Prediction of Underperforming Students Using Boosting Ensemble Methods

Key Points

• The aim is to predict underperforming students using boosting and ensemble methods on public datasets.
• Comparison of boosting and ensemble methods
• Evaluation of models including AdaBoost, CatBoost, and Random Forest
• Utilization of three public datasets
• Models provide early warning for at-risk students
• Proposed dashboard concept aids teachers in monitoring performance
• Supports interpretable educational interventions