On Accuracy of Community Structure Discovery Algorithms

TL;DR

This paper systematically compares eleven community detection algorithms on realistic artificial networks, revealing the influence of network size and link proportions on their performance, with Infomap leading overall.

Contribution

It provides a comprehensive performance evaluation of main community detection algorithms using realistic simulated data and the NMI measure.

Findings

Infomap outperforms other algorithms in accuracy.

Network size and intra/inter-community link ratio significantly affect performance.

Walktrap, SpinGlass, and Louvain also show strong results.

Abstract

Community structure discovery in complex networks is a quite challenging problem spanning many applications in various disciplines such as biology, social network and physics. Emerging from various approaches numerous algorithms have been proposed to tackle this problem. Nevertheless little attention has been devoted to compare their efficiency on realistic simulated data. To better understand their relative performances, we evaluate systematically eleven algorithms covering the main approaches. The Normalized Mutual Information (NMI) measure is used to assess the quality of the discovered community structure from controlled artificial networks with realistic topological properties. Results show that along with the network size, the average proportion of intra-community to inter-community links is the most influential parameter on performances. Overall, "Infomap" is the leading algorithm, followed by "Walktrap", "SpinGlass" and "Louvain" which also achieve good consistency.