Table 2.

AUC, accuracy (acc), sensitivity (sens), specificity (spec), NPV and PPV of LR, DT, GB, SVM and NN models on admission benchmark, last-value and time-varying models in test sets

Model	N	Admission						Last-value						Time-vary
		AUC	Acc	Sens	Spec	PPV	NPV	AUC	Acc	Sens	Spec	PPV	NPV	AUC	Acc	Sens	Spec	PPV	NPV
LR	864	0.79	0.80	0.34	0.94	0.66	0.82	0.98	0.95	0.23	0.97	0.90	0.97	0.88	0.83	0.66	0.89	0.65	0.89
DT	864	0.69	0.76	0.47	0.85	0.49	0.83	0.93	0.92	0.21	0.96	0.85	0.94	0.81	0.78	0.61	0.83	0.53	0.87
GB	864	0.83	0.82	0.53	0.91	0.64	0.86	0.99	0.96	0.24	0.97	0.90	0.98	0.93	0.88	0.81	0.90	0.72	0.94
SVM	864	0.77	0.74	0.56	0.80	0.47	0.85	0.99	0.93	0.23	0.94	0.83	0.97	0.85	0.80	0.68	0.84	0.57	0.89
NN	864	0.82	0.81	0.49	0.92	0.65	0.85	0.97	0.95	0.24	0.95	0.87	0.87	0.90	0.84	0.77	0.86	0.63	0.92

AUC, area under the receiver operating characteristic curve; DT, decision tree; GB, gradient boosting decision trees; LR, logistic regression; NN, neural network; NPV, negative predictive value; PPV, positive predictive value; SVM, support vector machine.