Search code examples
rr-caretglmnetstandardized

R caret glmnet standardize = FALSE

I'm trying to use the caret package to play with alpha levels for a glmnet model. The problem is that the data I'm using is all dummy variables and I don't want glmnet to standardize them. Usually if I was just using glmnet or cv.glmnet on its own, I'd just add 

standardize = FALSE

Is there a setting in caret to turn off standardize?

Solution

  • In caret, you can feed original function arguments into caret::train thanks to the ellipsis ... mechanism.

    For example this code will fit a regularized regression on non standardized data

    require(caret)
require(mlbench)
data(BostonHousing)

enet <-  train(medv ~ .,
               data = BostonHousing, 
               method = "glmnet",
               standardize = FALSE)