Skip to main content
. Author manuscript; available in PMC: 2017 Apr 1.
Published in final edited form as: Ann Stat. 2016 Apr;44(2):629–659. doi: 10.1214/15-AOS1380

Table 5.

Comparison of statistical performance. “Time” stands for the computational time in second. The numbers in parenthesis are the standard errors.

Method n = 100, d = 1000
AD FP FN Time
LASSO 2.558 (0.047) 5.700 (0.255) 0 (0) 2.332 (0.108)

GM1-SCAD 0.526 (0.017) 0.600 (0.084) 0 (0) 4.167 (0.254)

GM1-MCP 0.543 (0.018) 0.540 (0.073) 0 (0) 4.42 (0.874)

GM2-SCAD 3.816 (0.104) 18.360 (0.655) 0 (0) 3.968 (0.049)

GM2-MCP 0.548 (0.019) 0.610 (0.083) 0 (0) 3.916 (0.143)

SparseNet 1.012 (0.086) 5.850 (1.187) 0 (0) 2.154 (0.017)

Ncvreg-SCAD 1.068 (0.061) 9.220 (0.979) 0 (0) 0.733 (0.007)

Ncvreg-MCP 0.830 (0.045) 3.200 (0.375) 0 (0) 0.877 (0.009)

SCAD-LLA1 0.526 (0.017) 0.600 (0.084) 0 (0) 31.801 (1.533)

MCP-LLA1 0.543 (0.018) 0.540 (0.073) 0 (0) 28.695 (1.473)

MIPGO-SCAD 0.509 (0.017) 0 (0) 0 (0) 472.673 (97.982)

MIPGO-MCP 0.509 (0.017) 0 (0) 0 (0) 361.460 (70.683)

Oracle 0.509 (0.017)