Predicting Student Retention in Higher Education Using Data Mining Techniques: A Literature Review

Abstract

Many students drop out of higher education (HE) within the first year of enrollment, directly impacting academic achievement, student satisfaction, and the overall quality of the educational experience. Thus, improving student retention has become a top research priority in many HE institutions. This study aims to clarify the knowledge gap between the most effective data mining technique and the most influential factors that affect student retention. Several methods have been used to predict student retention, such as data analytics and machine learning (ML) techniques. The study reviewed and evaluated 10 current publications to increase student retention in HE. Moreover, the study examines the factors influencing student retention and the ML methods used. Additionally, the study discussed algorithms proven to work well and the challenges associated with using them. The study shows that the decision tree (DT), random forest (RF), logistic regression (LR), neural network (NN), AdaBoost, XGBoost, and long short-term memory (LSTM) algorithms performed the best out of the investigated algorithms when predicting student retention, with performance rates over 78.1%. The study concludes by discussing the limitations of this study and highlighting suggestions for future research directions that can be applied to increase student retention in HE.