AI-based learning style detection in adaptive learning systems: a systematic literature review

The integration of AI in education, particularly in adaptive learning, emphasizes the critical need for automatic detection of individual learning styles. Traditional methods such as tests or questionnaires, though reliable, face challenges including student reluctance and limited self-awareness of learning preferences. This underscores a research gap in learning style detection within adaptive learning systems, necessitating further investigation into the effectiveness of algorithms/models for automatic detection within AI-driven systems in real-world educational settings. Additionally, the parameters for adaptation experiments, the role of machine learning techniques, and the comparative analysis of different methodologies remain underexplored areas. Addressing these gaps, this study conducts a systematic review of articles from 2014 to 2022, using Web of Science and Scopus. Forty selected papers are rigorously evaluated to understand automatic learning style detection's current state, challenges, and future directions in diverse educational contexts. This study explores automatic learning style detection in diverse educational aspects, including techniques, approaches, models, and implementation. We find that AI techniques, especially data-driven approaches, enhance learning adaptation. The dominance of the Felder-Silverman model and the versatility of AI algorithms like Decision Trees and Artificial Neural Networks underscore their effectiveness across diverse contexts. Additionally, our analysis highlights the prevalence of Moodle in dataset mining and learning experiments, demonstrating its importance in research. Our research provides valuable insights into the design and implementation of AI-driven educational solutions, focusing on adapting course content according to learning styles. The aim is to enhance learning outcomes within educational environments.