Introduction

One of my colleagues asked me: "What's your most practical skill in network science?" I told him:" community detection"!

Whether you are a network science beginner, enthusiast, or expert, whether you study network data or other data networking, community detection technology will be able to accompany you at every stage of network analysis --- from Network Data Preprocessing and Analyse, Network Visualization, Network Advanced Insight Acquisition...

Community detection can be an organic part of your own models! The VGAER we developed provides an opportunity to combine cutting-edge GNN methods.

Come and try!

VGAER

Simple and efficient -- a novel unsupervised community detection with the fusion of modularity and network structure:

Requirement

dgl==0.8.0.post1

matplotlib==3.5.1

networkx==2.7.1

numpy==1.22.3

pandas==1.4.1

scikit_learn==1.0.2

scipy==1.8.0

seaborn==0.11.2

torch==1.11.0

Citation

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

@inproceedings{qiu2022VGAER,\ title={VGAER: Graph Neural Network Reconstruction based Community Detection},\ author={Qiu, Chenyang and Huang, Zhaoci and Xu, Wenzhe and Li, Huijia},
booktitle={AAAI: DLG-AAAI'22},
year={2022}
}