Table 2.

Diagnostic Performance of Each Model in the Training and Test Sets

Classifiers	Brier Loss	Log Loss	Acc.	Recall	F1	AUC	95% CI	Sen.	Spe.	N.p.v.	P.p.v.
Training set
RF	0.025	0.138	0.867	1.000	1.000	1.000	1.000–1.000	1.000	1.000	1.000	1.000
XGB	0.036	0.181	0.891	1.000	1.000	1.000	1.000–1.000	1.000	1.000	1.000	1.000
LGB	0.115	0.389	0.889	0.792	0.844	0.950	0.772–0.984	0.792	0.949	0.881	0.905
LDA	0.153	0.599	0.810	0.667	0.727	0.910	0.769–0.981	0.667	0.897	0.814	0.800
LR	0.162	0.506	0.794	0.500	0.649	0.920	0.733–0.949	0.500	0.974	0.760	0.923
SVC	0.117	0.367	0.794	0.542	0.667	0.930	0.736–0.952	0.542	0.949	0.771	0.867
KNN	0.117	0.354	0.841	0.708	0.773	0.910	0.71–0.961	0.708	0.923	0.837	0.850
NB	0.157	0.748	0.810	0.625	0.714	0.930	0.764–0.977	0.625	0.923	0.800	0.833
DT	0.046	0.143	0.937	0.958	0.920	1.00	0.875–0.988	0.958	0.923	0.973	0.885
Testing set
RF	0.170	0.514	0.786	0.636	0.700	0.800	0.506–0.867	0.636	0.882	0.789	0.778
XGB	0.170	0.507	0.714	0.636	0.636	0.840	0.521–0.866	0.636	0.765	0.765	0.636
LGB	0.174	0.523	0.679	0.545	0.571	0.820	0.478–0.839	0.545	0.765	0.722	0.600
LDA	0.231	0.945	0.750	0.636	0.667	0.830	0.546–0.894	0.636	0.824	0.778	0.700
LR	0.161	0.492	0.786	0.455	0.625	0.830	0.571–0.875	0.455	1.000	0.739	1.000
SVC	0.169	0.485	0.786	0.636	0.700	0.800	0.588–0.918	0.636	0.882	0.789	0.778
KNN	0.133	0.399	0.821	0.727	0.762	0.870	0.739–0.95	0.727	0.882	0.833	0.800
NB	0.224	1.170	0.750	0.636	0.667	0.880	0.551–0.889	0.636	0.824	0.778	0.700
DT	0.500	17.270	0.500	0.545	0.462	0.570	0.478–0.854	0.545	0.471	0.615	0.400

Abbreviations: RF, Random Forest; XGB, Extreme Gradient Boosting; LGB, LightGBM; LDA, Linear Discriminant Analysis; LR, Logistic Regression; SVC, Support Vector Classifier; KNN, k-Nearest Neighbors; NB, Naive Bayes; DT, Decision Tree; Acc., accuracy; AUC, area under the curve; CI, confidence interval; Sen., sensitivity; Spe., specificity; N.p.v., negative predictive value; P.p.v, positive predictive value.