Stable community cores in complex networks

By

Massoud Seifi, Jean-Loup Guillaume, Ivan Junier, Jean-Baptiste Rouquier and Svilen Iskrov

Proceedings of the 3rd Workshop on Complex Networks (CompleNet 2012), Melbourne, Florida

Abstract

Complex networks are generally composed of dense sub-networks called communities. Many algorithms have been proposed to automatically detect such communities. However, they are often unstable and behave non-deterministically. We propose here to use this non-determinism in order to compute groups of nodes on which community detection algorithms agree most of the time.We show that these groups of nodes, called community cores, are more similar to Ground Truth than communities in real and arti cial networks. Furthermore, we show that in contrary to the classical approaches, we can reveal the absence of community structure in random graphs.

This entry was posted in Papers and tagged ,