Virtual Exhibitions as Learning Media: A Study on Learning Path Mining and Instructional Intervention Strategies Based on VR/AR Interaction Logs

Abstract

Virtual exhibitions have emerged as highly promising informal learning environments, offering unprecedented opportunities for immersive and interactive education. Yet, understanding precisely how learners interact with and navigate through these complex immersive spaces remains significantly underexplored in current literature. This study systematically investigates the application of advanced learning path mining techniques, based on comprehensive virtual reality (VR) and augmented reality (AR) interaction logs, to uncover underlying behavioral patterns in virtual exhibition settings and subsequently develop targeted instructional intervention strategies. A rigorous quantitative analytical approach was employed, utilizing extensive, publicly available interaction log datasets derived from existing virtual museum deployments. Sophisticated process mining and sequence pattern mining techniques were meticulously applied to reconstruct, visualize, and analyze learners' intricate navigation trajectories. The empirical findings reveal distinct learning path archetypes that strongly correlate with varying levels of learning engagement, cognitive load, and ultimate knowledge acquisition outcomes. Based on these newly identified behavioral patterns, a comprehensive set of adaptive instructional intervention strategies was developed to dynamically personalize and optimize the virtual exhibition experience for diverse learner profiles. Ultimately, this study significantly contributes to the growing integration of learning analytics with immersive learning environments. It offers highly practical, data-driven insights for the future design and implementation of adaptive virtual exhibitions that can respond dynamically and intelligently to real-time learner behaviors, thereby maximizing educational efficacy.