Supplementary Materials for "Beyond grades: multi-target deep learning for early academic risk detection"
Miguel Angel Rodríguez Ortiz, Luis Anido-Rifón, Pedro C. Santana-Mancilla
This repository contains the supplementary materials associated with the article: “Beyond Grades: Multi-Target Deep Learning for Early Academic Risk Detection” The materials support the transparency, reproducibility, interpretability, and pedagogical analysis of the leakage-free multi-output modeling framework reported in the manuscript. Contents: Supplementary Appendix A:Systematic review coding matrix used to summarize the prevalence of variable categories across 55 empirical studies on academic-performance prediction in higher education. This appendix supports Figure 2 in the manuscript. Supplementary Appendix A1:Dataset variable dictionary, including formulas, units, source platforms, target/predictor roles, and leakage classification for the originally engineered variables and the final leakage-free predictor set. Supplementary Appendix B:Literature benchmark of reported models, prediction targets, and performance metrics from prior studies, used to contextualize the comparative analysis in the manuscript. Supplementary Appendix C:Extended SHAP interpretability visualizations, leakage-free behavioral clustering results, and per-cutoff SHAP summaries for early prediction analyses. These materials complement the main manuscript and provide additional transparency for the reported findings, including the data-leakage audit, feature construction decisions, model interpretation, and supplementary analyses.