Table 1.

Accuracy, sensitivity, specificity, and precision of the prediction models according to enterotypes.

Total Participants	AUROC	Accuracy	Sensitivity	Specificity	Precision	F1
XGBoost	1.0 ± 4.1 × 10−6	9.9 ± 1.7 × 10−4	0.99 ± 2.5 × 10−4	0.99 ± 2.0 × 10−4	1.0 ± 1.6 × 10−4	0.99 ± 1.52 × 10−4
Random forest	1.0 ± 0.001	1.0. ± 8.2 × 10−5	1.0 ± 0.0	0.99 ± 1.9 × 10−4	1.0 ± 1.5 × 10−4	1.0 ± 7.36 × 10−5
Linear regress	0.99 ± 2.0 × 10−4	0.97 ± 2.0 × 10−4	0.97 ± 3.4 × 10−4	0.96 ± 4.6 × 10−4	0.97 ± 4.6 × 10−4	0.97 ± 2.55 × 10−4
ET-B	AUROC	Accuracy	Sensitivity	Specificity	Precision	F1
XGBoost	1.0 ± 1.1 × 10−5	0.97 ± 3.9 × 10−4	0.96 ± 5.3 × 10−4	0.97 ± 1.3 × 10−4	0.97 ± 0.0005	0.97 ± 0.0004
Random forest	1.0 ± 3.0 × 10−5	0.97 ± 3.9 × 10−4	0.98 ± 3.9 × 10−4	0.97 ± 6.4 × 10−4	0.97 ± 0.0005	0.97 ± 0.0003
Linear regress	0.98 ± 2.8 × 10−4	0.95 ± 5.2 × 10−4	0.97 ± 5.4 × 10−4	0.93 ± 7.7 × 10−4	0.94 ± 0.0007	0.96 ± 0.0005
ET-L	AUROC	Accuracy	Sensitivity	Specificity	Precision	F1
XGBoost	1.0 ± 0.0	0.99 ± 3.4 × 10−6	0.99 ± 3.7 × 10−4	1.0 ± 0.0	1.0 ± 0.0	0.99 ± 1.9 × 10−4
Random forest	1.0 ± 0.0	0.99 ± 1.9 × 10−4	0.99 ± 4.3 × 10−4	1.0 ± 0.0	1.0 ± 0.0	0.99 ± 2.3 × 10−4
Linear regress	1.0 ± 0.0001	0.98 ± 3.9 × 10−4	0.99 ± 4.4 × 10−4	0.97 ± 6.8 × 10−4	0.98 ± 5.2 × 10−4	0.98 ± 3.2 × 10−4
ET-P	AUROC	Accuracy	Sensitivity	Specificity	Precision	F1
XGBoost	1.0 ± 0.0	0.98 ± 6.9 × 10−4	0.954 ± 0.001	1.0 ± 0.0	1.0 ± 0.0	0.98 ± 8.3 × 10−5
Random forest	1.0 ± 0.0	0.96 ± 9.2 × 10−4	0.911 ± 0.002	1.0 ± 0.0	1.0 ± 0.0	0.95 ± 0.002
Linear regress	0.95 ± 0.001	0.96 ± 0.001	0.954 ± 0.001	0.957 ± 0.001	0.954 ± 0.001	0.95 ± 0.001

ET-B, enterotype Bacteroidaceae; ET-L, enterotype Lachnospiraceae; and ET-P, enterotype Prevotellaceae. AUROC, the area under the receiver operating characteristic. Precision = TP/(TP + FP). F1 score = 2 × (precision × recall)/(precision + recall) in recall = TP/(TP + FP). TP, true positive value; FP, false positive value; FN, false negative.