Community Detection Based on Symbiotic Organisms Search and Neighborhood Information
Jing Xiao,Chao Wang,Xiao-Ke Xu
DOI: https://doi.org/10.1109/tcss.2019.2941988
2019-01-01
IEEE Transactions on Computational Social Systems
Abstract:Modularity optimization methods based on nature-inspired metaheuristics are popular and competent for community detection in complex networks. However, on some real-world networks with complex and vague structures, most contemporary algorithms are difficult to obtain the global optimal partition. There are two key factors that are seriously affecting the global optimization capability: one is the convergence performance of the incorporated optimization strategy and the other is the sufficient and rational utilization of network topological information. In this article, a novel community detection method is proposed, named symbiotic organisms search community detection (SOSCD). The bio-inspired metaheuristic [symbiotic organisms search (SOS)] is discretized and utilized as the optimization strategy to improve global convergence performance of modularity optimization. Meanwhile, by utilizing neighborhood information of each node to guide community optimization, two different local search (LS) schemes are designed to intensify exploitation and, thus, assisting the global search, including the neighbor-based community modification (NCM) and the neighbor-based LS (NLS). Experimental results on both synthetic and real-world networks have validated the effectiveness and superiority of the proposed operations in SOSCD. Moreover, SOSCD can significantly improve the precision and stability of the identified optimal partition, comparing with many state-of-the-art modularity optimization algorithms.