Overlapping Communities based on Relaxed Cliques

Abstract

Detection of overlapping communities is one of the most important tasks in network science that allows to simulate various real-world objects like social networks. This paper describes an algorithm for finding overlapping communities using relaxed cliques. The use of relaxed cliques makes it possible to find communities (including overlapping ones) with a denser internal structure due to the requirement for more tight connectivity of nodes inside the relaxed clique. The developed algorithm has shown its efficiency in the analysis of communities in synthetic networks built with the LFR (Lancichinetti–Fortunato–Radicchi) graph generator.