Ontology-Driven Dynamic Learner Profiling: Towards Fluid Learning
Bousalem, Samia Chelghoum, Massinissa Benchikha, Fouzia
Bousalem, Samia
Chelghoum, Massinissa
Benchikha, Fouzia
Type
Supervisor
Subject
Date
2026-02-26
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
Recent advances in adaptive learning have empha-sized the importance of learner profiles that are not only accurate but also capable of evolving over time. However, most existing profiling approaches rely on static representations that fail to capture the dynamic nature of learners’ preferences and performances. This limitation reduces the ability of adaptive systems to provide truly individualized and responsive learning experiences. In this paper, we propose a dynamic learner profiling approach grounded in an enriched ontology. Rather than introducing new features, our method reuses two existing dimensions—Evaluation and Learning Style (based on the VAK model)—to enable conditional and continuous updates of learner profiles. Evaluation results serve as semantic triggers for re-predicting the learner’s VAK distribution, which is represented as percentages to reflect fluid and overlapping tendencies. SWRL rules formalize the triggering conditions, while a machine learning model recomputes the learning style distribution when necessary. By integrating semantic modeling with adaptive predictions, the proposed framework moves beyond static learner representations and lays the foundation for Fluid Learning, where learner profiles evolve alongside their progress and setbacks.
Department
Publisher
Sponsor
na