Predicting Postgraduate Students’ Performance using Decision Tree Algorithms
Keywords:
Educational data mining, Student performance prediction, Class imbalance problemAbstract
Abstract
In this study, a data mining model that predicts postgraduate students’ performance using decision tree algorithms was developed. Postgraduate student data collected from the postgraduate school, University of Ibadan and a case study department were pre-processed adequately. Seven different feature selection techniques in Waikato Environment for Knowledge Analysis (WEKA) were used to determine the major attributes that contribute to the prediction of Postgraduate students’ performance. The highest-ranked attributes were used for the analysis using RandomTree, RepTree andJ48 decision tree algorithms in WEKA. The work was evaluated using the AUROC performance metrics for the major classes of interest. Results obtained gave insight into the optimal algorithm for the analysis and rules that could predict postgraduate students’ performance were generated from the developed model.
