Table 2. Summary of the training set results for the multimodel classification.

ML model	AUC (95% CI)	Cutoff (95% CI)	Accuracy (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)	F1 score (95% CI)	Kappa (95% CI)
XGBoost	0.985 (0.981–0.989)	0.294 (0.274–0.314)	0.948 (0.945–0.951)	0.912 (0.905–0.919)	0.964 (0.958–0.971)	0.926 (0.912–0.940)	0.958 (0.955–0.962)	0.919 (0.915–0.922)	0.880 (0.874–0.885)
Logistic regression	0.816 (0.798–0.834)	0.296 (0.272–0.319)	0.740 (0.731–0.750)	0.824 (0.800–0.847)	0.702 (0.679–0.726)	0.563 (0.552–0.575)	0.896 (0.885–0.907)	0.668 (0.665–0.671)	0.467 (0.458–0.476)
LightGBM	0.986 (0.981–0.990)	0.328 (0.304–0.352)	0.951 (0.949–0.953)	0.908 (0.899–0.917)	0.971 (0.965–0.976)	0.938 (0.924–0.953)	0.957 (0.953–0.961)	0.923 (0.919–0.926)	0.886 (0.881–0.891)
Random forest	0.985 (0.981–0.990)	0.340 (0.324–0.356)	0.948 (0.946–0.950)	0.918 (0.912–0.924)	0.959 (0.954–0.964)	0.926 (0.914–0.938)	0.959 (0.956–0.962)	0.922 (0.918–0.926)	0.880 (0.876–0.884)
SVM	0.823 (0.805–0.840)	0.197 (0.186–0.207)	0.710 (0.702–0.719)	0.904 (0.889–0.919)	0.621 (0.603–0.638)	0.526 (0.515–0.536)	0.933 (0.924–0.941)	0.664 (0.659–0.670)	0.439 (0.428–0.450)
KNN	0.948 (0.938–0.959)	0.400 (0.400–0.400)	0.900 (0.897–0.903)	0.871 (0.863–0.880)	0.899 (0.891–0.908)	0.882 (0.870–0.894)	0.907 (0.904–0.911)	0.876 (0.871–0.881)	0.763 (0.758–0.769)
MLP	0.849 (0.832–0.866)	0.297 (0.266–0.328)	0.748 (0.737–0.758)	0.861 (0.824–0.897)	0.696 (0.664–0.728)	0.574 (0.552–0.597)	0.915 (0.901–0.929)	0.686 (0.684–0.688)	0.490 (0.480–0.500)

ML, machine learning; AUC, area under the curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; XGBoost, extreme gradient boosting; LightGBM, light gradient-boosting machine; SVM, support vector machine; KNN, k-nearest neighbors; MLP, multilayer perceptron.