Abstract

Many complex systems in the real world such as social networks, biological networks can be represented by networks. These networks own a common key property, called community or cluster. In some cases, communities may overlap with each other. Detecting overlapping communities can give insights into the structures and functions of these systems. Many overlapping community detection algorithms find communities based on betweeness centrality. But the calculation of betweenness suffers high computational complexity. Steve proposed a kind of local betweenness, while it can not be suitable for different networks with community structures in different scales. In this paper, we propose a novel local betweenness index, called community betweenness. Based on this index, we design an efficient overlapping community detection algorithm. Experiments on synthetic and real-world networks show that our method can give better results than other methods using local betweenness.

Paper Details