Table 3.

Performance of different models using one-hot vectors on the test set of training cohort.

Model	Sensitivity (mean ± std)	Specificity (mean ± std)	PPV (mean ± std)	NPV (mean ± std)	F1-score (mean ± std)	AUC (mean ± std)
FNN	0.8023 ± 0.0270	0.7796 ± 0.0293	0.0935 ± 0.0095	0.9929 ± 0.0008	0.1676 ± 0.0149	0.8641 ± 0.0118
LightGBM	0.7652 ± 0.0160	0.8799 ± 0.0036	0.1530 ± 0.0064	0.9925 ± 0.0005	0.2550 ± 0.0097	0.9052 ± 0.0067
LR	0.7573 ± 0.0161	0.8714 ± 0.0029	0.1430 ± 0.0033	0.9922 ± 0.0005	0.2406 ± 0.0052	0.8943 ± 0.0104
RF	0.7945 ± 0.0477	0.8094 ± 0.0182	0.1057 ± 0.0053	0.9929 ± 0.0013	0.1865 ± 0.0066	0.8783 ± 0.0085
SVM	0.7867 ± 0.0079	0.8659 ± 0.0042	0.1426 ± 0.0044	0.9931 ± 0.0003	0.2414 ± 0.0066	0.9104 ± 0.0090
XGBoost	0.7867 ± 0.0142	0.8623 ± 0.0028	0.1393 ± 0.0042	0.9930 ± 0.0004	0.2367 ± 0.0067	0.9092 ± 0.0064