ANALYZING HYBRID MACHINE LEARNING APPROACHES FOR PREDICTING STUDENT PERFORMANCE IN HIGHER EDUCATION

Abstract

Higher education has undergone a transformation in recent years due to the incorporation of blended learning models, which combine traditional classroom approaches with online educational resources. This study analyzes and predicts student performance in higher education through the implementation of a hybrid machine-learning approach. Since student learning incorporate various educational and behavioral patterns which needs to be analyzed through appropriate method. First, it implements cluster analysis, infer cluster labels, and apply hybrid classifiers. Through the use of several machine learning techniques, such as ensemble methods, this study hopes to improve performance forecasts over time relative to single algorithms. There are several features were used that includes academic records, online activity logs, discussion forum posts and views. The findings present a comparative analysis of various hybrid techniques and demonstrate that combining models enables more efficient analysis of the different factors influencing student performance.