Beta-RecSys an open source project for Building, Evaluating and Tuning Automated Recommender Systems. Beta-RecSys aims to provide a practical data toolkit for building end-to-end recommendation systems in a standardized way. It provided means for dataset preparation and splitting using common strategies, a generalized model engine for implementing recommender models using Pytorch with a lot of models available out-of-the-box, as well as a unified training, validation, tuning and testing pipeline. Furthermore, Beta-RecSys is designed to be both modular and extensible, enabling new models to be quickly added to the framework. It is deployable in a wide range of environments via pre-built docker containers and supports distributed parameter tuning using Ray.

Details of each folder in this repository are

• beta_rec: data processing, model definition and utility functions
• configs: configurations of system information, data splitting and automatic hyper-parameter tuning
• docs: documentation for the Beta-RecSys system
• examples: practices of training various neural recommendation models
• tests: testing codes with the Pytest framework

For a more detailed overview of the repository, please see the documents at the wiki page.

Currently we are supporting Python 3 and PyTorch.

1. Install Anaconda with Python >= 3.6. Miniconda is a quick way to get started.

2. Install and record the installed files

   $ git clone https://github.com/beta-team/beta-recsys.git
   $ cd $project_path$
   $ python setup.py install --record files.txt
   

To uninstall the Beta-RecSys

$ cd $project_path$
$ xargs rm -rf < files.txt

1. To train a Neural Graph Collaborative Filtering with default configurations, you can run

python examples/train_ngcf.py

If you want to change training configurations, such as the used dataset and the range of hyper-paramters, you can change the default NGCF configuration file or create a new one.

2. To try new datasets, you can ceate a new dataset script in beta-rec/datasets by referring to how the movielens dataset is dealt with.
3. To define a new model, you can ceate a new model script in beta-rec/models by referring to how the NGCF model is defined.

Note - To conveniently check system information during model running, we implement funcnationlities of output logging and system monitoring in beta-rec/utils.

The following is a list of recommender models currently available in the repository.

• MF is an implementation of Matrix Factorization
• NCF is an implementation of Neural Collaborative Filtering.
• NGCF is an implementation of Neural Graph Collaborative Filtering.
• Triple2Vec is an implementation of Representing and Recommending Shopping Baskets with Complementarity, Compatibility, and Loyalty.
• VBCAR is an implementation of Variational Bayesian Context-aware Representation for Grocery Recommendation.
• LightGCN is an implementation of LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation.

This project welcomes contributions and suggestions. Please make sure to read the Contributing Guide before creating a pull request.

If you use Beta-RecSys in you research, we would appreciate citations to the following paper:

• Zaiqiao Meng, Richard McCreadie, Craig Macdonald, Iadh Ounis, Shangsong Liang, Siwei Liu, Guangtao Zeng, Junhua Liang, Yucheng Liang, Qiang Zhang, Yaxiong Wu, Xi Wang. BETA-Rec: Build, Evaluate and Tune Automated Recommender Systems. RecSys 2020.