“…In particular, we evaluate a wide range of machine learning based models (i.e., the evaluating machine learning models of SVM, kNN, and RF are configured as shown in Table I) with six different normalization methods (i.e., the proposed normalization methods of MinMaxScaler [49], MaxAbsScaler [50], StandardScaler [49], RobustScaler [51], Normalizer [52], QuantileTransformer [53], and PowerTransformers [54] implemented by the Scikit-Learn toolbox [55]) and three types of features (i.e., only time-domain features, only frequency features, and both time and frequency domain features). The results of the machine learning based models are presented in Table VII.…”