Fig. 6

(a) Comparison of K-fold and LOO Cross-Validation Results for Classification Algorithms. The chart illustrates the performance of various classification algorithms, including Random Forest, KNN, Decision Tree, SVM, Logistic Regression, Naive Bayes, XGBoost, and LightGBM. Results are displayed for two validation methods: (a) K-fold and (b) LOO. The accuracy varies across methods, with Random Forest and KNN performing best in K-fold, while ensemble models like XGBoost excel in LOO. LightGBM shows significant improvement in LOO compared to K-fold, where it performed the weakest. (b) Accuracy of Random Forest Classification Using K-Fold and LOO Validation. This chart illustrates the classification accuracy of the Random Forest algorithm based on various parameter combinations, evaluated using K-Fold and LOO methods. The combination of CT, Q, and DI yields the highest accuracy of 76.10% in K-Fold validation, showcasing the effectiveness of these features in enhancing classification performance.