/UCODE

Primary LanguagePython

UCoDe: Unified Community Detection with Graph Convolutional Networks

Implementation of UCoDe for overlapping and non-overlapping community detection as proposed in the paper:

This implementation is written in Python 3.

Requirements

  • networkx
  • numpy
  • pandas
  • scikit-learn
  • scipy
  • torch==1.10.0

Installation

You can install all the required packages using the following command:

$conda create --name --file requirements.txt

Run the code

Run main.py for an example result of non-overlapping community detection.
Run mainOUCODE.py for an example result of overlapping community detection.

Cite

@article{moradan2023ucode,
author = {Moradan, Atefeh and Draganov, Andrew and Mottin, Davide and Assent, Ira},
title = {UCoDe: unified community detection with graph convolutional networks},
journal = {Machine Learning},
pages={1--24},
year = {2023},
publisher = {Springer},
}