A novel fuzzy clustering algorithm with between-cluster information for categorical data

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

35 Scopus Citations
View graph of relations

Author(s)

Detail(s)

Original languageEnglish
Pages (from-to)55-73
Journal / PublicationFuzzy Sets and Systems
Volume215
Publication statusPublished - 16 Mar 2013

Abstract

In this paper, we present a new fuzzy clustering algorithm for categorical data. In the algorithm, the objective function of the fuzzy k-modes algorithm is modified by adding the between-cluster information so that we can simultaneously minimize the within-cluster dispersion and enhance the between-cluster separation. For obtaining the local optimal solutions of the modified objective function, the corresponding update formulas of the membership matrix and the cluster prototypes are strictly derived. The convergence of the proposed algorithm under the optimization framework is proved. On several real data sets from UCI, the performance of the proposed algorithm is studied. The experimental results illustrate that the algorithm is effective and suitable for categorical data sets. © 2012 Elsevier B.V.

Research Area(s)

  • Categorical data, Fuzzy clustering, Optimization objective function, The fuzzy k-modes algorithm