Table 4.

Performance summary of algorisms used in the present study.

Model	Accuracy	Sensitivity	Specificity	F1 Score	AUC
SVM	0.52 (0.44–0.56)	0.52 (0.45–0.62)	0.52 (0.37–0.60)	0.49 (0.46–0.54)	0.53 (0.44–0.56)
LGBM	0.59 (0.52–0.69)	0.57 (0.41–0.79)	0.60 (0.29–0.89)	0.56 (0.52–0.60)	0.59 (0.52–0.67)
XGB	0.59 (0.48–0.69)	0.55 (0.42–0.69)	0.61 (0.40–0.86)	0.55 (0.44–0.64)	0.56 (0.49–0.68)
XGB based on RF	0.61 (0.50–0.67)	0.59 (0.46–0.66)	0.62 (0.38–0.76)	0.58 (0.51–0.62)	0.61 (0.51–0.66)
CatBoost	0.63 (0.56–0.69)	0.70 (0.52–0.86)	0.58 (0.44–0.80)	0.63 (0.56–0.74)	0.60 (0.57–0.70)
iRF	0.76 (0.70–0.80)	0.69 (0.62–0.77)	0.83 (0.65–0.94)	0.73 (0.71–0.76)	0.77 (0.73–0.83)

SVM, support vector machine; LGBM, light gradient boosting machine; XGB, extreme gradient boosting; RF, random forest; CatBoost, categorical boosting; iRF, improved random forest; AUC, area under the receiver operating characteristic curve.