Table 6.

ICU case: results of the statistical test using the 20-top relevant features.

Best classifier	Other classifiers	Meandif	lower	upper	Reject
DeepNetwork	Knn	− 0.0309	− 0.0799	0.0181	False
DeepNetwork	Linear SVM	− 0.1272	− 0.1762	− 0.0783	True
DeepNetwork	LogisticReg	− 0.1206	− 0.1696	− 0.0716	True
DeepNetwork	MLP	− 0.092	− 0.141	− 0.043	True
DeepNetwork	Multinomial Log Reg	− 0.0937	− 0.1427	− 0.0447	True
DeepNetwork	Nonlinear SVM	− 0.1272	− 0.1762	− 0.0783	True
DeepNetwork	Random Forest	− 0.1006	− 0.1495	− 0.0516	True
DeepNetwork	AdaBoost	0.0409	− 0.0081	0.0899	False
DeepNetwork	Bagging	0.1206	0.0716	0.1696	True