A Review: Predicting Student Success at Various Levels of their Learning Journey in a Science Programme

Abstract

This paper examines how features affect student persistence or dropout at South African higher education institutions, based on three previous studies. In the previous studies, high school grades were used as a valid predictor of student success. The quality of a high school's learning environment has an effect on almost every aspect of higher education success. Students who are better prepared coming out of high school are ideally suited to do well in higher education institutions, who they are, how much money they have, and where they go don't matter. This review aims to identify effective features that warrant student success from high school grades and choice of academic courses during registration in higher education. The following questions are used to guide this review: How can we define student success? Which features should we focus on? Which models work? Based on data mining techniques such as machine learning models that the previous studies have used to predict student success, it has been revealed that the most important features that influence student success in a Computer Science programme are Prior Computer experience, Mathematics, English from High school and the choice of a course.