Thesis (BS Computer Science) -- University of the Philippines Mindanao, 2009

This study focused on implementing a system which clusters categorical data with or without missing values using two modified K-modes algorithms namely, the available case analysis and the adaptive imputation. Files with .txt extensions are the only ones accepted the system. The file or data set must only contain the data points or the instances of a given data. The original K-modes algorithm was first implemented and then modified according to the available case analysis algorithm and adaptive imputation algorithm. The available case analysis has a modified distance measure and computation for the few cluster centers to cater data sets with missing values. The adaptive imputation also has a different dissimilarity measure and computation for the new cluster center; it applied imputation on the data during the first iteration. The results were displayed in tabular form along with its final cluster centers. File uploading was made available in the system for the users ease. It can upload at most four data sets at a time