Citation

Shannaq, Boumedyen and AlMaqbali, Said and Ali, Oualid (2026) Leakage-Aware Explainable AI Framework for ROI Optimization in Distance Learning: A Predictive Modeling and Scenario Simulation Approach. In: 2026 ASU International Conference in Emerging Technologies for Sustainability and Intelligent Systems, ICETSIS 2026, 6 May 2026 - 7 May 2026, Manama.

Text 10.pdf - Published Version Restricted to Repository staff only Download (820kB)

Abstract

The recent rapid development of artificial intelligence (AI) as a part of distance-learning systems has triggered a parallel increase in institutional funding of digital pedagogy. However, solid approaches to measuring the return-on-investment (ROI) of these systems are limited and are vulnerable to bias due to data leakage. To respond to this, this work proposes a leakage-conscious, explainable AI model that predicts and maximizes institutional ROI using behavioral and structural features derived out of vast volumes of online course data. The empirical study uses a cohort of 291 course records of HarvardX and MITx which were obtained in the Kaggle repository named Online Courses from Harvard and MIT. An array of machine-learning models was put to test, and XGBoost has achieved the most desirable baseline outcome (Test \(R^2 = 0.963\)). Nevertheless, diagnostic testing based on SHAP values indicated that there existed outcome-embedded variables, which add data-artefacts associated with leakage. A subsequent elimination of these predictors due to leakage resulted in a clean model, which produced a strong Test R 2 = 0.869 (RMSE = 0.0252), thus proving valid external validity.The explainable AI diagnostics additionally found the efficiency of certification, the quality of learning-performance, and the level of learnerengagement as the main ROI drivers. The scenariosimulation experiments showed that middle-range interventions that would facilitate engagement had a potential of increasing the average ROI by about 18.48 per cent on average. The proposed framework syncretically incorporates leakage-control schemes, explicable modeling tools, and interactive scenario simulation, thus empowering educational executives with information-based insights in order to support informed decision making in AI-enabled distance-learning frameworks

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Artificial Intelligence, Data Leakage Detection
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	04 Aug 2026 06:25
Last Modified:	04 Aug 2026 06:25
URII:	http://shdl.mmu.edu.my/id/eprint/16491

Downloads

Edit (login required)