The Fuzzy k-representatives Algorithm for Clustering Categorical Data

Abstract

This paper presents an algorithm which extends the fuzzy k-means algorithm for clustering categorical objects. In this algorithm, use a simple matching dissimilarity measure to deal with categorical objects and uses a frequency-based method to update the representative objects in the clustering process to minimize the clustering cost function. The fuzzy krepresentative algorithm is that it not only partition objects into clusters but also shows how confident an object is assigned to a cluster. The confidence is determined by the dissimilarity measures of an object to all cluster representatives. This paper is provided a method to find the fuzzy cluster representatives when the simple matching dissimilarity measure is used for categorical objects.