The KNN algorithm with feature selection achieved the highest accuracy of 74.44% and an Area Under the Curve (AUC) of 0.8212. This model showed a balanced accuracy improvement compared to its performance using the dataset with complete features, which had an accuracy of 72.83% and an AUC of 0.8071. Similarly, the Random Forest model with feature selection showed an accuracy of 72.00% and an AUC of 0.7741, compared to an accuracy of 70.52% and an AUC of 0.7672 with all features. The SVM model with feature selection also improved, reaching an accuracy of 72.28% and an AUC of 0.7812, compared to an accuracy of 69.80% and an AUC of 0.774 with all features. Logistic Regression showed minimal change, with an accuracy of 69.14% and an AUC of 0.7644 after feature selection, compared to an accuracy of 69.25% and an AUC of 0.7645 with all features.