Table 3. Average metrics of six models trained with stepwise backward elimination.

Model	AUC	Accuracy	Precision	Sensitivity	Specificity	F1-score
ANN	0.813	0.706	0.961	0.697	0.780	0.808
Random forest	0.772	0.686	0.958	0.676	0.769	0.792
AdaBoost	0.762	0.673	0.956	0.662	0.759	0.782
Stochastic gradient boosting	0.803	0.707	0.956	0.701	0.748	0.809
XGBoost	0.808	0.696	0.958	0.688	0.764	0.801
SVM	0.760	0.771	0.969	0.774	0.739	0.860

The highest value was bolded.

AUC: area under the receiver operating characteristic curve; ANN: artificial neural network; SVM: support vector machine