TABLE 2.

AUCs for the performance of the microbiome-based predictive models.

Cohort	Classifiers	AUC	Accuracy	Sensitivity	Specificity	PPV	NPV	F1 score
Training	LR	0.915 (0.87–0.961)	0.853	0.784	0.891	0.800	0.882	0.792
	RF	0.944 (0.910–0.979)	0.881	0.863	0.891	0.815	0.921	0.838
	XGBoost	0.964 (0.937–0.990)	0.916	0.824	0.967	0.933	0.908	0.875
	SVM	0.868 (0.810–0.926)	0.804	0.765	0.826	0.709	0.864	0.736
	MLP	0.960 (0.931–0.990)	0.895	0.961	0.859	0.790	0.975	0.867
	KNN	0.909 (0.931–0.990)	0.860	0.882	0.848	0.763	0.929	0.818
	DT	0.776 (0.700–0.852)	0.769	0.725	0.793	0.661	0.839	0.692
	LightGBM	0.994 (0.987–1.000)	0.965	0.980	0.957	0.926	0.989	0.952
Test	LR	0.742 (0.618–0.865)	0.677	0.56	0.757	0.609	0.718	0.583
	RF	0.706 (0.559–0.854)	0.710	0.640	0.757	0.640	0.757	0.640
	XGBoost	0.756 (0.626–0.885)	0.742	0.600	0.838	0.714	0.756	0.652
	SVM	0.815 (0.710–0.920)	0.710	0.720	0.703	0.621	0.788	0.667
	MLP	0.785 (0.669–0.901)	0.694	0.760	0.649	0.594	0.800	0.667
	KNN	0.744 (0.619–0.868)	0.645	0.640	0.649	0.552	0.727	0.593
	DT	0.549 (0.417–0.680)	0.548	0.400	0.649	0.435	0.615	0.417
	LightGBM	0.749 (0.625–0.873)	0.645	0.640	0.649	0.552	0.727	0.593

SVM: support vector machine, LR: logistic regression, RF: random forest, KNN: k-nearest neighbor, AUC: area under the curve, NPV: negative predictive value, PPV: positive predictive value.