Table 6.

Test result comparison with ensemble approaches.

Model	TNa	FPb	FNc	TPd	Sene	Spef	Accg	BAh	AUCi
5-layer deep neural network (DNN) (proposed)	967	103	5	46	0.9020	0.9037	0.9037	0.9028	0.9617
DNN + linear regression (LR)	976	94	6	45	0.8824	0.9121	0.9108	0.8973	0.9589
DNN + random forest (RF)	967	103	5	46	0.9020	0.9037	0.9037	0.9028	0.9572
DNN + AdaBoost	965	105	5	46	0.9020	0.9019	0.9019	0.9019	0.9607
DNN + eXtreme Gradient Boosting (XGBoost)	963	107	6	45	0.8824	0.9000	0.8992	0.8912	0.9490
DNN + support vector machine (SVM)	962	108	5	46	0.9020	0.8991	0.8992	0.9005	0.9563
RF + AdaBoost	954	116	5	46	0.9020	0.8916	0.8921	0.8968	0.9515
DNN + RF + AdaBoost	967	103	5	46	0.9020	0.9037	0.9037	0.9028	0.9579
DNN + RF + SVM	962	108	5	46	0.9020	0.8991	0.8992	0.9005	0.9556
DNN + RF + LR	963	107	5	46	0.9020	0.9000	0.9001	0.9010	0.9585
DNN + RF + AdaBoost + XGBoost	944	126	5	46	0.9020	0.8822	0.8831	0.8921	0.9571
DNN + RF + AdaBoost + SVM	959	111	5	46	0.9020	0.8963	0.8965	0.8991	0.9562
DNN + RF + AdaBoost + XGBoost + SVM	978	92	6	45	0.8824	0.9140	0.9126	0.8982	0.9572

^aTN: true negative.

^bFP: false positive.

^cFN: false negative.

^dTP: true positive.

^eSen: sensitivity.

^fSpe: specificity.

^gAcc: accuracy.

^hBA: balanced accuracy.

ⁱAUC: area under the curve.