yiyang-wang / ie-hgcn

Interpretable and Efficient Heterogeneous Graph Convolutional Network, IEEE TKDE 2021

ie-HGCN

This is the source code of our paper published in IEEE TKDE 2021 paper: Interpretable and Efficient Heterogeneous Graph Convolutional Network.

IEEE Xplore: https://ieeexplore.ieee.org/document/9508875

Pre-print version: https://arxiv.org/abs/2005.13183

Datasets

The original dataset of IMDB is downloaded from HetRec 2011

The original datasets of ACM and DBLP are provided by the authors of [WWW 2019] [HAN] Heterogeneous Graph Attention Network

The preprocessed dataset are too large to be posted on GitHub. You can download preprocessed dataset and preprocessing script from the following links:

Baidu cloud drive: https://pan.baidu.com/s/1uTqp2H9a0bQImcjEE3HUjw, extraction code "dqbc".

OneDrive cloud drive: https://tinyurl.com/tkde-hgcn.

We also provide a concise version of source code in both of the two above cloud drives.

Please feel free to E-mail me [email protected] if you cannot download them successfully.

Requirements

• PyTorch
• NumPy
• scikit-learn

Reference

If you make advantage of ie-HGCN in your research, please cite our work as follows:

@article{yang2021interpretable,
  title={Interpretable and efficient heterogeneous graph convolutional network},
  author={Yang, Yaming and Guan, Ziyu and Li, Jianxin and Zhao, Wei and Cui, Jiangtao and Wang, Quan},
  journal={IEEE Transactions on Knowledge and Data Engineering},
  year={2021},
  publisher={IEEE}
}