Overlapping community detection in social networks

Abstract

In recent years, complex networks such as social networks have received great attention due to their popularity, also the need to understand their structure and their usefulness in several domains such as healthcare. We find commonly a community structure, i.e. groups of vertices are more densely connected than to other vertices in the network. Frequently in these communities, we observe some vertex which can occur in more than one community, this phenomenon is called overlapping community. The identification of overlapping community is a crucial task. Existing methods present a high complexity as the size of the network increases. We aim in this paper to present a new method allowing overlapping community detection based on the principle of edge betweenness. Moreover, we perform some comparative experiments to show the effectiveness of our algorithm.