ embedding_1         │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m16\u001b[0m)     │        \u001b[38;5;34m288\u001b[0m │ genres_input[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m…\u001b[0m │\n",
       "│ (\u001b[38;5;33mEmbedding\u001b[0m)         │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ flatten (\u001b[38;5;33mFlatten\u001b[0m)   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m320\u001b[0m)       │          \u001b[38;5;34m0\u001b[0m │ embedding[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ flatten_1 (\u001b[38;5;33mFlatten\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m48\u001b[0m)        │          \u001b[38;5;34m0\u001b[0m │ embedding_1[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m] │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ avg_rating_input    │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)         │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ concatenate         │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m369\u001b[0m)       │          \u001b[38;5;34m0\u001b[0m │ flatten[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],    │\n",
       "│ (\u001b[38;5;33mConcatenate\u001b[0m)       │                   │            │ flatten_1[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],  │\n",
       "│                     │                   │            │ avg_rating_input… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense (\u001b[38;5;33mDense\u001b[0m)       │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m)        │     \u001b[38;5;34m11,840\u001b[0m │ concatenate[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m] │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m)        │        \u001b[38;5;34m528\u001b[0m │ dense[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]       │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_2 (\u001b[38;5;33mDense\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m)        │        \u001b[38;5;34m544\u001b[0m │ dense_1[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m)        │        \u001b[38;5;34m528\u001b[0m │ dense_2[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ prediction (\u001b[38;5;33mDense\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)         │         \u001b[38;5;34m17\u001b[0m │ dense_3[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]     │\n",
       "└─────────────────────┴───────────────────┴────────────┴───────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">44,657</span> (174.44 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m44,657\u001b[0m (174.44 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">44,657</span> (174.44 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m44,657\u001b[0m (174.44 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the model summary\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 2: Split the dataset into training and validation sets\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# Assuming df contains the entire dataset (movies and ratings)\n",
    "df_train, df_val = train_test_split(df, test_size=0.1, random_state=seed_value)  # 90% train, 10% validation\n",
    "\n",
    "# Prepare validation data (use the same padding and conversion as for the training set)\n",
    "X_val_keywords_padded = pad_sequences(df_val['keywords_encoded'], maxlen=MAX_KEYWORDS, padding='post', truncating='post')\n",
    "X_val_genres_padded = pad_sequences(df_val['genres_encoded'], maxlen=MAX_GENRES, padding='post', truncating='post')\n",
    "\n",
    "X_val_avg_rating = np.array(df_val['average_rating'], dtype=np.int32)\n",
    "\n",
    "X_avg_rating= np.array(df['average_rating'], dtype=np.int32)\n",
    "y_val_rating = np.array(df_val['rating_given'], dtype=np.float32)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 24ms/step - loss: 9.8417 - mean_absolute_error: 3.0258 - val_loss: 1.3113 - val_mean_absolute_error: 0.9620\n",
      "Epoch 2/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 8ms/step - loss: 0.8749 - mean_absolute_error: 0.7552 - val_loss: 0.3007 - val_mean_absolute_error: 0.4589\n",
      "Epoch 3/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step - loss: 0.3570 - mean_absolute_error: 0.4809 - val_loss: 0.1891 - val_mean_absolute_error: 0.3449\n",
      "Epoch 4/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 10ms/step - loss: 0.2494 - mean_absolute_error: 0.3880 - val_loss: 0.1204 - val_mean_absolute_error: 0.2639\n",
      "Epoch 5/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - loss: 0.1753 - mean_absolute_error: 0.3169 - val_loss: 0.0820 - val_mean_absolute_error: 0.2051\n",
      "Epoch 6/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.1071 - mean_absolute_error: 0.2407 - val_loss: 0.0618 - val_mean_absolute_error: 0.1656\n",
