ML-Based Adaptive Dashboard for Personalized Feedback: A Practical Solution to Enhance Self-Regulation

This study investigates a practical solution that supports teachers, learners, and administrators in open distance learning through academic analytics. The objective is to enhance self-regulation and motivations by providing personalized and adaptive insight to learners by designing and implementing a prototype intelligent student facing learning analytics dashboard (SF-LAD) grounded in educational theory, leveraging a multi-tier computational pipeline in developing and employing an iterative, user-centered design (UCD) methodology. The system architecture integrates learning management system (LMS) data with an analytics engine utilizing a sentiment classification model (XLNet) and a predictive model (K-means clustering). The dashboard is intelligent in the sense that it embeds machine-learning models that autonomously interpret data to drive downstream analytical behavior. Research finding demonstrates that intelligent SF-LAD has decent usability. The study exemplifies the potential of intelligent frameworks and provides a foundation for long-term research into the impact of automated insights on distance learner success.