Table 7.

Top results (with threshold accuracy set at 80%) of numerical experiments for the STFT approach with hyperparameter optimization.

Model	Transformation	n_train	n_test	% for Training	Acc. [%]
Gradient Boosting	NONE	30	45	60	88.89
Random Forest	STD	30	45	60	88.89
Random Forest	MINMAX	52	23	30	86.96
Random Forest	NONE	52	23	30	86.96
Gradient Boosting	MINMAX	37	38	50	86.84
Gradient Boosting	MINMAX	30	45	60	86.67
SVC	STD	22	53	70	84.91
SVC	MINMAX	22	53	70	84.91
SVC	NONE	22	53	70	84.91
SVC	MINMAX	30	45	60	84.44
SVC	STD	30	45	60	84.44
SVC	NONE	30	45	60	84.44
Gradient Boosting	STD	22	53	70	83.02
XGB	MINMAX	30	45	60	82.22
XGB	STD	30	45	60	82.22
XGB	NONE	30	45	60	82.22
Gradient Boosting	NONE	37	38	50	81.58
Random Forest	MINMAX	37	38	50	81.58
SGD	STD	37	38	50	81.58
GaussianNB	STD	22	53	70	81.13
GaussianNB	MINMAX	22	53	70	81.13
GaussianNB	NONE	22	53	70	81.13
XGB	STD	22	53	70	81.13
XGB	MINMAX	22	53	70	81.13
XGB	NONE	22	53	70	81.13
Decision Tree	MINMAX	45	30	40	80.00
LGBM	MINMAX	30	45	60	80.00
LGBM	STD	30	45	60	80.00
LGBM	NONE	30	45	60	80.00
Random Forest	NONE	30	45	60	80.00
SGD	STD	30	45	60	80.00
SVC	MINMAX	45	30	40	80.00
SVC	STD	45	30	40	80.00
Decision Tree	NONE	45	30	40	80.00
SVC	NONE	45	30	40	80.00
Random Forest	MINMAX	60	15	20	80.00