摘要

Hybrid methods of spatial credibilistic clustering and particle swarm optimization (SCCPSO) (Wen et al. in Int J Fuzzy Syst 10:174-184, 2008) are validated to be effective, and produce better results than other common methods. In this paper, SCCPSO is further investigated and a modified SCCPSO is put forward by discussing the membership functions and presenting a pre-selection method based on proving an evaluation criterion on the clustering results. The analysis of computational complexity demonstrates the feasibility of the modified SCCPSO. Experiments verify the discussion on the membership functions, the correctness of the evaluation criterion, as well as the effectiveness of the pre-selection method and the modified SCCPSO.

全文