Systematic Review of AI-Based Early Warning System as the Basis for Preventing Dropouts and Improving Retention of History Education Students

Abstract

The phenomenon of student dropout poses a serious challenge for higher education institutions in achieving SDG 4 on quality education. This study aims to assess the effectiveness of an Artificial Intelligence (AI)-based Early Warning System (EWS) as a basis for developing an early detection system for dropout risk and improving student retention, particularly in the History Education Study Program. The method used was a Systematic Literature Review (SLR) with the PRISMA 2020 framework, through a search on Google Scholar between 2018 and 2026, resulting in 11 final articles that met the inclusion criteria. The study results indicate that the AI-based EWS is capable of detecting students at risk of dropping out with an accuracy of 82%–96.8% and an AUC of up to 0.9615. Initial Semester Grade Point Average (GPA) was consistently the most important predictor, and accuracy increased significantly when integrated with socio-economic data. The adaptive EWS proved superior to the static system. These findings provide a scientific basis for the History Education Study Program to develop a contextual, effective, and sustainable dropout prediction model.