COVID-19 patients classification using Graph neural network on a Heterogeneous graph

Introduction

This repository contains code for node classification on a heterogeneous graph, concretely, patient node classification on a COVID-19 graph.

Requirements

scipy
numpy
pandas
pytorch==1.6.0
dgl==0.4.3post2

Installation

Clone the repository and install the package

git clone https://github.com/KienMN/COVID-19-in-Korea-graph.git
cd STGNN-for-Covid-in-Korea
pip install -e .

Install package using pip

pip install git+https://github.com/KienMN/COVID-19-in-Korea-graph.git

Main components

Preprocessing

Process dataset from CSV file to DGL graph data structure. Check graph_neural_network/preprocessing.py for more details.

Models

This module contains Relational Graph convolution network model for Heterogeneous graph which conducts graph convolution on each relationship. Check graph_neural_network/models.py for more details.

Citation

This source code is for the paper:

Kien Mai Ngoc, Minho Lee. "COVID-19 patients classification using Graph neural network on a Heterogeneous graph". In: Proc. of the International Conference on Convergence Content 2020, The Korea Contents Society, 2020, 13-14. URL: https://www.dbpia.co.kr/journal/articleDetail?nodeId=NODE10506109

Bibtex:

@inproceedings{mai2020graph,
  author={Kien, Mai Ngoc and Minho, Lee},
  title={COVID-19 patients classification using Graph neural network on a Heterogeneous graph},
  booktitle={Proc. of the International Conference on Convergence Content 2020},
  year={2020},
  pages={13--14},
  publisher={The Korea Contents Society},
  url={https://www.dbpia.co.kr/journal/articleDetail?nodeId=NODE10506109}
}

References

DGL Documentation: https://docs.dgl.ai/en/0.4.x/tutorials/basics/5_hetero.html
COVID-19 in Korea dataset: https://www.kaggle.com/kimjihoo/coronavirusdataset