Learning Style Classification for Adaptive MOOCs: A Data-Driven Study of Omani Learners

Despite the global creation of MOOCs, their one-size-fits-all delivery fails to address regional learning preferences, particularly in understudied populations like Oman. This paper presents a machine learning-ready dataset and analysis framework for adaptive MOOC personalization, focusing on Omani learners' characteristics. From a survey of 211 participants, we extract and statistically validate four key learning style features namely visual, verbal, sensing and sequential using clustering and feature importance analysis. As we know, MOOCs have become a necessity in higher education, their one-size- fits-all approach leads to very low completion rates. Hence in this work, the proposed computational analysis demonstrates that preference-based adaptation particularly for visual was 76% video preference and sequential learners was 58% linear content preference could significantly improve engagement. The resulting feature set and classification schema provide foundational parameters for building AI -driven learner models in MOOCs, with direct applications in content recommendation systems and adaptive learning path. Findings from this study could overcome a critical gap in educational technology by offering data-driven information for personalized MOOC architectures in the Omani context.