      "Epoch 7/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.0552 - mean_absolute_error: 0.1607 - val_loss: 0.0548 - val_mean_absolute_error: 0.1361\n",
      "Epoch 8/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 10ms/step - loss: 0.0267 - mean_absolute_error: 0.0944 - val_loss: 0.0532 - val_mean_absolute_error: 0.1187\n",
      "Epoch 9/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.0145 - mean_absolute_error: 0.0649 - val_loss: 0.0534 - val_mean_absolute_error: 0.1206\n",
      "Epoch 10/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 10ms/step - loss: 0.0094 - mean_absolute_error: 0.0523 - val_loss: 0.0498 - val_mean_absolute_error: 0.1214\n",
      "Epoch 11/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - loss: 0.0072 - mean_absolute_error: 0.0467 - val_loss: 0.0468 - val_mean_absolute_error: 0.1265\n",
      "Epoch 12/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.0089 - mean_absolute_error: 0.0669 - val_loss: 0.0701 - val_mean_absolute_error: 0.2225\n",
      "Epoch 13/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.0174 - mean_absolute_error: 0.1149 - val_loss: 0.1155 - val_mean_absolute_error: 0.2922\n",
      "Epoch 14/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.0303 - mean_absolute_error: 0.1547 - val_loss: 0.1013 - val_mean_absolute_error: 0.2659\n",
      "Epoch 15/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - loss: 0.0532 - mean_absolute_error: 0.1842 - val_loss: 0.2003 - val_mean_absolute_error: 0.3304\n",
      "Epoch 16/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.1513 - mean_absolute_error: 0.3473 - val_loss: 0.7788 - val_mean_absolute_error: 0.8205\n",
      "Epoch 17/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step - loss: 0.3943 - mean_absolute_error: 0.5823 - val_loss: 0.9989 - val_mean_absolute_error: 0.9090\n",
      "Epoch 18/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step - loss: 0.6473 - mean_absolute_error: 0.7480 - val_loss: 0.2365 - val_mean_absolute_error: 0.3856\n",
      "Epoch 19/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 10ms/step - loss: 0.5742 - mean_absolute_error: 0.6410 - val_loss: 0.6428 - val_mean_absolute_error: 0.7291\n",
      "Epoch 20/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 15ms/step - loss: 0.7015 - mean_absolute_error: 0.7583 - val_loss: 0.0899 - val_mean_absolute_error: 0.2435\n",
      "Epoch 21/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step - loss: 0.2095 - mean_absolute_error: 0.3877 - val_loss: 0.0836 - val_mean_absolute_error: 0.2042\n",
      "Epoch 22/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 8ms/step - loss: 0.0821 - mean_absolute_error: 0.2371 - val_loss: 0.0649 - val_mean_absolute_error: 0.1592\n",
      "Epoch 23/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 16ms/step - loss: 0.0230 - mean_absolute_error: 0.1163 - val_loss: 0.0480 - val_mean_absolute_error: 0.1195\n",
      "Epoch 24/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - loss: 0.0048 - mean_absolute_error: 0.0437 - val_loss: 0.0423 - val_mean_absolute_error: 0.1193\n",
      "Epoch 25/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - loss: 0.0040 - mean_absolute_error: 0.0415 - val_loss: 0.0412 - val_mean_absolute_error: 0.1220\n",
      "Epoch 26/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - loss: 0.0037 - mean_absolute_error: 0.0422 - val_loss: 0.0414 - val_mean_absolute_error: 0.1142\n",
      "Epoch 27/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 15ms/step - loss: 0.0025 - mean_absolute_error: 0.0329 - val_loss: 0.0430 - val_mean_absolute_error: 0.1041\n",
      "Epoch 28/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - loss: 0.0018 - mean_absolute_error: 0.0246 - val_loss: 0.0447 - val_mean_absolute_error: 0.1035\n",
      "Epoch 29/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step - loss: 0.0015 - mean_absolute_error: 0.0255 - val_loss: 0.0442 - val_mean_absolute_error: 0.0996\n",
      "Epoch 30/30\n",
      "\u001b[1m27/27\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 16ms/step - loss: 0.0012 - mean_absolute_error: 0.0207 - val_loss: 0.0419 - val_mean_absolute_error: 0.0940\n"
     ]
    }
   ],
   "source": [
    "# Step 3: Train the model\n",
    "history = model.fit(\n",
    "    [X_keywords_padded, X_genres_padded,X_avg_rating],  \n",
    "    y_rating,                              # Target: rating_given\n",
    "    batch_size=8,                          # Batch size (small dataset, so small batch size)\n",
    "    epochs=30,                             # Number of epochs\n",
    "    verbose=1,                             # Show training progress\n",
    "    validation_data=([X_val_keywords_padded, X_val_genres_padded,X_val_avg_rating], y_val_rating)  # Validation data\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Step 4: Visualize training and validation MAE\n",
