[R] About lm.ridge in the MASS package

baghaliforoosh@Yahoo.com baghaliforoosh at yahoo.com
Tue Jun 8 03:43:37 CEST 2010


I had a questions about doing ridge regression in R. 

Why is it that when I try this on datasets with more predictors than samples (p>n) using lambda=0, it still finds coefficients for all predictors? I thought when lambda=0, it should be like ordinary regression and therefore not find coefficients for all due to singularity? 

I would greatly appreciate your help.

Thank you,

--James K.

More information about the R-help mailing list