Overlapping Community Detection Based on Information Dynamics

Identifying overlapping communities is essential for analyzing network structures, exploring the interactions of groups, studying network functions, and obtaining insight into the dynamics of networks. Many algorithms have been proposed for detecting overlapping communities but identifying the intrinsic communities is still a non-trivial problem because of the difficulties with parameter tuning, user bias criteria, and the lack of ground truth information. In this paper, we propose a new model called OCDID (Overlapping Community Detection based on Information Dynamics) to uncover the overlapping communities, which treats the network as a dynamical system that allows an individual to communicate and share information with its neighbors. The information flow in the network is controlled by the underlying topology structure (e.g., the community structure), and the community structure is also reflected by the information dynamics. Overlapping nodes act as bridges between multiple communities and the information from multiple communities flows through these nodes. Thus, the overlapping nodes can be identified by analyzing the information flow among communities. In addition, we use the monotone convergence theorem to confirm the convergence of our model. Experiments based on synthetic and real-world networks demonstrate that in most cases, our proposed approach is superior to other representative algorithms in terms of the quality of overlapping community detection.