Community detection in complex networks with an ambiguous structure using central node based link prediction

Abstract

Community detection in complex networks has aroused wide attention, since it can find some useful information hidden in the networks. Many different community detection algorithms have been proposed to detect the communities in a variety of networks. However, as the ratio of each node connecting with the nodes in other communities increases, namely, the community structure of networks becomes unclear, the performance of most existing community detection algorithms will considerately deteriorate. As a method of finding missing information, link prediction can predict undiscovered edges in the networks. However, the existing link prediction based community detection algorithms cannot deal with the networks with an ambiguous community structure, namely, the networks having a mixing parameter greater than 0.5. In this paper, we design a new strategy of link prediction and propose a community detection algorithm based on this strategy to detect the communities in complex networks, especially for the networks with an ambiguous community structure. Experimental results on synthetic benchmark networks and real-world networks indicate that the proposed community detection algorithm outperforms five state-of-the-art community detection algorithms, especially for those without a clear community structure.