indicators = FALSE behavior

[tidymodels]

indicators = FALSE behavior

[DavisVaughan]

Maybe not because we still want to do things like log(var)

[DavisVaughan]

Hard because you might want to allow I(x * y) to be computed algebraically and not represent interactions

[DavisVaughan]

Adding some notes on what led me to the right answer:

when indicators = FALSE:

• check that factor columns are only by themselves (factorcol + x) not factorcol:x or paste0(factorcol)
  • uses terms() directly on the formula, along with original factor classes to determine this
  • throws error if factor columns are used not by themselves
• remove the factor columns from the data
• remove the factor columns from the formula? (append - factorcol)
• run model.matrix(formula_no_factors, data_no_factors)
• matrix -> tibble
• add the factor columns back onto the data
  • Original positions are not kept, but this seems ok

[DavisVaughan]

Closed by #42

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.