A Competence-Based Architecture for Personalized Adaptive Learning

Abstract

Traditional e-learning systems often provide limited personalization and adaptation to individual student needs. This paper presents a competence-based architecture for personalized adaptive learning in higher education that integrates an ontology-based content model, practical-skill verification, a learning-analytics-based student model, and explainable adaptive recommendations. The novelty lies in integrating these components into a single assessment-driven architecture in which a competence-and-depth content ontology, a mastery-and-attainment student model, and an ontology-grounded explainable recommendation layer operate together on real course data. The architecture was evaluated as a proof of concept using records from two undergraduate computing courses representing different computing subdomains. Student activity and assessment data were used to estimate mastery, distinguish demonstrated competence mastery from depth-level attainment, and generate prerequisite-aware recommendations from ontology relations. The results show that the proposed approach provides interpretable mastery estimates and ontology-grounded recommendations, offering initial evidence for the feasibility of competence-based adaptive learning in real course settings. A student feedback survey (n = 86) indicated positive perceptions of the clarity, relevance, and usefulness of the generated recommendations. The main contribution is a working, interpretable pipeline that transforms verified assessment evidence into mastery estimates, behavioral student types, and explainable learning recommendations.