Table 2.

Accuracy of the estimated number of true positives (TP) obtained using the proposed algorithm in the simulation study. Average of a tuning parameter (λ), number of selected genes ($\#\left\{j;{\widehat{\mathrm{\beta }}}_j\left(\mathrm{\lambda }\right)\ne \mathrm{0}\right\}$) in the lasso, true number of true positives (True TP), estimated number of TP ($\widehat{\mathrm{T}\mathrm{P}}$), and false positives ($\widehat{\mathrm{F}\mathrm{P}}$) are reported at λ_k (k = 5,10,50,100,150) of the solution path.

p 1	ρ	k	λ	$\#\left\{j;{\widehat{\mathrm{\beta }}}_j\left(\mathrm{\lambda }\right)\ne \mathrm{0}\right\}$	True TP	$\widehat{\mathrm{T}\mathrm{P}}$	$\widehat{\mathrm{F}\mathrm{P}}$
30	0	5	47.0	5.0	4.4	2.9	2.2
		10	40.8	10.1	8.0	5.8	4.3
		50	22.9	48.6	25.6	28.5	20.1
		100	12.6	86.7	29.9	32.1	54.7
		150	8.6	124.5	30.0	30.7	93.9
	0.5	5	48.6	5.0	4.1	2.8	2.2
		10	42.1	10.0	7.5	5.8	4.2
		50	23.5	48.1	25.2	31.9	16.3
		100	12.4	84.9	29.9	35.3	49.6
		150	8.4	121.2	30.0	31.6	89.6
5	0	5	66.9	5.0	5.0	3.0	2.0
		10	26.3	10.4	5.0	5.2	5.2
		50	17.2	50.1	5.0	5.2	44.9
		100	12.7	93.9	5.0	5.0	88.9
		150	9.8	128.4	5.0	5.0	123.4
	0.5	5	66.8	5.0	5.0	3.0	2.0
		10	26.5	10.3	5.0	5.2	5.1
		50	16.9	49.5	5.0	5.1	44.4
		100	12.4	92.1	5.0	5.0	87.1
		150	9.6	125.2	5.0	5.0	120.2