    "fig, ax = plt.subplots(figsize=(15, 6))\n",
    "ax.plot(history.epoch, history.history['val_mean_absolute_error'], label='Validation MAE')\n",
    "ax.plot(history.epoch, history.history['mean_absolute_error'], label='Training MAE')\n",
    "ax.set_xlabel('Epoch')\n",
    "ax.set_ylabel('Mean Absolute Error')\n",
    "ax.set_xlim(left=0, right=history.epoch[-1])\n",
    "baseline_mae = 0.73  # Baseline MAE (replace with the actual baseline you calculated)\n",
    "ax.axhline(baseline_mae, ls='--', label='Baseline', color='#002255', alpha=0.5)\n",
    "ax.grid()\n",
    "fig.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 334ms/step\n",
      "R² score for validation set: 0.9106921390879748\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "\n",
    "# Predict ratings for the validation set\n",
    "y_val_pred = (model.predict([X_val_keywords_padded,X_val_genres_padded,X_val_avg_rating]))\n",
    "\n",
    "# Calculate R² score for validation\n",
    "r2_val = r2_score(y_val_rating, y_val_pred)\n",
    "print(f'R² score for validation set: {r2_val}')\n",
    "\n",
    "# Create the scatter plot for validation data\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.scatter(y_val_rating, y_val_pred, color='blue', alpha=0.6, label='Predicted Ratings')\n",
    "plt.plot([y_val_rating.min(), y_val_rating.max()], \n",
    "         [y_val_rating.min(), y_val_rating.max()], \n",
    "         color='red', linestyle='--', label='Ideal Line')  # Ideal line\n",
    "plt.title('Actual Ratings vs. Predicted Ratings (Validation Set)')\n",
    "plt.xlabel('Actual Ratings')\n",
    "plt.ylabel('Predicted Ratings')\n",
    "plt.xlim(y_val_rating.min(), y_val_rating.max())\n",
    "plt.ylim(y_val_rating.min(), y_val_rating.max())\n",
    "plt.grid()\n",
    "plt.legend()  \n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Model Evaluation**:\n",
    "- *r2 score* = 0.91\n",
    "- *mae* = 0.09\n",
    "\n",
    " Our model looks pretty decent. Most of the predictions are very near to the actual rating."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Checking movies with a rating difference over 0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "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>Name</th>\n",
       "      <th>Actual Rating</th>\n",
       "      <th>Predicted Rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Dune</td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.662421</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Name  Actual Rating  Predicted Rating\n",
       "13  Dune            4.5          3.662421"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Flatten y_val_pred to make it 1-dimensional\n",
    "y_val_pred_flat = y_val_pred.ravel()  # or y_val_pred[:, 0]\n",
    "\n",
    "# Create a DataFrame with actual and predicted ratings side by side\n",
    "df_results = pd.DataFrame({\n",
    "    'Actual Rating': y_val_rating,\n",
    "    'Predicted Rating': y_val_pred_flat  # Use the flattened array\n",
    "})\n",
    "\n",
    "# Display rows with significant error in prediction\n",
    "high_diff=df_results[abs(df_results[\"Actual Rating\"]-df_results[\"Predicted Rating\"])>0.5]\n",
    "\n",
    "\n",
    "(df_val.reset_index(drop=True)\n",
    " .merge(high_diff,left_index=True,right_index=True)\n",
    " [[\"Name\",\"Actual Rating\",\"Predicted Rating\"]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our model's prediction says i shouldnt have liked Dune as much as i did. But Denis Villenueve outclassed our model, lol!\n",
    "\n",
    "Justifying this prediction:\n",
    "- The essence of Dune is rooted in cinematography, editing and it just has a different vibe? \n",
    "this is something which is not translated in our feature variables.\n",
    "- so it may be fair to say this might be the reason our model predicted a lower rating than the actual rating given by me.\n",
    "\n",
    "- bit of stretch maybe, but hey, we gotta infer what we gotta infer, right?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Step 6: Save the model\n",
    "# model.save('final_model.keras')\n",
    "\n",
    "# Step 7: Load the model\n",
    "# # from keras.models import load_model\n",
    "# # loaded_model = load_model('movie_rating_model.h5')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model.save_weights('final_model.weights.h5')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "let's fetch movie details from tmdb api for predictions!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "genres_ids=pd.read_pickle(\"genres.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "import numpy as np\n",
    "from keras_preprocessing.sequence import pad_sequences\n",
