Applications. Developments in overlapping clustering has mainly been driven by the concrete needs of applications. For instance, driven by the need to cluster microarray gene expression data,various methods for overlapping clustering[37],[23] and overlapping bi-clustering [38], [39] have been proposed. Even though detecting communities in social networks is a problem that has been studied extensively, only few researchers have addressed the problem of detecting overlapping communities; for a survey, see Fortunato [40, Section 11]. The best known approach to detect overlapping communities is the CFinder algorithm based on clique percolation [41]. According to the CFinder method, communities are discovered by finding k-cliques and merging them when they share k−1 nodes. As a node can belong to multiple cliques, the method generates overlapping communities.