This repository contains a Python 3.6 implementation of a Graph Summarization framework based on Szemerédi's Regulairty Lemma for the task of separating structure from noise in large graphs, as described in:
Marco Fiorucci, Francesco Pelosin and Marcello Pelillo. Separating Structure from Noise in Large Graphs Using the Regularity Lemma. Pattern Recognition 2020
Please cite our paper if you use this code in your own work:
@article{Fiorucci2020,
title = "Separating Structure from Noise in Large Graphs Using the Regularity Lemma",
journal = "Pattern Recognition",
volume = "98",
pages = "107070",
year = "2020",
author = "Marco Fiorucci and Francesco Pelosin and Marcello Pelillo"
}
The packages required are in the file requirements.txt
We suggest to create a virtualenv and install the packages by just running pip install -r requirements.txt.
To replicate the experiments in the paper just run sh experiment.sh if you are in a Windows system you can just sequentially run the commands specified in the latter file.