Ensemble and Multimodal Approaches for Analyzing Student Engagement and Flexibility in Online Learning: A Review

Abstract

The rapid growth of online learning environments has intensified the need to understand student engagement in order to enhance learning effectiveness and academic outcomes. Student engagement is typically inferred from digital traces such as login frequency, video viewing behavior, discussion participation, and assessment performance. However, another critical dimension of online learning is student flexibility, encompassing time management, self-regulation, and adaptability to learning schedules has received comparatively limited attention in existing research. This review paper systematically examines recent studies on student engagement analysis, with a particular focus on data modalities, Machine Learning (ML) and Deep Learning (DL) models, ensemble learning techniques, and multimodal learning strategies. The surveyed literature reveals extensive use of traditional ML algorithms and deep neural networks, often relying on single-source data and standalone models. While these approaches demonstrate promising predictive performance, they frequently fail to capture the complex and dynamic nature of student behavior in online settings. Recent trends indicate a growing interest in ensemble learning methods, which integrate multiple models to enhance prediction accuracy, robustness, and generalizability. Despite this progress, the review identifies significant research gaps, including the limited incorporation of flexibility-related features and the absence of comprehensive ensemble-based multimodal frameworks that jointly model engagement and flexibility. Based on the analysis, this paper argues that integrating engagement indicators with flexibility attributes through ensemble and multimodal learning approaches can provide a more holistic and reliable understanding of student behavior. Such frameworks have the potential to support early intervention, personalized learning, and improved decision-making in online education systems.