Student Achievement Prediction Models: A PRISMA-Based Systematic Literature Review

Student achievement prediction has become an important research area in educational data mining because it supports early intervention, academic monitoring, and evidence-based decision-making in educational institutions. This study aims to identify research trends, commonly used methods, predictive variables, and potential research gaps in student achievement prediction models. A Systematic Literature Review (SLR) was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework. Articles published between 2020 and 2024 were collected from seven reputable databases, namely Scopus, ScienceDirect, IEEE Xplore, SpringerLink, IOP, Wiley, and MDPI. After applying the inclusion and exclusion criteria, 52 articles were selected for final analysis. The findings show that classification-based machine learning methods dominate this research area, with Random Forest being the most frequently used algorithm. Academic data, such as grades, GPA, and attendance, remain the most common predictive variables, while non-academic variables are still rarely explored. This study highlights the need for multi-source data integration, hybrid or ensemble modeling, and broader variable selection to improve prediction accuracy and applicability. The novelty of this study lies in its structured synthesis of recent studies and its proposed direction for developing more comprehensive student achievement prediction models.