Table 1.

Results of ML models for COVID-19 (95% confidence interval (CI)).

Score/Model	XGBoost	LR	RF	SVM
Accuracy	0.93	0.912	0.912	0.877
	(0.864–0.996)	(0.839–0.986)	(0.839–0.986)	(0.792–0.962)
Specificity	0.926	0.893	0.923	0.833
	(0.757–0.991)	(0.718–0.977)	(0.749–0.991)	(0.653–0.944)
Sensitivity	0.933	0.931	0.903	0.926
	(0.779–0.992)	(0.772–0.992)	(0.742–0.98)	(0.757–0.991)
F1-score	0.933	0.915	0.918	0.877
	(0.869–0.998)	(0.843–0.988)	(0.847–0.989)	(0.792–0.962)
Negative predictive value	0.926	0.926	0.889	0.926
	(0.757–0.991)	(0.757–0.991)	(0.708–0.976)	(0.757–0.991)
Positive predictive value	0.933	0.9	0.933	0.833
	(0.779–0.992)	(0.735–0.979)	(0.779–0.992)	(0.653–0.944)