Student Academic Performance Prediction in Higher Education By Means of Machine Learning: A Review

Abstract

Predicting student performance is one of the most significant themes in learning contexts because it aids in the creation of successful mechanisms to improve academic results and prevent dropout. Instructors can more correctly distribute resources and teaching using effective performance prediction tools. As a result, identifying the factors affecting students' performance in higher education, particularly through the use of predictive data mining tools, is still scarce. Hence, the primary goal of this research is to identify the most widely investigated factors that influence student performance, as well as the most commonly used data mining tools to find these characteristics. Although there are various EDM surveys in the literature, we were only able to discover a few that focused on student performance analysis and prediction. In this area, researchers are looking for features that may be used to create predictions, algorithms that can improve forecasts, and ways to quantify aspects of student performance. Furthermore, research into forecasting student performance aims to find interconnected features as well as the underlying reasons why some aspects work better than others.