Table 4. Predictive performance of different approaches based on the variance-stability transformation for dataset 1.

	Dataset 1, including random effect W			Dataset 1, excluding random effect W
Method	ER (SD)	AUC (SD)	AUPRC (SD)	ER (SD)	AUC (SD)	AUPRC (SD)
Oracle	0.04 (0.02)	0.98 (0.02)	0.97 (0.03)	0.04 (0.02)	0.98 (0.02)	0.32 (0.1)
LASSO	0.21 (0.05)	0.87 (0.04)	0.87 (0.07)	0.22 (0.05)	0.86 (0.05)	0.86 (0.07)
ZINB+GLM	0.24 (0.06)	0.82 (0.06)	0.82 (0.09)	0.25 (0.06)	0.81 (0.07)	0.81 (0.09)
TPNB+GLM	0.2 (0.05)	0.87 (0.05)	0.86 (0.08)	0.21 (0.05)	0.86 (0.05)	0.85 (0.09)
NB+GLM	0.23 (0.05)	0.83 (0.05)	0.83 (0.08)	0.24 (0.06)	0.82 (0.06)	0.82 (0.09)
ZINB+LASSO	0.24 (0.06)	0.82 (0.06)	0.82 (0.09)	0.25 (0.06)	0.81 (0.07)	0.81 (0.09)
TPNB+LASSO	0.2 (0.05)	0.87 (0.05)	0.86 (0.08)	0.21 (0.05)	0.86 (0.05)	0.85 (0.09)
NB+LASSO	0.23 (0.05)	0.83 (0.05)	0.83 (0.08)	0.24 (0.06)	0.82 (0.06)	0.82 (0.09)