A novel possibilistic fuzzy leader clustering algorithm

Abstract

The clusters tend to have vague or imprecise boundaries in some fields such as web mining, since clustering has been widely used. Fuzzy clustering is sensitive to noises and possibilistic clustering is sensitive to the initialization of cluster centers and generates coincident clusters. Based on combination of fuzzy clustering and possibilistic clustering, a novel possibilistic fuzzy leader (PFL) clustering algorithm is proposed in this paper to overcome these shortcomings. Considering the advantages of the leader algorithm in time efficiency and the initialization of cluster, the framework of the leader algorithm is used. In addition, a λ-cut set is defined to process the overlapping clusters adaptively. The comparison of experimental results shows that our proposed algorithm is valid, efficient, and has better accuracy. © 2009 Springer-Verlag Berlin Heidelberg.