This paper investigates the use of deep learning ensemble strategies in outcome-based education (OBE) and examines how ideology and politics influence vocational training outcomes. The study evaluates three ensemble strategies, namely bagging, boosting, and stacking, for predicting vocational outcomes and explores how political and institutional contexts relate to training effectiveness. The dataset comprises vocational training results together with political-context information linked to vocational programs. Political ideology was quantified using policy and governance indicators and encoded as contextual features aligned with vocational training records. A deep learning ensemble framework was developed to predict outcomes such as skill acquisition, employment rates, and overall learner success. The results show that ensemble learning improves predictive performance, with the best-performing model achieving an accuracy of 98.45% in predicting student performance from training-related parameters. Model robustness was assessed using 10-fold cross-validation, and performance is reported as the cross-validated mean with confidence intervals. The findings further indicate that political and ideological factors are associated with variations in vocational outcomes, highlighting the importance of policy context in program design and alignment with labor market demands. Overall, the integration of ensemble learning with vocational education systems offers a promising approach for improving outcome prediction and supporting more effective and inclusive vocational training initiatives.
Alamoudi et al. (Tue,) studied this question.