Machine learning approach for the prediction of 30-day mortality in patients with sepsis-associated encephalopathy

Table 2 Analysis of sensitivity and specificity

Model	Accuracy	Sensitivity	Specificity	PPV	NPV	AUC	Operating threshold	95% CI
NNET	0.840	0.802	0.733	0.391	0.946	0.833	0.164	(0.816, 0.849)
NB	0.833	0.767	0.800	0.450	0.941	0.816	0.058	(0.799, 0.833)
LR	0.843	0.808	0.731	0.391	0.947	0.833	0.162	(0.816, 0.848)
GBM	0.844	0.805	0.699	0.360	0.944	0.824	0.141	(0.807, 0.840)
Ada	0.846	0.786	0.737	0.390	0.942	0.834	0.148	(0.817, 0.849)
RF	0.840	0.856	0.642	0.338	0.954	0.825	0.150	(0.808, 0.841)
BT	0.836	0.715	0.745	0.375	0.925	0.804	0.240	(0.786, 0.820)
XGB	0.844	0.808	0.712	0.374	0.945	0.830	0.157	(0.814, 0.846)
CatBoost	0.842	0.789	0.741	0.394	0.943	0.830	0.165	(0.813, 0.846)

PPV Positive Predictive Values, NPV Negative Predictive Values, AUC Area Under the Curve, CI Confidence Interval, NNET artificial Neural Network, NB Naïve Bayes, LR Logistic Regression, GBM Gradient Boosting Machine, Ada Adapting boosting, RF Random Forest, BT Bagged Trees, XGB eXtreme Gradient Boosting

ISSN: 1471-2288