This repository contains Pytorch codes and datasets for the paper:

Zixuan Yi, Iadh Ounis and Craig MacDonald (2023). Graph Contrastive Learning with Positional Representation for Recommendation. In ECIR'23, Dublin, Ireland, April 2-6, 2023.

If you want to use our codes and datasets in your research, please cite:

@inproceedings{yi2023graph,
  title={Graph Contrastive Learning with Positional Representation for Recommendation},
  author={Yi,Zixuan and Ounis, Iadh and Macdonald, Craig},
  booktitle={European Conference on Information Retrieval},
  year={2023},
  organization={Springer}
}

The codes of PGCL are implemented and tested under the following development environment:

• numba==0.53.1
• numpy==1.20.3
• scipy==1.6.2
• torch>=1.7.0

• Configure the xx.conf file in the directory named conf. (xx is the name of the model you want to run)
• Run main.py and choose the model you want to run.

Three datasets are adopted to evaluate PGCL: Yelp, Gowalla, and Amazon-Kindle. The user-item pair $(u_i, v_j)$ in the adjacent matrix is set as 1, if user $u_i$ has rated item $v_j$ in Yelp, or if user $u_i$ has check in venue $v_j$ in Gowalla, or if user $u_i$ has purchased item $v_j$ in Amazon-Kindle. We filtered out users and items with too few interactions.

Will upload the instructions recently.