Extending k-Representative Clustering Algorithm with an Information Theoretic-based Dissimilarity Measure for Categorical Objects

This paper aims at introducing a new dissimilarity measure for categorical objects into an extension of k-representative algorithm for clustering categorical data. Basically, the proposed dissimilarity measure is based on an information theoretic definition of similarity introduced by Lin [15] that considers the amount of information of two values in the domain set. In order to demonstrate the efficiency of the extended k-representative algorithm with the new dissimilarity measure, we conduct a series of experiments on real datasets from UCI Machine Learning Repository and compare the result with several previously developed algorithms for clustering categorical data.