Table 2.

Model performance.

Model	AUROCa (95% CI)
Ensemble learning	0.827 (0.824-0.830)
Least absolute shrinkage and selection operator	0.819 (0.816-0.822)
Extreme gradient boosting	0.808 (0.805-0.811)
Feed-forward neural network	0.807 (0.804-0.810)
Boosting	0.775 (0.771-0.778)
Random forest	0.695 (0.691-0.699)
k-Nearest neighbor	0.631 (0.627-0.635)
Naïve Bayes	0.611 (0.607-0.614)

^aAUROC: area under the receiver operating characteristic curve.