Soheila Gheisari, Hamid Salarian (2026) — arXiv:2606.20617 (cs.CY; cs.LG) 📄 Full text (arXiv)
Gheisari and Salarian (2026) apply supervised machine learning classification to identify at-risk students before they withdraw from higher education programs. The study evaluates Logistic Regression, Random Forest, Support Vector Machines (SVM), and K-Nearest Neighbors (KNN) using academic performance, demographic data, and enrollment records. Logistic Regression and linear SVM achieved the highest predictive accuracy, demonstrating ML's capability to detect at-risk students for proactive intervention. This learning-analytics research contributes to student-modeling for dropout-reduction in higher-ed contexts, providing a data-driven foundation for strategic retention decisions and connecting to the broader ai-adoption-training-public-sector discussion on AI-supported institutional decision-making.
Related Pages
- at-risk-students-ml-prediction -- This page