CLUSTERING ALGORITHMS FOR CATEGORICAL DATA USING CONCEPTS OF SIGNIFICANCE AND DEPENDENCE OF ATTRIBUTES

Abstract

Clustering categorical data is an essential and integral part of data mining. In this paper, we propose two new algorithms for clustering categorical data, namely, the Standard Deviation of Standard Deviation Significance and Standard Deviation of Standard Deviation Dependence algorithms. The proposed techniques are based mainly on rough set theory, taking into account the significance and dependence of attributes of database concepts. Analysis of the performance of the proposed algorithms compared with others shows their efficiency as well as ability to handle uncertainty together with categorical data.