“…Therefore, no conclusive empirical evidence from their experimental results showed that using SMOTE, RUS, and ROS could significantly positively impact machine learning-based CSD models. Alkharabsheh et al Alkharabsheh et al (2022) used the machine learning classifiers (i.e., LDA, Quadratic Discriminant Analysis (QDA), NB, Multi-Layer Perceptron (MLP), SVM, DT, GB, CatBoost, Light Gradient Boosting Machine (LGBM), XGBoost, XGBoost with Random Forest (XGBRF), AdaBoost, Bagging, RF, Extra Trees (ET), KNN, Nearest Centroid (NC), Gaussian Process (GP), Ridge, LR, Perceptron, Passive Aggressive (PA), and Stochastic Gradient Descent (SGD)) to compare whether using SMOTE would improve the detection performance on God Class detection. Their results showed that SMOTE could not improve the God Class detection performance.…”