Interval type-2 fuzzy clustering algorithm constructs asymmetric fuzzy boundaries through two fuzzy sets,effectively handling clustered data of varying sizes and densities.The combination of clustering centers and fuzzy fuzzifiers directly affects the clustering outcome and cannot be adaptively determined.To address this issue,a multi-objective alternate particle swarm optimization-based interval type-2 fuzzy clustering algorithm(MOAPSO-IT2FCM)is proposed.Initially,the particle swarm optimization algorithm is introduced to separately encode and optimize the clustering centers and fuzzy fuzzifiers.Subsequently,considering the details and completeness of the clustering,a multi-objective optimization approach is utilized to improve the particle swarm optimization algorithm.Finally,an alternate particle swarm optimization strategy is employed to alternately fix and optimize the clustering centers and fuzzy fuzzifiers,effectively determining their combination.The MOAPSO-IT2FCM algorithm is applied to the segmentation of Berkeley color nature images,and the results demonstrate its ability to effectively determine suitable combinations of clustering centers and fuzzy fuzzifiers,with clustering performance superior to other comparative algorithms.
维普
万方数据
