A case study of knowledge discovery on academic achievement, student desertion and student retention

Abstract

This paper presents an applied research study on knowledge discovery based on academic data analysis. The main objectives of this research study were to get knowledge about academic achievement success and failure, student retention and student desertion. Automatic clustering and decision rule data mining techniques were used. The application of the C-mean algorithm on statistically homogeneous data subsets provided a group of clusters, which have been qualitatively described. Using selected clusters a decision rule study based on C4.5 algorithm generated a set of decision rules for the four research topics of the study. These decision rules have been interpreted, finding global content and specific content knowledge. The findings of the research study have been well evaluated by experts in aspects such as validity, novelty, simplicity and utility. In the last stage of the knowledge discovery process, and based on the knowledge findings, various strategic actions to improve academic processes are proposed.