An interpretable hybrid ensemble model for early academic risk detection with shap-based explainability. (2026). Journal of Soft Computing Exploration, 7(3), 489-501. https://doi.org/10.52465/joscex.v7i3.50