Table 3.

Model performance in predicting poor neurologic outcome after extracorporeal cardiopulmonary resuscitation.

Algorithm	Accuracy	AUC	Sensitivity	Specificity	Kappa
XGboost	0.739	0.837	0.700	0.740	0.450
RF	0.734	0.860	0.600	0.851	0.443
GBM	0.723	0.807	0.750	0.730	0.430

AUC, area under the curve; GBM, stochastic gradient boosting; RF, random forest; XGBoost, eXtreme gradient boosting.