Machine Learning in Higher Education: Predicting and Mitigating Student Dropout

The purpose of this study is to address the challenge of student attrition in higher education institutions by developing an early warning system (EWS) using machine learning algorithms. A combination of Random Forest (RF), Decision Tree (DT), and Support Vector Machine (SVM) models was employed to predict student dropout rates. A dataset of 12,220 students, including enrollment information, financial standing, and academic indicators, was utilized. Data preprocessing involved handling missing values and feature engineering. The Random Forest model achieved the highest accuracy at 93%, followed by the Decision Tree at 91% and the Support Vector Machine at 89%. Feature importance analysis identified key predictors of attrition, such as Current Earned Hours, Overall CGPA, and financial standing. The findings suggest that targeted interventions and support strategies can improve student retention. The study highlights the potential of machine learning techniques in predicting student attrition and providing actionable insights for effective interventions. The cross-validation process ensured the reliability and generalizability of the models, contributing valuable knowledge to the field of educational data analytics.