Abstract—Clustering Categorical data is more complicated process than numerical clustering. In this paper the traditional K-Modes algorithm is extended with the weighted measure based on the probability of respective matching attribute value in the data set. The proposed method is experimented with the data sets obtained from UCI data repository. Results prove that the proposed weighted measure is superior to K-Modes.
Index Terms—Clustering, Categorical Data, K-Modes, Probability, Weighted measure
Aranganayagi S. is with the J. K. K. Nataraja College of Arts & Science, Komarapalayam, Tamilnadu, India and doing research in the Department of Computer Science and Applications, Gandhigram Rural University, Gandhigram Tamilnadu India. Member of IAENG, IACSIT: Corresponding author, phone: 0424-2230855, 9842723085; e-mail: firstname.lastname@example.org.
Dr. K. Thangavel is with the Periyar University, Salem, Tamilnadu, Indiaas Professor in Computer Science. (email@example.com).
Cite: Aranganayagi S. and Thangavel K., "Extended K-Modes with Probability Measure," International Journal of Computer Theory and Engineering vol. 2, no. 3, pp. 431-435, 2010.