Algorithm of Detecting Overlapping Communities in Complex Networks

Huangbin You,Xuewu Zhang,Huaiyong Fu,Zhuo Zhang,Min Li,Xinnan Fan
DOI: https://doi.org/10.1109/icspcc.2014.6986151
2014-01-01
Abstract:As an important property of complex networks the research of community structure has never stopped. Recently people found some nodes may belong to several communities, so more and more people try to present algorithms to detect the overlapping communities in network. In addition, for a large scale complex network, a algorithm with lower time complexity and higher classification accuracy is required. Thus we propose an algorithm for detecting overlapping community structure based on LFM algorithm. And the fitness function is applied to the local optimization. The first step of our method is to choose maximum degree node and its some special adjacent nodes (fitness function satisfies the condition) as the initial community, then expand the initial community by repeatedly adding qualified nodes to it. Many classical network datasets are employed to test the presented method. Experimental results reveal that overlapping communities can be successfully detected from complex networks by our algorithm, and also demonstrate that our method has higher division accuracy and a running time of O(n(2)) in the worst case.
What problem does this paper attempt to address?