Overlapping Community Structure Detection and Evolution in Networks

Nan Du,Bin Wu,Bai Wang
2008-01-01
Abstract:Recent researches have seen that rich interactions among entities in nature and society bring about complex networks with community structures which could represent interacting patterns in protein-protein networks, or circles of friends and families in social networks. Although the investigation of the community structure has motivated many kinds of algorithms, most of them only find separated communities. However, for the vast majority of real-world networks, few communities really disjoint with each other. As a matter of fact, they often overlap to some extent. Being contrary to earlier methods, in this paper, we propose a new algorithm COCD(Clique-based Overlapping Community Detection) to efficiently mine overlapping communities in large-scale networks. COCD is only based on the network topology and does not require any priori knowledge about the original partition of the network. We find that the extent to which communities overlap with each other actually depends on the specific types of the links in the network. Moreover, due to the frequent changes of the interactions among entities, the associated networks are able to develop and evolve over time. Consequently, uncovering the underlying micro mechanisms that govern the dynamics of the various communities is critical to have a better understanding of the macro network evolution as a whole. Therefore, based on COCD, we propose a novel method to track the evolution of communities in a series of time snapshots. We show that the community’s lifetime and persistence depend not only on its internal structure, but also on the constitution of the network. Our studies on the real networks from different domains demonstrate that the existence and evolution of overlapping communities has meaningful and non-trivial statistics. ∗This work is supported by the National Science Foundation of China under grant number 60402011, and the National Science and Technology Support Program of China under Grant No.2006BAH03B05. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Copyright 200X ACM X-XXXXX-XX-X/XX/XX ...$5.00.
What problem does this paper attempt to address?