Community Detection in networks by Dynamical Optimal Transport Formulation

TL;DR

This paper introduces a flexible optimal transport-based method for community detection in networks, leveraging recent advances to improve accuracy and alignment with node metadata across synthetic and real datasets.

Contribution

The paper presents a novel OT-based community detection approach that incorporates traffic penalization, enhancing flexibility and performance over existing OT methods.

Findings

Achieves comparable or better results than existing OT methods on synthetic networks.

Finds communities more aligned with node metadata in real-world networks.

Demonstrates the effectiveness of geometric approaches in complex network analysis.

Abstract

Detecting communities in networks is important in various domains of applications. While a variety of methods exists to perform this task, recent efforts propose Optimal Transport (OT) principles combined with the geometric notion of Ollivier-Ricci curvature to classify nodes into groups by rigorously comparing the information encoded into nodes' neighborhoods. We present an OT-based approach that exploits recent advances in OT theory to allow tuning for traffic penalization, which enforces different transportation schemes. As a result, our model can flexibly capture different scenarios and thus increase performance accuracy in recovering communities, compared to standard OT-based formulations. We test the performance of our algorithm in both synthetic and real networks, achieving a comparable or better performance than other OT-based methods in the former case, while finding communities more aligned with node metadata in real data. This pushes further our understanding of geometric approaches in their ability to capture patterns in complex networks.