Table 3.

Performance summary of machine learning models in Non-CMP-IV group.

Model	Accuracy (95% CI)	Precision (95% CI)	Recall (95% CI)	F1 (95% CI)	AUC (95% CI)
LGBM	0.847 (0.830, 0.871)	0.857 (0.848, 0.877)	0.894 (0.889, 0.899)	0.895 (0.885, 0.908)	0.738 (0.699, 0.777)
XGBoost	0.832 (0.800, 0.861)	0.857 (0.843, 0.872)	0.888 (0.878, 0.898)	0.886 (0.867, 0.902)	0.732 (0.694, 0.770)
CatBoost	0.851 (0.840, 0.871)	0.856 (0.849, 0.870)	0.894 (0.888, 0.901)	0.897 (0.891, 0.908)	0.724 (0.668, 0.789)
RF	0.852 (0.843, 0.861)	0.853 (0.850, 0.861)	0.882 (0.866, 0.898)	0.898 (0.892, 0.903)	0.716 (0.656, 0.774)
GBDT	0.851 (0.843, 0.861)	0.853 (0.851, 0.858)	0.871 (0.846, 0.891)	0.898 (0.893, 0.903)	0.696 (0.647, 0.746)
Bagging	0.817 (0.779, 0.818)	0.859 (0.840, 0.865)	0.844 (0.815, 0.873)	0.876 (0.844, 0.883)	0.688 (0.622, 0.754)
SVM	0.843 (0.812, 0.861)	0.852 (0.845, 0.861)	0.868 (0.854, 0.882)	0.863 (0.844, 0.873)	0.672 (0.649, 0.740)
AdaBoost	0.807 (0.761, 0.830)	0.862 (0.838, 0.846)	0.855 (0.848, 0.862)	0.868 (0.842, 0.883)	0.682 (0.624, 0.742)
MLP	0.817 (0.796, 0.836)	0.849 (0.840, 0.858)	0.825 (0.766, 0.884)	0.874 (0.836, 0.893)	0.668 (0.634, 0.702)
LR	0.820 (0.800, 0.830)	0.851 (0.843, 0.864)	0.855 (0.839, 0.863)	0.878 (0.867, 0.884)	0.646 (0.608, 0.684)

LGBM, light gradient boosting machine; XGBoost, extreme gradient boosting; CatBoost, category boosting; RF, random forest; GBDT, Gradient boosting decision tree; Bagging, bootstrap aggregation; LR, logistic regression; SVM, support vector machine; AdaBoost, adaptive boosting; MLP, multi-layer perceptron; AUC, the area under the receiver operating characteristic curve.