This PR simplifies the linear baseline model by removing the effect encoder step used in its preprocessing pipeline and replacing it with standard one-hot features. This is to reduce the dependency weight of the pipeline, as the embed recipes package has enormous sub-dependencies.
This PR also bumps all the R/renv dependencies used by each of the pipeline profiles to their latest version available on CRAN.
This PR simplifies the linear baseline model by removing the effect encoder step used in its preprocessing pipeline and replacing it with standard one-hot features. This is to reduce the dependency weight of the pipeline, as the
embed
recipes package has enormous sub-dependencies.This PR also bumps all the R/renv dependencies used by each of the pipeline profiles to their latest version available on CRAN.