e-LION: Data integration semantic model to enhance predictive analytics in e-Learning.

Abstract

The surge in online education emphasizes Learning Management Systems' (LMSs) crucial role in organizing learning resources and enabling teacher-learner communication. COVID-19 accelerated this, spiking engagement and substantial learning data. Academic institutions now have extensive data for comprehensive analysis to inform educational planning. However, integrating this diverse, sizable dataset from heterogeneous sources with semantic inconsistencies is challenging. Standardized integration schemes are needed for efficient utilization in machine learning models. Semantic web technologies offer a promising framework for semantic integration of e-learning data, enabling systematic consolidation, linkage, and advanced querying.

We propose the e-LION (e-Learning Integration ONtology) semantic model to consolidate diverse e-learning knowledge bases and enhance analytical capabilities. Populated with real-world data from various LMSs, focusing on Software Engineering courses from the University of Malaga (Spain) and the Open University Learning, we validate it through four in-depth case studies. Advanced semantic querying techniques feed predictive models, perform time-series forecasting of student interactions based on final grades, and develop SWRL reasoning rules for student behavior classification.

Validation study results are highly promising, suggesting e-LION as an ontological mediator scheme for integrating future semantic models within the e-learning domain. This opens exciting possibilities for leveraging the e-LION model to enhance educational planning, predictive modeling, and behavioral analysis, ultimately advancing e-learning through effective semantic integration and diverse learning-related data utilization.