SR-HGN

This repo is for source code of Expert Systems with Applications paper "SR-HGN: Semantic-and Relation-Aware Heterogeneous Graph Neural Network". paper

Environment Settings

python==3.8.0
scipy==1.6.2
torch==1.11.0
scikit-learn==1.0.2
torch_geometric==2.0.4
dgl==0.6.1

Dataset

We utilize three benchmark datasets in the paper to perform node classification and node clustering. We provide ACM, DBLP, and IMDB in GoogleDrive.

ACM
DBLP
IMDB

You can create the "data" folder in the root directory, then put the datasets in. Like "/SRHGN/data/acm/...".

How to run

For example, if you want to run SR-HGN on ACM dataset, execute

python main.py --dataset acm

Citation

@article{wang2023sr,
  title={SR-HGN: Semantic-and Relation-Aware Heterogeneous Graph Neural Network},
  author={Wang, Zehong and Yu, Donghua and Li, Qi and Shen, Shigen and Yao, Shuang},
  journal={Expert Systems with Applications},
  volume={224},
  pages={119982},
  year={2023},
  publisher={Elsevier}
}

Contact