Bundling x$coefs as well currently just appends the bundle class and isn't technically necessary, but allows for a stacks extension of allowing users to blend predictions with a meta-learner other than glmnet that may actually need bundling. Same story for classing this bundler as model_stack rather than linear_stack. :)
Closes #20.
Bundling
x$coefs
as well currently just appends the bundle class and isn't technically necessary, but allows for a stacks extension of allowing users to blend predictions with a meta-learner other than glmnet that may actually need bundling. Same story for classing this bundler asmodel_stack
rather thanlinear_stack
. :)Related to tidymodels/stacks#145!