Table 5.

The results of repeated 10-fold cross-validation experiments for predicting remission using multilayer feedforward neural networks (MFNNs) and logistic regression with 16 biomarkers and 6 clinical biomarkers only.

Algorithm	AUC	Sensitivity	Specificity	Number of biomarkers
MFNN with 1 hidden layer	0.8042 ± 0.0729	0.7689 ± 0.0579	0.6580 ± 0.0839	16
MFNN with 2 hidden layers	0.8047 ± 0.0727	0.7734 ± 0.0593	0.6643 ± 0.0832	16
MFNN with 3 hidden layers	0.8060 ± 0.0722	0.7732 ± 0.0583	0.6623 ± 0.0853	16
Logistic Regression	0.7985 ± 0.0772	0.7722 ± 0.0645	0.6753 ± 0.0932	16
MFNN with 1 hidden layer	0.6089 ± 0.0848	0.6698 ± 0.0073	0.3302 ± 0.0073	6
MFNN with 2 hidden layers	0.6135 ± 0.0871	0.6698 ± 0.0073	0.3302 ± 0.0073	6
MFNN with 3 hidden layers	0.6116 ± 0.0872	0.6698 ± 0.0073	0.3302 ± 0.0073	6
Logistic Regression	0.5922 ± 0.0878	0.6501 ± 0.0292	0.3330 ± 0.0290	6

AUC, the area under the receiver operating characteristic curve.

Data are presented as mean ± standard deviation.