| dc.description.abstract | One of the main problems in Universitas Terbuka (UT) is a significant number of non-active students. This problem may potentially be minimized if the students’ characteristics are understood. The educational databases often have hidden knowledge about the students. This knowledge can be utilized as the materials to improve the treatment program for students. This research proposes a data mining method by means of cluster analysis to analyze student characteristics using cluster ensemble algorithm, so called, algCEBMDC, a clustering algorithm for mixed categorical and numerical data. The experiment data obtained from the Computer Center of UT in first semester 2008/2009, consists of 4,132 non-active students of Mathematics Study Program of UT. The experiment results show that there are certain groups of students with similar characteristics which could provide the management with insights as bases to construct further strategy of treatment for non-active students. | id |
