<![CDATA[Ensemble learning methods, which combine multiple models, have shown superior performance in various prediction tasks by leveraging the strengths of different algorithms. This study presents an application of a stacking ensemble machine learning method to predict the success of applicants in the Kominfo Scholarship program. By utilizing historical administrative data of scholarship applicants, we build a predictive model to identify candidates with a high potential to be selected and successfully complete the sponsored graduate studies. The proposed approach combines multiple base learners in an ensemble, addressing class imbalance with SMOTE oversampling and optimizing model parameters via grid search. The best-performing stacked model (combining Random Forest and XGBoost with a logistic regression meta-learner) achieved an Area Under the ROC Curve (AUC) of 0.93, outperforming individual classifiers. This paper details the data preparation, model building, and evaluation process, and discusses the implications for fair and efficient scholarship selection. The findings demonstrate that the stacking ensemble approach can enhance accuracy and objectivity in candidate selection, ensuring that deserving applicants are identified more reliably compared to conventional methods.]]>
Copyrights © 2025
