Table 2.

Performance comparison of four machine learning models with different sampling ratios.

Model		AUCa, mean (SD)		F1, mean (SD)		Accuracy, mean (SD)		Recall, mean (SD)		Precision, mean (SD)
Ratio=1
	LGBMb	0.75 (0.12)	0.7 (0.14)		0.72 (0.08)		0.71 (0.19)		0.73 (0.18)
	LRc	0.75 (0.1)	0.73 (0.13)		0.73 (0.1)		0.76 (0.16)		0.72 (0.15)
	RFd	0.76 (0.11)	0.72 (0.09)		0.72 (0.1)		0.73 (0.11)		0.76 (0.2)
	XGBe	0.75 (0.11)	0.71 (0.13)		0.71 (0.1)		0.73 (0.13)		0.7 (0.16)
Ratio=3
	LGBM	0.73 (0.07)	0.52 (0.08)		0.72 (0.07)		0.62 (0.16)		0.47 (0.1)
	LR	0.74 (0.07)	0.54 (0.06)		0.7 (0.09)		0.72 (0.11)		0.46 (0.11)
	RF	0.75 (0.08)	0.5 (0.08)		0.69 (0.08)		0.64 (0.14)		0.44 (0.11)
	XGB	0.76 (0.08)	0.53 (0.11)		0.71 (0.12)		0.65 (0.19)		0.48 (0.12)
Ratio=5
	LGBM	0.74 (0.09)	0.42 (0.11)		0.7 (0.1)		0.66 (0.18)		0.35 (0.14)
	LR	0.73 (0.09)	0.43 (0.11)		0.68 (0.13)		0.68 (0.13)		0.34 (0.15)
	RF	0.74 (0.07)	0.43 (0.11)		0.71 (0.14)		0.65 (0.15)		0.35 (0.13)
	XGB	0.73 (0.08)	0.41 (0.1)		0.62 (0.1)		0.8 (0.12)		0.29 (0.09)
Ratio=10
	LGBM	0.75 (0.09)	0.32 (0.09)		0.74 (0.13)		0.64 (0.2)		0.23 (0.07)
	LR	0.73 (0.1)	0.29 (0.08)		0.68 (0.13)		0.66 (0.11)		0.19 (0.07)
	RF	0.75 (0.08)	0.29 (0.07)		0.69 (0.12)		0.69 (0.13)		0.19 (0.05)
	XGB	0.75 (0.07)	0.32 (0.06)		0.72 (0.12)		0.66 (0.13)		0.22 (0.08)
Ratio=20
	LGBM	0.72 (0.07)	0.18 (0.07)		0.68 (0.13)		0.67 (0.13)		0.11 (0.05)
	LR	0.73 (0.06)	0.2 (0.06)		0.72 (0.14)		0.65 (0.2)		0.13 (0.06)
	RF	0.72 (0.06)	0.18 (0.04)		0.72 (0.09)		0.63 (0.12)		0.11 (0.03)
	XGB	0.74 (0.06)	0.17 (0.03)		0.67 (0.13)		0.69 (0.15)		0.1 (0.02)

^aAUC: area under the receiver operating characteristic curve.

^bLGBM: light gradient boosting machine.

^cLR: logistic regression.

^dRF: random forest.

^eXGB: extreme gradient boosting.