Movie-Recommendation System.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[2 1]\n",
      " [1 2]]\n",
      "[[1.  0.8]\n",
      " [0.8 1. ]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer #import for counting number of words into text.\n",
    "\n",
    "from sklearn.metrics.pairwise import cosine_similarity #method for finding similarity or cos theta using dot product of two vectors\n",
    "\n",
    "text = [\"London Paris London\" , \"Paris Paris London\"] #array of texts\n",
    "\n",
    "cv = CountVectorizer()   # creation of object.\n",
    "count_matrix = cv.fit_transform(text) # will give sparse matrix representation.\n",
    "print(count_matrix.toarray()) # coverts into 2d-array.first array has [2,1] --> means 2 times london and 1 time paris.similarly second\n",
    "\n",
    "\n",
    "#finding cosine similarity between two texts.\n",
    "\n",
    "similarity_scores = cosine_similarity(count_matrix) # get [1 , 0.8] --> means first sentence is similar to itself 100% and 80% similar to second sentence.similar logic to second array\n",
    "print(similarity_scores)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avatar\n",
      "Guardians of the Galaxy\n",
      "Aliens\n",
      "Star Wars: Clone Wars: Volume 1\n",
      "Star Trek Into Darkness\n",
      "Star Trek Beyond\n",
      "Alien\n",
      "Lockout\n",
      "Jason X\n",
      "The Helix... Loaded\n",
      "Moonraker\n",
      "Planet of the Apes\n",
      "Galaxy Quest\n",
      "Gravity\n",
      "Alien³\n",
      "Jupiter Ascending\n",
      "The Wolverine\n",
      "Silent Running\n",
      "Zathura: A Space Adventure\n",
      "Trekkies\n",
      "Cargo\n",
      "Wing Commander\n",
      "Star Trek\n",
      "Lost in Space\n",
      "Babylon A.D.\n",
      "The Fifth Element\n",
      "Oblivion\n",
      "Titan A.E.\n",
      "AVP: Alien vs. Predator\n",
      "The Empire Strikes Back\n",
      "Dragonball Evolution\n",
      "Superman Returns\n",
      "Divergent\n",
      "John Carter\n",
      "The Black Hole\n",
      "The Ice Pirates\n",
      "Memoirs of an Invisible Man\n",
      "Starship Troopers\n",
      "The Astronaut's Wife\n",
      "Machete Kills\n",
      "Soldier\n",
      "The Abyss\n",
      "Damnation Alley\n",
      "Men in Black\n",
      "Space Cowboys\n",
      "Space Dogs\n",
      "The Time Machine\n",
      "Sheena\n",
      "Captain America: Civil War\n",
      "Star Trek: Insurrection\n",
      "Oz: The Great and Powerful\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "###### helper functions. Use them when needed #######\n",
    "def get_title_from_index(index):\n",
    "\treturn df[df.index == index][\"title\"].values[0]\n",
    "\n",
    "def get_index_from_title(title):\n",
    "\treturn df[df.title == title][\"index\"].values[0]\n",
    "##################################################\n",
    "\n",
    "##Step 1: Read CSV File\n",
    "df = pd.read_csv(\"movie_dataset.csv\")   # df stands for data frame.read csv file\n",
    "#print(df.head())\n",
    "#print(df.columns)  # df.colmun --> prints all columns\n",
    "\n",
    "\n",
    "##Step 2: Select Features\n",
    "features = ['keywords','cast','genres','director']\n",
    "##Step 3: Create a column in DF which combines all selected features\n",
    "for feature in features:\n",
    "    df[feature] = df[feature].fillna('')\n",
    "def combine_features(row):\n",
    "    return row['keywords'] +\" \"+ row['cast'] +\" \"+ row['genres'] +\" \"+ row['director']\n",
    "\n",
    "df[\"combined_features\"] = df.apply(combine_features, axis = 1) # combined column of all features\n",
    "#print(df[\"combined_features\"].head())\n",
    "\n",
    "##Step 4: Create count matrix from this new combined column\n",
    "\n",
    "cv = CountVectorizer()\n",
    "\n",
    "count_matrix = cv.fit_transform(df[\"combined_features\"])\n",
    "#print(count_matrix)\n",
    "##Step 5: Compute the Cosine Similarity based on the count_matrix\n",
    "cosine_sim = cosine_similarity(count_matrix)\n",
    "\n",
    "movie_user_likes = \"Avatar\"\n",
    "\n",
    "## Step 6: Get index of this movie from its title\n",
    "movie_index = get_index_from_title(movie_user_likes)\n",
    "similar_movies = list(enumerate(cosine_sim[movie_index]))\n",
    "\n",
    "\n",
    "## Step 7: Get a list of similar movies in descending order of similarity score\n",
    "sorted_similar_movies = sorted(similar_movies , key = lambda x : x[1], reverse = True)\n",
    "\n",
    "## Step 8: Print titles of first 50 movies\n",
    "i = 0\n",
    "for movie in sorted_similar_movies:\n",
    "    print(get_title_from_index(movie[0]))\n",
    "    i = i + 1\n",
    "    if(i > 50):\n",
    "        break\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}