CM-HGNN

Code

This is the source code for KBS 2022 Paper: Category-aware Multi-relation Heterogeneous Graph Neural Networks for session-based recommendation

Requirements

• Python 3
• PyTorch >= 1.3.0
• tqdm

Usage

Data preprocessing:

The code for session data preprocessing can refer to SR-GNN. In addition, we extracted category labels from the three original datasets and recorded them in category.txt. The source of the original datasets can be found in the paper

Train and evaluate the model:

python main.py --dataset diginetica

Citation

@article{xu2022category,
  title={Category-aware Multi-relation Heterogeneous Graph Neural Networks for session-based recommendation},
  author={Xu, Hao and Yang, Bo and Liu, Xiangkun and Fan, Wenqi and Li, Qing},
  journal={Knowledge-Based Systems},
  pages={109246},
  year={2022},
  publisher={Elsevier}
}