Predicting Students’ Academic Performance Using Utility Based Educational Data Mining

Abstract

Knowledge extracted from educational data can be used by the educators to obtain insights about how the quality of teaching and learning must be improved, how the factors affect the performance of the students and how qualified students can be trained for the industry requirements. This research focuses on classifying a knowledge based system using a set of rules. The main purpose of the study is to analyze the most influencing attributes of the students for their module performance in tertiary education in Sri Lanka. The study has gathered data about students in a reputed degree awarding institute in Sri Lanka and used three different data mining algorithms to predict the influential factors and they have been evaluated for interestingness using objective oriented utility based method. Subsequently, age of the students, their family background with regard to parents’ occupations, average monthly income of the family, their English language fluency level and knowledge of Mathematics were identified as the interesting factors. The findings of this study will positively affect the future decisions made regarding the progress of the students’ performance, quality of the education process and the future of the education provider.