<![CDATA[Predicting student graduation is one of the challenges in the field of education that requires a data-driven approach. Not only final grades play a role, but also other factors such as attendance rate, weekly study hours, previous exam scores, and extracurricular activities. This study compares the performance of three classification algorithms—Decision Tree, Random Forest, and k-Nearest Neighbor (k-NN)—in predicting student graduation status based on the Student Performance dataset from Kaggle, which contains 708 student records. The modeling process was conducted using Orange Data Mining through a visual workflow approach. The models were evaluated using 20-fold cross-validation and assessed with performance metrics including AUC, accuracy, precision, recall, F1-score, and MCC. The results show that the Random Forest algorithm achieved the best performance, with an AUC of 97.1%, accuracy of 94.1%, F1-score of 94.2%, precision of 94.2%, recall of 94.1%, and MCC of 79.7%. While Decision Tree and k-NN also performed well, their results were still below those of Random Forest. These findings indicate that Random Forest is the most accurate and stable model for classifying student graduation and demonstrate that Orange Data Mining is an effective tool for applying data mining techniques in the educational field.]]>
Copyrights © 2025
