Detecting Communities from Bipartite Networks Based on Bipartite Modularities

Discovering communities from networks is one of the important and challenging research topics of social network analysis. Although Newman's modularity is often used for evaluating division of unipartite networks, it is not suitable for evaluating division of bipartite networks that are composed of two types of vertices. To compensate for the situation, Guimera and Barber propose bipartite modularities. This paper discusses the characteristics of these bipartite modularities and proposes another bipartite modularity. Experimental results show that our new bipartite modularity allows one-to-many correspondence between communities of different vertex types.