https://github.com/ravindramohith/movie_recommender_system

A movie recommendation system utilizing a Graph Neural Network (GNN) framework implemented in Jupyter Notebook
https://github.com/ravindramohith/movie_recommender_system

bipartite-graphs graph graph-neural-networks lightgcn pygeometric pytorch self-supervised-learning supervised-learning

Last synced: 15 days ago
JSON representation

A movie recommendation system utilizing a Graph Neural Network (GNN) framework implemented in Jupyter Notebook

• Host: GitHub
• URL: https://github.com/ravindramohith/movie_recommender_system
• Owner: ravindramohith
• License: apache-2.0
• Created: 2024-08-12T22:16:56.000Z (6 months ago)
• Default Branch: main
• Last Pushed: 2024-12-12T12:33:10.000Z (about 1 month ago)
• Last Synced: 2024-12-12T13:34:52.202Z (about 1 month ago)
• Topics: bipartite-graphs, graph, graph-neural-networks, lightgcn, pygeometric, pytorch, self-supervised-learning, supervised-learning
• Language: Jupyter Notebook
• Homepage:
• Size: 119 KB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Movie Recommender System

This repository contains the implementation of an advanced movie recommender system using LightGCN in PyTorch Geometric (PyG). The system integrates message passing and embeddings from a bipartite graph of users and movies, utilizing both supervised and self-supervised learning techniques to enhance recommendation quality.

## Features

- **Graph Neural Networks**: Built using LightGCN in PyG, integrating message passing and embeddings.

- **Learning Methods**: Utilizes both supervised and self-supervised learning to improve recommendation quality.

- **Loss Functions**: Trained using BPR (Bayesian Personalized Ranking) and RMSE (Root Mean Squared Error) loss functions.

- **Optimizer**: Uses Adam optimizer with learning rate decay for optimized convergence.

- **Data Processing**: Efficiently processes and evaluates graph data with sparse matrices and edge indices.

- **Evaluation Metrics**: Uses recall, precision, and NDCG (Normalized Discounted Cumulative Gain) metrics for evaluation.

## Installation

1. Clone the repository:

   ```sh

   git clone https://github.com/ravindramohith/movie_recommender_system.git

   ```

2. Navigate to the project directory:

   ```sh

   cd movie_recommender_system

   ```

## Usage

- **Supervised Learning**: 

  - The supervised learning implementation can be found in `recommender-system-using-supervised-gnn_modified.ipynb`.

  - Follow the notebook to train and evaluate the supervised recommender system.

- **Self-Supervised Learning**:

  - The self-supervised learning implementation can be found in `recommender-system-using-self-supervised-gnn_modified.ipynb`.

  - Follow the notebook to train and evaluate the self-supervised recommender system.

## Evaluation

The model performance is evaluated using the following metrics:

- **Recall**: Measures the ability of the recommender system to capture relevant items.

- **Precision**: Measures the accuracy of the recommended items.

- **NDCG**: Measures the ranking quality of the recommendations.

## Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

## Contact

For any questions or suggestions, please contact me through [[email protected]](https://github.com/ravindramohith).