Skip to main content
. Author manuscript; available in PMC: 2017 Mar 21.
Published in final edited form as: J Mach Learn Res. 2015 Mar;16:553–557.

Table 2.

Average timing performance (in seconds) with standard errors in the parentheses on sparse linear regression and sparse precision matrix estimation.

Sparse Linear Regression
Method d = 375 d = 750 d = 1500 d = 3000
LAD Lasso 1.1713(0.2915) 1.1046(0.3640) 1.8103(0.2919) 3.1378(0.7753)
ℓ1.5 Lasso 12.995(0.5535) 14.071(0.5966) 14.382(0.7390) 16.936(0.5696)
Dantzig selector 0.3245(0.1871) 1.5360(1.8566) 4.4669(5.9929) 17.034(23.202)
SQRT Lasso (flare) 0.4888(0.0264) 0.7330(0.1234) 0.9485(0.2167) 1.2761(0.1510)
SQRT Lasso (glmnet) 0.6417(0.0341) 0.8794(0.0159) 1.1406(0.0440) 2.1675(0.0937)
Sparse Precision Matrix Estimation
Method d = 100 d = 200 d = 300 d=400
TIGER 1.0637(0.0361) 4.6251(0.0807) 7.1860(0.0795) 11.085(0.1715)
CLIME 2.5761(0.3807) 20.137(3.2258) 42.882(18.188) 112.50(11.561)