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

作者：Wang Li Na^*; Liu Qian; Zhou Yuan

来源：Information Technology Journal, 2013, 12(20): 5482-5486.

DOI：10.3923/itj.2013.5482.5486

• Between-cluster information
• Categorical data
• Dissimilarity measures
• Fuzzy centroid
• Fuzzy k-modes algorithm
• Misclassifications
• Novel algorithm
• Objective functions

In this study, a new fuzzy centroids clustering for categorical data is presented. The objective function of the fuzzy k-modes algorithm is modified by adding the between-cluster information so as to simultaneously minimize the wifhin-cluster dispersion and enhance the between-cluster separation. Due to the misclassification by using the hard centroids, a fuzzy centroids clustering with the between-cluster information for categorical data is provided. Furthermore, the dissimilarity measure between an object and the centroid at the feature level is given as 1 minus the frequency of the feature value of the object. On several real data sets from UCI, the proposed algorithm is effective and the performance of the novel algorithm outperforms the one with hard-type centroids.

• 出版日期2013
• 单位南京大学