Exploring the Use of Video Analytics to Support Adaptive Learning and Student Engagement in Thailand

Abstract

This paper presents a video-based learning system enhanced with learning analytics to monitor and analyze Thai learner behavior during video consumption. The system logs detailed interactions, including pausing, skipping, and rewinding, along with their timestamps. These interactions are quantified by frequency and duration to derive meaningful behavioral insights. An adaptive mechanism is integrated to provide personalized prompts, encouraging students to seek instructor feedback or alternative learning strategies when rapid progression through content is detected. Experimental results reveal distinct behavioral patterns of Thai students including linear viewers, pausers, skippers, rewinders, and neutral viewers. From the results, we found that 34, 24, 19, 13, and 10 percentage of participants are linear viewers, pausers, skippers, rewinders, and neutral viewers, respectively. Among these, rewinders achieved the highest average post-test score of 4.38, underscoring the benefit of content review. Frequent pausers also performed well, suggesting that reflective engagement enhances understanding. In contrast, students who skipped over half the video content scored the lowest, with an average of 1.91. These findings highlight the potential of learning analytics to support adaptive and effective video-based education for Thai students.