Purpose: This study aims to identify trends, key research areas, popular predictive features, machine learning algorithms used, and gaps in the existing literature. Design/Methodology/Approach: This study selected 60 articles published between 2018 and August 31, 2023, from Google Scholar and Scopus databases to address the identified knowledge gap. A systematic literature review and bibliographic analysis were conducted using both quantitative and qualitative approaches, with PRISMA chosen as the reporting format. Findings: The study reveals that machine learning models, particularly Decision Trees, followed by Random Forests, Artificial Neural Networks, Support Vector Machines, and Naïve Bayes, have significantly contributed to predicting student academic performance. The datasets predominantly utilised by researchers include students' academic records, demographics, activities, and behaviour, which are crucial for predicting academic performance and evaluating model effectiveness. Research Limitation: These limitations highlight the challenges in conducting comprehensive reviews of machine learning applications in educational contexts while acknowledging the evolving nature of both technology and educational assessment methods. Practical Implication: This research has the potential to inform evidence-based decision-making, promote personalised learning experiences and enable early interventions for at-risk students Social Implication: The study underscores the potential of machine learning algorithms to promote equity and inclusion, provide targeted student support, empower individuals, raise ethical considerations, foster community engagement, and support lifelong learning initiatives. Originality and Value: This research uniquely explores machine learning applications to predict student academic performance, identifying classroom participation and examination scores as key predictors. It offers a pragmatic approach to educational practice, identifies future research opportunities, and advocates for data-driven decision-making in higher education.
Read full abstract