PRISMA-Guided Systematic Review on Machine Learning for University Student Dropout Prediction

Abstract

This systematic review examines the application of machine learning techniques to predict students dropout.
The prisma 2020 guidelines were followed to ensure a comprehensive and transparent review process. As the behaviour of
students who drop out becomes increasingly complex due to factors such as academic performance, personal characteristics
and socio-economic conditions, machine learning offers promising solutions for the early identification of students at risk.
This review summarises findings from peer-reviewed studies published between 2014 and 2024 and indexed in the scopus
database. The focus is on the performance, strengths and limitations of different machine learning models such as decision
trees, support vector machines and neural networks. The selection of the 2014-2024 timeframe reflects the significant
advances in machine learning technologies, the improved quality and availability of educational data, and the evolving
research trends in education. This timeframe also coincides with changes in education policy and ensures that the study
captures current and relevant findings. The report concludes with recommendations for future research, including the
integration of complex data characteristics and the development of universal models that can be adapted to different student
populations.