Exploring Machine Learning Approaches for Predicting Student Performance in E-Learning: A Systematic Literature Review

In the learning environment, business intelligence through data mining techniques changed the whole sphere of education with new approaches and key findings that emerged from the implementation of machine learning. The objective of this systematic literature review is to discuss new developments in this field that enable gaining an insight into the dynamics of student learning environments with the help of analysis based on data. Educational institutions operate in a highly competitive and complex setup and face issues ranging from performance appraisal to the provision of quality education and assessment of student progress. Core to these challenges is the development of student intervention strategies that can effectively support diverse learning journeys. One of the key points synthesised from the review of studies in which the ML algorithms were applied to predict student outcomes, most of whom had dropout risks, was that, drawing on a good number of diverse datasets sources, institutional student databases, and online learning platforms, the researchers established the efficacy of ML methods in spotting students at elevated risk of dropping out, enhancing student performance overall. It consolidates existing knowledge on the contribution of ML in shaping proactive educational strategies while emancipating paths to future endeavours in the optimization of student success in e-learning environments.