Affiliations: [a] State Key Lab of Software Engineering, School of Computer, Wuhan University, Wuhan, Hubei, China | [b] School of Information Science and Technology, Jiujiang University, Jiujiang, Jiangxi, China | [c] Vietnam Academy of Science and Technology, Hanoi, Vietnam
Correspondence:
[*]
Corresponding author: Dang Cong Tran, State Key Lab of Software Engineering, School of Computer, Wuhan University, Wuhan, Hubei 430072, China. Tel.: +86 13554284874; E-mail:[email protected]
Abstract: This paper presents an improved approach of particle swarm optimization (PSO) based on new neighborhood search strategy with diversity mechanism and Cauchy mutation operator (denoted EPSONS). Firstly, with a test on thirteen well-known benchmark functions, the proposed algorithm has significant improvement over several other PSO variants for global numerical optimization. The proposed approach is then applied to data clustering. The experimental results on fourteen benchmark data sets including artificial and real-world data sets show that the proposed method outperforms than other comparative clustering algorithms in terms of accuracy and convergence speed.
Keywords: Data clustering, K-means, neighborhood search, global optimization, particle swarm optimization
