Content-Based Filtering Model

This project implements a content-based filtering model for recommending movies. The model uses various features extracted from a dataset of the top 1000 movies from IMDb to compute similarities and recommend similar movies.

Overview

The goal of this project is to recommend movies based on their content. The key features considered include:

• Overview: The movie's plot summary.
• Director: The director of the movie.
• Stars: The main actors in the movie.
• Gross: The gross revenue of the movie.
• IMDB Rating: The rating of the movie on IMDb.

Installation

1. Clone the repository:

   git clone https://github.com/faisal-fida/Content-Based-Filtering-Model.git

2. Install the required libraries:

   pip install -r requirements.txt

Usage

1. Load the Jupyter Notebook content-based filtering model.ipynb in your Jupyter environment.
2. Run the notebook cells sequentially to preprocess the data, extract features, compute similarities, and make movie recommendations.
3. Use the recommend_movies function to get movie recommendations based on a given movie title.

Challenges and Solutions

• Data Cleaning: Missing values in the dataset were filled or removed to ensure data consistency.
• Feature Extraction: Various features were one-hot encoded and normalized to be used in the similarity computation.
• Similarity Computation: Cosine similarity was used to compute the similarity between movies based on the extracted features.

Example

Here is an example of how to get movie recommendations:

movie_title = 'Pulp Fiction'
recommend_movies(movie_title, n=5)

Output:

['Brokeback Mountain', 'Forrest Gump', 'Naked', 'The Green Mile']

Conclusion

This content-based filtering model provides a straightforward way to recommend movies based on their content. By leveraging various features and computing similarities, the model can successfully recommend movies that are similar to a given movie.

For more details, please refer to the Jupyter Notebook in this repository.