Navigating the Ensemble Landscape: A Dynamic Analysis of Predictive Classifiers and Recommendations in TVET

The incorporation of user personality traits into computing systems has catalyzed the development of new research areas, including automatic personality recognition and behavioral analysis. Personality-matched education not only enhances trainees’ course performance but also reduces dropout rates. In this study, we have evaluated Ensemble methods of voting, bagging, boosting, and stacking for Technical and Vocational Education and Training (TVET) courses predictions based on Big Five Personality Traits and Trainees demographics of Age and Gender. The results indicated that ensemble voting and stacking methods achieved a 97% accuracy rate in predicting TVET courses, with a low error rate of 0.12 for MAE and 0.01 for NMAE. When compared to existing studies, both methods surpassed previous findings. This successful prediction lays the groundwork for course recommendations, early dropout predictions, and career guidance.