This project is a clean and lightweight implementation of two popular recommendation system algorithms: Neural Collaborative Filtering (NCF) and Deep Factorization Machine (DFM) in PyTorch. It is designed to be easy to understand and use for anyone interested in the field of recommendation systems.
Make sure you have Python installed on your system. All required packages are listed in requirements.txt.
Clone the repository to your local machine:
git clone https://github.com/ChenXuanting/RecommenderSystem.gitNavigate to the cloned repository, and install the required packages:
pip install -r requirements.txtThe project comes with a preprocessed version of the classic MovieLens 100K dataset in CSV format, located in the dataset folder.
Additionally, you can use Amazon review datasets by modifying the ratings_name variable in main.py. The available Amazon dataset names can be checked at Amazon Review Datasets.
To train a model, run the following command:
python main.py <model_name>Available model names are:
NCF: For Neural Collaborative FilteringDFM: For Deep Factorization Machine
For example, to train the Neural Collaborative Filtering model, run:
python main.py NCFYou can tune the hyperparameters of the models directly in the main.py file.
The performance of the models is evaluated using two metrics:
- Hit Rate at K
- Normalized Discounted Cumulative Gain (NDCG) at K
We have implemented negative sampling and leave-one-out evaluation as proposed in the original papers. Related parameters, such as the number of negative samples, can be adjusted in the main.py file.
The performance is evaluated as a 10-time average on the MovieLens 100K dataset. The matrix sparsity is 0.937.
| Model | Hit Rate @ 10 | NDCG @ 10 |
|---|---|---|
| NCF | 0.5815 | 0.460 |
| DFM | 0.6036 | 0.486 |
Contributions to improve the project are welcome. Feel free to fork the repository and submit pull requests.
This project is open-source and available under the MIT LICENSE.