Signed Bipartite Graph Neural Networks

This is our PyTorch implementation code for our paper:

Signed Bipartite Graph Neural Networks (CIKM2021)

arXiv

Introduction

Method

Dataset

For bonanza, house, senate, you can download it from this repository. For review dataset, you can download it in experiments-data folder.

Dependency

In order to run this code, you need to install following dependencies:

pip install torch numpy sklearn tqdm tensorboard

Run Example

python sbgnn.py --lr 5e-3 --seed 222 \
                --dataset_name house1to10-1 --gnn_layer 2 \
                --epoch 2000 --agg AttentionAggregator

Results:

test_auc 0.8498742632577166 
test_f1 0.8592910848549948 
test_macro_f1 0.848896372204643 
test_micro_f1 0.8496114447191806

Citation

Please cite our paper if you use this code in your own work

@inproceedings{huang2021signed,
  title     = {Signed Bipartite Graph Neural Networks},
  author    = {Huang, Junjie and Shen, Huawei and Cao, Qi and Tao, ShuChang and Cheng, Xueqi},
  booktitle = {{CIKM} '21: The 30th {ACM} International Conference on Information
               and Knowledge Management, Virtual Event, Queensland, Australia, November
               1 - 5, 2021},  
  year      = {2021},
  pages     = {740--749},
  publisher = {{ACM}},
  year      = {2021},
  url       = {https://doi.org/10.1145/3459637.3482392},
  doi       = {10.1145/3459637.3482392},
}