Is there a way to force clogitL1 or cv.clogitL1 to keep certain covariates in all the models? I'm thinking of something analogous to the penalty.factor argument in glmnet.
(Or, if there isn't a way to do it, I guess that's good to know too.)
↧
clogitL1 or cv.clogitL1 -- force variables into model?
↧