Learning Analytics and Predictive Modelling in E-Learning Platforms: A Review for Early Identification of At-Risk Students

Abstract

The rapid expansion of digital education platforms and e-learning environments has transformed modern education by providing flexible, scalable, and accessible learning opportunities for students worldwide. However, a major challenge in online learning systems is the high rate of student dropout and academic failure. Unlike traditional classrooms, where instructors can closely monitor student progress and provide immediate support, online environments often lack direct interaction, which may lead to low engagement, poor motivation, ineffective time management, and limited understanding of course materials. To address these challenges, learning analytics and predictive modelling have emerged as powerful tools for identifying students who are at risk of poor academic performance or dropping out. Learning analytics involves the collection and analysis of data generated through student interactions with digital learning systems, such as login frequency, participation in discussion forums, time spent on course materials, assessment results, and assignment submission patterns. Predictive modelling applies statistical and machine learning techniques—including logistic regression, decision trees, random forests, and neural networks—to analyse these data and forecast potential learning outcomes. By identifying at-risk students early, educational institutions can implement targeted interventions and personalized support strategies. Although predictive analytics can significantly improve student retention and engagement, challenges such as data privacy, algorithm bias, and model interpretability must be carefully addressed for effective implementation.