Betweenness-driven overlapping label propagation community detection

Tahalea, Sylvert Prian and Krész, Miklós (2025) Betweenness-driven overlapping label propagation community detection. ANNALES MATHEMATICAE ET INFORMATICAE, 61. pp. 229-247. ISSN 1787-6117

Preview

Text 229_247_krész.pdf - Published Version Download (563kB) | Preview

Abstract

Community detection holds significant value in discovering hidden structures in complex networks. In this paper, we propose a betweennessdriven community detection based on the label propagation algorithm. First, at the multiple labels’ assignment phase, we detect communities using label propagation and apply labels for the nodes using the betweenness and degree centrality as references. Second, we refine the modularity and stability using several configurations, such as global modularity and stability pruning, to avoid nodes that have not changed for several iterations. This algorithm was tested with the most common datasets, such as Zachary’s Karate Club Network, Polbooks, Football, and 12 LFR synthetic datasets, which resulted in improved scores on modularity, overlapping normalised mutual information, omega index, generalized F1-score, and extended various pieces of information.

Actions (login required)

Edit Item