Line Graph Neural Networks for Link Prediction

Introduction

Pytorch implementation of Line Graph Neural Networks for Link Prediction.

If using this code , please cite our paper.

@article{cai2020line,
  title={Line graph neural networks for link prediction},
  author={Cai, Lei and Li, Jundong and Wang, Jie and Ji, Shuiwang},
  journal={arXiv preprint arXiv:2010.10046},
  year={2020}
}

System requirement

Programming language

Python 3.5 +

Python Packages

Pytorch , Numpy , Networkx, PyTorch Geometric

Datasets

Raw datasets are obtained from https://noesis.ikor.org/datasets/link-prediction. Datasets are processed and saved into mat file.

Training

Train the network

cd LGLP/Python
python Main.py --data-name=BUP --test-ratio=0.5

Acknowlegdements

Part of code borrow from https://github.com/muhanzhang/SEAL and https://github.com/muhanzhang/DGCNN. Thanks for their excellent work!