Adaptive M-Learning Content on the Moodle Platform Using M-Learner Modeling Approach, Ontologies, and Machine Learning Techniques

Background/purpose. Mobile learning (M-learning) has become a crucial component of higher education due to the increasing demand for flexible and adaptive learning environments. However, ensuring personalized and effective M-learning experiences remains a challenge. This study aims to enhance M-learning effectiveness by introducing an AI-driven Moodle plugin that personalizes learning experiences using a learner-modelling approach. Materials/methods. The proposed model leverages data from Ibn Tofail University's Moodle platform and integrates DBSCAN and K-means clustering techniques to classify learners based on behavioural, contextual, and emotional factors. The generated clusters enable the delivery of personalized content structured within a Moodle course ontology, optimizing engagement and learning outcomes. Results. Experimental results on the "Algorithmics and Programming in Python" course show that DBSCAN achieves better clustering accuracy than K-means, as evidenced by cohesion, separation, and silhouette scores. Additionally, an analysis of mobile device characteristics highlights the importance of screen size and RAM capacity in optimizing M-learning experiences. Conclusion. This study demonstrates the significance of learner modelling and technological adaptation in fostering effective mobile learning environments. By integrating clustering algorithms and ontologies, the proposed approach contributes to the development of intelligent, personalized M-learning applications that enhance accessibility and educational performance.