Modeling academic achievement of UUM graduate using descriptive and predictive data mining

Abstract

The selection of data mining approaches is based on the ability of data mining as a powerful tool for academic analysis purposes. In higher educational institution, data mining can be used for the process of uncovering hidden trends and patterns that help the institutions in forecasting the students’ achievement. Today, the abilities such as intelligence, skill and CGPA are identified as a main factor for academic achievement. In essence, it is a common practice to use the CGPA as an indicator of students’ academic achievement. However, measuring the academic achievement is not an easy task. The purpose of this study is to investigate factors that associated with academic achievement for the undergraduate students of University Utara Malaysia (UUM) based on College of Arts and Sciences (CAS), College of Business (COB), and College of Law, Governance and International Studies (COLGIS) using descriptive and predictive data mining. Prior research indicates that students and faculty shared a common perception of the skills necessary for success in the degree programs. Based on the results extracted from descriptive and predictive data mining, empirical investigation using logistic and neural networks reveal that factors such as family income, race and language skill have significant association with academic achievement.