A Survey of Learning Style Detection Method using Eye-Tracking and Machine Learning in Multimedia Learning

Abstract

Current utilization of multimedia learning environment focuses on student-centered approach. This approach is based on a theory stating that learning styles affect individuals in information processing. Based on prior works, there are three main approaches to distinguish learning styles: conventional approach-such as interview and self-reporting, artificial-intelligence-based approach, and sensor-based approach. Unfortunately, there is no comparative analysis that addresses strengths and limitations of these approaches. Thus, there is no information on how and when to use these approaches appropriately. To address this limitation, we present a brief literature review of several studies in distinguishing learning styles, including their strengths and limitations. We also present insights on potential methods of detecting learning styles in multimedia learning based on eye movement data and machine learning algorithms. Our paper is useful as a guideline for developing intelligent e-learning systems based on eye tracking and machine learning.