Predictive Analytics for Educational Equity: A Machine Learning Approach to Identifying Learning Gaps in Low-Resource Schools

Abstract: Educational inequality remains a persistent and systemic challenge in low-resource school systems, where limited instructional capacity, unstable learning environments, and delayed identification of academic difficulties contribute to widening learning gaps. In such contexts, traditional assessment practices often detect failure only after significant learning loss has occurred, reducing the effectiveness of remedial interventions. This study proposes a predictive analytics framework that applies machine learning techniques to proactively identify, characterize, and prioritize learning gaps in low-resource schools, with a specific focus on supporting equity-driven educational decision-making. The framework integrates multi-source educational data, including academic performance records, attendance histories, and engagement indicators, to capture both instructional outcomes and structural conditions influencing learning. Feature engineering emphasized temporal and behavioral instability measures such as attendance entropy, assessment decay, and engagement volatility, reflecting patterns commonly associated with sustained underperformance in underserved settings. Multiple machine learning models were evaluated, including logistic regression, random forest, gradient boosting, and long short-term memory architectures. Model evaluation prioritized equity-sensitive metrics, particularly recall and F1-score, to minimize false negatives and ensure early identification of at-risk learners.

Results indicate that ensemble-based models, especially gradient boosting, achieved the highest predictive performance, with strong recall and balanced precision across learning-gap categories. Attendance-linked and engagement-driven gaps emerged as the most severe and prevalent, while foundational literacy and numeracy gaps showed comparatively lower volatility. Equity-stratified analysis revealed a clear socioeconomic gradient, with learners in low-resource contexts experiencing disproportionately higher learning-gap intensity across all dimensions. To operationalize predictive insights, the study developed a targeted intervention prioritization framework that aligned learning-gap structure with intervention type, cost, and time-to-effect. Findings suggest that low-cost stabilization strategies, such as attendance nudges and engagement coaching, yield the greatest short-term equity gains, while intensive remediation is most effective when selectively applied. Overall, the study demonstrates that machine learning-driven predictive analytics can enable proactive, scalable, and equity-focused interventions in low-resource educational systems.

Keywords: Predictive Analytics; Educational Equity; Machine Learning; Learning Gaps; Low-Resource Schools.

Title: Predictive Analytics for Educational Equity: A Machine Learning Approach to Identifying Learning Gaps in Low-Resource Schools

Author: Linda Aluso, Stella Awo Kpogli, Joy Onma Enyejo

International Journal of Recent Research in Interdisciplinary Sciences (IJRRIS)

ISSN 2350-1049

Vol. 13, Issue 1, January 2026 - March 2026

Page No: 12-26

Paper Publications

Website: www.paperpublications.org

Published Date: 27-January-2026

DOI: https://doi.org/10.5281/zenodo.18390393

Paper Download Link (Source)

https://www.paperpublications.org/upload/book/Predictive%20Analytics%20for%20Educational%20Equity-27012026-3.pdf