Sequential RecSys: Open-Source Library Development and Algorithms Evaluation

Welcome to the Sequential RecSys project! This open-source library is dedicated to the implementation and evaluation of three recommendation algorithms on three diverse datasets: MovieLens 1M, Amazon Beauty, and Steam.

Datasets

Dataset Variety: Our library supports three distinct datasets for algorithm evaluation:

• MovieLens 1M: A well-known movie rating dataset.
• Amazon Beauty: A dataset containing product reviews and ratings.
• Steam Dataset: Data from the popular gaming platform Steam, including user interactions with games.

Three Recommendation Algorithms:

We have implemented three state-of-the-art recommendation algorithms, each designed to cater to different recommendation scenarios.

To train and run the algorithms, use the following commands:

• SASRec Algorithm:

  python train_sasrec.py --dataset=path/to/the/dataset --train_dir=default

• SR-GNN Algorithm:

  python train_srgnn.py --dataset_path=path/to/the/dataset

• CASER Algorithm:

  python train_caser.py

  Contact

If you have any questions, suggestions, or need assistance, please feel free to contact us:

• Email : [email protected]