Table 2.

Performance metrics of the prediction models validated on the NHANES cohort.

Model	Accuracy	Precision	Recall	F1-score	AUC (95% CI)	DeLong’s test P
LR	0.743	0.326	0.602	0.423	0.767 (0.758, 0.777)	<0.001
LR (+enzymes)	0.762 (+1.9%)	0.340 (+1.4%)	0.559 (−4.3%)	0.423 (0%)	0.775 (0.765, 0.784)
CART	0.562	0.244	0.859	0.380	0.709 (0.699, 0.719)	0.002
CART (+enzymes)	0.554 (−0.8%)	0.243 (−0.1%)	0.876 (+1.7%)	0.381 (+0.1%)	0.698 (0.688, 0.708)
RF	0.631	0.279	0.863	0.422	0.794 (0.785, 0.803)	<0.001
RF (+enzymes)	0.667 (+3.6%)	0.300 (+2.1%)	0.849 (−1.4%)	0.443 (+2.1%)	0.810 (0.801, 0.818)
GBDT	0.709	0.316	0.737	0.442	0.792 (0.783, 0.800)	<0.001
GBDT (+enzymes)	0.717 (+0.8%)	0.328 (+1.2%)	0.773 (+3.6%)	0.461 (+1.9%)	0.808 (0.799, 0.816)
SVM	0.641	0.237	0.584	0.337	0.651 (0.640, 0.662)	0.090
SVM (+enzymes)	0.649 (+0.8%)	0.242 (+0.5%)	0.586 (+0.2%)	0.343 (+0.6%)	0.660 (0.648, 0.673)
KNN	0.619	0.270	0.840	0.408	0.771 (0.763, 0.780)	0.030
KNN (+enzymes)	0.607 (−1.2%)	0.265 (−0.5%)	0.853 (+1.3%)	0.404 (−0.4%)	0.778 (0.769, 0.788)
LightGBM	0.678	0.300	0.799	0.437	0.798 (0.790, 0.806)	<0.001
LightGBM (+enzymes)	0.745 (+6.7%)	0.344 (+4.4%)	0.698 (−10.1%)	0.461 (+2.4%)	0.810 (0.801, 0.818)
XGBoost	0.696	0.314	0.797	0.450	0.803 (0.794, 0.811)	<0.001
XGBoost (+enzymes)	0.743 (+4.7%)	0.346 (+3.2%)	0.723 (−7.4%)	0.468 (+1.8%)	0.814 (0.805, 0.822)