RHGN

This is the code for the CIKM 2021 Paper: Relation-aware Heterogeneous Graph for User Profiling.

Usage

Raw data and more details in JD-Dataset and Alibaba-dataset

1、Download processed data:

Processed data download link: Dataset

python tb_tmain.py --data_dir ../taobao_data/ --model RHGN --label gender --graph G_ori --gpu 3  # Alibaba-Dataset

python jd_tmain.py --data_dir ../data/ --model RHGN --graph G_ori  --label age --gpu 2  # JD-Dataset

2、Process raw data

python tbdata_process.py    #Alibaba-Dataset

python jddata_process.py     #JD-Dataset

More command can refer run.sh and baseline.sh

Requirements

• torch==1.6.0
  
• torchvision==0.7.0
  
• dgl==0.7.1
  
• scikit-learn==0.23.2
  
• numpy==1.19.1
  
• scipy ==1.5.2
  
• pandas==1.1.2
  
  

Note: The code require your gpu memory not less than 20 GB.

Citation

Please cite our paper if you use the code:

@inproceedings{yan2021relation,
  title={Relation-aware heterogeneous graph for user profiling},
  author={Yan, Qilong and Zhang, Yufeng and Liu, Qiang and Wu, Shu and Wang, Liang},
  booktitle={Proceedings of the 30th ACM International Conference on Information \& Knowledge Management},
  pages={3573--3577},
  year={2021}
}

Contact

qilongy@foxmail.com