Movie Recommendation System

Overview

This is a Movie Recommendation System that suggests similar movies based on content similarity. It uses Natural Language Processing (NLP) and cosine similarity to find movies similar to the one selected by the user.

Features

• Recommends top 5 similar movies based on content analysis.
• Uses cosine similarity to measure movie similarity.
• Fetches movie posters using the OMDb API.
• Provides an interactive user interface using Streamlit.

Technologies Used

• Python
• Pandas
• NumPy
• Scikit-learn
• NLTK (Natural Language Toolkit)
• Streamlit (for UI)
• OMDb API (for fetching posters)

---

Setup and Installation

1. Clone the Repository

git clone https://github.com/purva-8/Movie-Recommendation-System.git
cd Movie-Recommendation-System

2. Install all dependencies

pip install -r requirements.txt

3. Run the Model Preprocessing

Execute main.py to generate the necessary files (movies.pkl and similarity.pkl).

python main.py

4. Run the Streamlit App

Once the preprocessing is done, start the recommendation system:

streamlit run app.py

---

How It Works

1. Data Preprocessing (main.py)

• Merges the movies and credits datasets.
• Extracts important features (genres, keywords, cast, crew, overview).
• Cleans and processes text (removes spaces, lowercase, and stems words).
• Converts text data into numerical vectors using CountVectorizer.
• Computes cosine similarity between movie vectors.
• Saves processed data into movies.pkl and similarity.pkl.

2. Recommendation System (app.py)

• Loads the movies and similarity matrix.
• Provides a dropdown to select a movie.
• Computes top 5 similar movies.
• Fetches and displays movie posters using the OMDb API.

---

Example Output

When a user selects "Iron Man", the system may recommend:

1. The Avengers
2. Iron Man 2
3. Captain America: Civil War
4. Avengers: Age of Ultron
5. Thor

Each recommended movie is displayed along with its poster.

---