Academic Success Prediction based on Important Student Data Selected via Multi-objective Evolutionary Computation

Abstract

This paper proposes an academic success prediction modeling approach that can be used for student advising, in which a multi-objective evolutionary computation approach is applied that automatically selects important explanatory variables suitable to predict academic success and construct multiple predictive models based on machine learning. Numerical experiments using actual student data suggest that it is possible to construct predictive models in considering the trade-off of prediction performance and model interpretability.