David Reiss
2007-Jan-06 23:25 UTC
[R] has anyone implemented LARS with the "positive lasso"?
Hi, I am interested in a modification to LARS that allows for positive-only constraints in the variables (with details about how to implement this as described in section 3.4 of the Efron et al (2003) LARS paper). Before I dive into the "lars" package code myself, I was wondering if anyone knew of a version where this is available, or if another package that I have not found can do this? Thanks very much for any pointers. -David [[alternative HTML version deleted]]