Recommendation System for Personalized Contextual Pedagogical Resources Based on Learning Style

E-learning systems have undergone a significant transformation in the field of education, providing valuable support to learners and trainers. Among the emerging technologies, recommender systems have proven essential in supporting learners in selecting learning resources that best fit their preferences and requirements. However, learning is not limited to specific times or physical environments, making it necessary to customize resources that reflect learners’ contextual backgrounds. To address this challenge, this paper proposes an intelligent recommendation approach that offers highly personalized learning objects using contextual information and the Felder-Silverman Learning Style Model (FSLSM). In order to automatically model the learner, we analyze the learner’s interaction traces and sensor log data. By extracting relevant features, we cluster a population of 624 learners enrolled in two agronomy courses at IAV Hassan II, based on their preferred learning style (global or sequential). To achieve the objective of this study, we developed an innovative rule-based classification model. The results of this study outperform all alternative recommendation methods on both quality and accuracy, with a high precision rate of over 96%. This performance reflects the seamless incorporation of current learner contextual factors.