An interpretable feature-augmented stacked ensemble framework for student dropout prediction
The results show that the eSDP framework achieves good predictive performance and high interpretability and comparative performance assessment for predicting student dropout and providing educational decision support.