    "\n",
    "# Dictionary for genre ID to name mapping\n",
    "id_to_name = {genre['id']: genre['name'] for genre in genres_ids['genres']}\n",
    "\n",
    "# Function to fetch details (keywords, genres, avg_rating) using the movie name\n",
    "def fetch_details(movie_name):\n",
    "    movie_url = f\"https://api.themoviedb.org/3/search/movie?query={movie_name}&include_adult=false&language=en-US&page=1\"\n",
    "    headers = {\n",
    "    \"accept\": \"application/json\",\n",
    "    \"Authorization\": \"YOUR_TMDB_API_HERE\"\n",
    "    }\n",
    "\n",
    "    response = requests.get(movie_url, headers=headers).json()[\"results\"][0]\n",
    "    movie_id = response[\"id\"]\n",
    "    title=response[\"title\"]\n",
    "    date=response[\"release_date\"][:4]\n",
    "    avg_rating = response[\"vote_average\"]\n",
    "    genre_ids = response[\"genre_ids\"]\n",
    "    genres = [id_to_name[genre_id] for genre_id in genre_ids]\n",
    "\n",
    "    # Get keywords\n",
    "    keyword_url = f\"https://api.themoviedb.org/3/movie/{movie_id}/keywords\"\n",
    "    keyword_response = requests.get(keyword_url, headers=headers).json()[\"keywords\"]\n",
    "    keywords = [key[\"name\"] for key in keyword_response]\n",
    "\n",
    "    return keywords, genres, avg_rating, movie_id,title,date\n",
    "\n",
    "# Function to encode the inputs and prepare for the model\n",
    "def prepare_input(keywords, genres, avg_rating, max_keywords_length, max_genres_length, keyword_encoder, genre_encoder):\n",
    "    \n",
    "    # Encode keywords with handling for unseen labels\n",
    "    encoded_keywords = []\n",
    "    for kw in keywords:\n",
    "        try:\n",
    "            encoded_keywords.append(keyword_encoder.transform([kw])[0])  # Transform each keyword\n",
    "        except ValueError:\n",
    "            encoded_keywords.append(-1)  # Assign -1 for unseen labels\n",
    "    \n",
    "    # Pad the keywords to the max length\n",
    "    padded_keywords = pad_sequences([encoded_keywords], maxlen=max_keywords_length, padding='post')\n",
    "    \n",
    "    # Encode genres with handling for unseen labels\n",
    "    encoded_genres = []\n",
    "    for genre in genres:\n",
    "        try:\n",
    "            encoded_genres.append(genre_encoder.transform([genre])[0])  # Transform each genre\n",
    "        except ValueError:\n",
    "            encoded_genres.append(-1)  # Assign -1 for unseen labels\n",
    "    \n",
    "    # Pad the genres to the max length\n",
    "    padded_genres = pad_sequences([encoded_genres], maxlen=max_genres_length, padding='post')\n",
    "\n",
    "    avg_rating_input = np.array([avg_rating])  \n",
    "    \n",
    "    return padded_keywords, padded_genres, avg_rating_input"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I mostly logged english movies, so to check if our model is well generalized, let's predict a movie from a different sample than we've trained on (a *bollywood* movie)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### predicting the rating user(me, in this case) would give by taking *movie title* as input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "Predicted rating for 'Andhadhun',2018 (Movie ID: 534780): 4.50\n"
     ]
    }
   ],
   "source": [
    "# Step 1: Define the movie name\n",
    "movie_name = \"andhadhun\"  # Replace with your desired movie name\n",
    "\n",
    "# Step 2: Fetch details using the movie name\n",
    "keywords, genres, avg_rating, movie_id,title,date = fetch_details(movie_name)\n",
    "\n",
    "# Step 3: Prepare input for the model\n",
    "max_keywords_length = 20  # Example max length for keywords\n",
    "max_genres_length = 3     # Example max length for genres\n",
    "\n",
    "# Assume you have trained encoders for keywords and genres\n",
    "padded_keywords, padded_genres, avg_rating_input = prepare_input(\n",
    "    keywords, genres, avg_rating, max_keywords_length, max_genres_length, \n",
    "    keyword_encoder=le_keywords, genre_encoder=le_genres)\n",
    "\n",
    "# Step 4: Make the prediction using the model\n",
    "predicted_rating = model.predict([padded_keywords, padded_genres, avg_rating_input])\n",
    "\n",
    "# Print the predicted rating\n",
    "print(f\"Predicted rating for '{title}',{date} (Movie ID: {movie_id}): {np.ceil(predicted_rating[0][0] * 2) / 2:.2f}\") #taking the ceil since we can only rate in .5 multiples\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "yeppp 4.5 is exactly the rating i'd give to Andhadhun!(I have a good take on movies, don't i lol?)\n",
    "\n",
    "our prediction model works, phew!"
   ]
  }
 ],
 "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": "3.10.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}