High Dimensional Variable Selection with Error Control

Table 1

The true positive rate (TP), the false positive rate (FP), and the false discovery rate (FDR), the average number of false positives (ANFP) in the final models, the final model size (size), the area under the curve (AUROC), and the number of filtered truly important variables from FDR and RF (# filter) under the low correlation coefficients, = 0.1 among variables.

Screening	Method	TP	FP	FDR	ANFP	Size	AUROC	# filter
	LASSO	0.22336; (0.00203)	(0)	0.03086; (0.00303)	0.216; (0.02199)	5.8; (0.05816)	0.80285; (0.00157)	5.826 (0.05985)
	SCAD	0.22336; (0.00203)	(0)	0.03071; (0.003)	0.214; (0.02157)	5.798; (0.05791)	0.80286; (0.00157)
	MCP	0.22336; (0.00203)	(0)	0.03071; (0.003)	0.214; (0.02157)	5.798; (0.05791)	0.80286; (0.00157)
	LASSO	0.2676; (0.00229)	(0)	0.12618; (0.0055)	1.17; (0.05914)	7.86; (0.09344)	0.81965; (0.00154)	7.904 (0.09499)
	SCAD	0.2676; (0.00229)	(0)	0.12522; (0.00547)	1.158; (0.05866)	7.848; (0.09288)	0.81964; (0.00154)
	MCP	0.2676; (0.00229)	(0)	0.12355; (0.00541)	1.134; (0.05698)	7.824; (0.09096)	0.81967; (0.00154)
	LASSO	0.42112; (0.00236)	0.0001237; (0)	0.53987; (0.00265)	12.364; (0.06298)	22.892; (0.01389)	0.83244; (0.00149)	13.286 (0.06807)
	SCAD	0.42336; (0.00243)	0.0001226; (0)	0.53635; (0.00276)	12.26; (0.06622)	22.844; (0.01624)	0.83196; (0.00147)
	MCP	0.42656; (0.00249)	0.0001217; (0)	0.53261; (0.00283)	12.17; (0.06797)	22.834; (0.01666)	0.83191; (0.00147)

( ): standard deviation.