Figure - PMC

Skip to main content

View full-text article in PMC

. Author manuscript; available in PMC: 2014 Dec 11.

Published in final edited form as: J R Stat Soc Series B Stat Methodol. 2012 Mar;74(2):245–266. doi: 10.1111/j.1467-9868.2011.01004.x

Fig. 10 — Gaussian lasso setting, N = 200, p =20000, pairwise correlation between features of 0.7 (the first 50 predictors have positive, decreasing coefficients): Shown are the number of predictors left after applying the strong sequential rule (6)(2) and the number that have ever been active (1) (i.e. had a non-zero coefficient in the solution) for values of λ larger than the current value ( , unit slope line for reference); (b) is a zoomed version of the full scale plot (a)

Inline graphic — Gaussian lasso setting, N = 200, p =20000, pairwise correlation between features of 0.7 (the first 50 predictors have positive, decreasing coefficients): Shown are the number of predictors left after applying the strong sequential rule (6)(2) and the number that have ever been active (1) (i.e. had a non-zero coefficient in the solution) for values of λ larger than the current value ( , unit slope line for reference); (b) is a zoomed version of the full scale plot (a)