<![CDATA[Student graduation prediction is an important aspect in supporting academic decision-making in higher education. However, conventional evaluation approaches have not been able to identify the risk of early graduation delays. This study aims to compare the performance of two ensemble learning approaches, namely Bagging using Random Forest and Boosting using XGBoost, in predicting student graduation. The study used the Predict Students' Dropout and Academic Success dataset consisting of 4,424 student data. Both models were trained on the same data and evaluated using the Accuracy, Precision, Recall, F1-Score, and ROC-AUC metrics. The results of the experiment showed that both models had almost equal accuracy, i.e. 82.6% for Random Forest and 82.5% for XGBoost. However, XGBoost showed better performance on Recall (0.878) and F1-Score (0.834), which indicated a higher ability to detect students who actually graduated. Based on these results, this study concludes that XGBoost is more effective than Random Forest in the context of predicting student graduation and is more suitable to be applied to the Academic Early Warning System in universities]]>
Copyrights © 2026
