A FRAMEWORK FOR STUDENTS’ ACADEMIC PERFORMANCE ANALYSIS USING NAÏVE BAYES CLASSIFIER

Abstract

Educational database of Higher Learning Institutions holds an enormous amount of data that increases every semester. Data mining technique is usually applied to this database to discover underlying information about the students. This paper proposed a framework to predict the performance of first year bachelor students in Computer Science course. Naïve Bayes Classifier was used to extract patterns using WEKA as a Data mining tool in order to build a prediction model. The data were collected from 6 year period intakes from July 2006/2007 until July 2011/2012. From the students’ data, six parameters were selected that are race, gender, family income, university entry mode, and Grade Point Average. By using Naïve Bayes Classifier, it would predict the class label “Grade Point Average” as a categorical value; Poor, Average, and Good. Result from the study shows that the students’ family income, gender, and hometown parameter contribute towards students’ academic performance. The prediction model is useful to the lecturers and management of the faculty in identifying students with weak performance so that they will be able to take necessary actions to improve the students’ academic performance.