Predictive Learning Analytics Using SAP and AI for Early Identification of Academic Disparities

Abstract

Academic disparities — systematic performance gaps rooted in socioeconomic, demographic, and institutional factors — represent one of the most persistent challenges in modern education. Traditional approaches to identifying at-risk learners rely on lagging indicators such as end-of-term grades or attendance records, by which point meaningful intervention windows have often closed. This paper presents a Predictive Learning Analytics (PLA) framework that integrates SAP Student Lifecycle Management (SAP SLCM), SAP Analytics Cloud (SAC), and advanced Artificial Intelligence models — including Gradient Boosting classifiers, LSTM temporal networks, and Explainable AI techniques — to enable early, accurate, and equitable identification of academic disparity trajectories. Implemented across three higher education institutions in India and the Gulf region involving 28,740 enrolled students over three academic years, the proposed framework achieved a prediction sensitivity of 89.4% for at-risk identification at the eight-week mark of a semester, a false-positive rate of 6.2%, and a 31.7% reduction in end-of-year academic failure rates among cohorts receiving AI-triggered interventions. Demographic fairness analysis confirmed equitable prediction accuracy across gender, socioeconomic, and regional subgroups (disparity index < 0.06). The study establishes a replicable, ethically governed model for AI-augmented educational equity in SAP-integrated institutional environments.