Add new variables using dplyr

step_mutate() creates a specification of a recipe step that will add variables using dplyr::mutate().

Usage

step_mutate(
  recipe,
  ...,
  .pkgs = character(),
  role = "predictor",
  trained = FALSE,
  inputs = NULL,
  skip = FALSE,
  id = rand_id("mutate")
)

Arguments

recipe: A recipe object. The step will be added to the sequence of operations for this recipe.
...: Name-value pairs of expressions. See dplyr::mutate().
.pkgs: Character vector, package names of functions used in expressions .... Should be specified if using non-base functions.
role: For model terms created by this step, what analysis role should they be assigned? By default, the new columns created by this step from the original variables will be used as predictors in a model.
trained: A logical to indicate if the quantities for preprocessing have been estimated.
inputs: Quosure(s) of ....
skip: A logical. Should the step be skipped when the recipe is baked by bake()? While all operations are baked when prep() is run, some operations may not be able to be conducted on new data (e.g. processing the outcome variable(s)). Care should be taken when using skip = TRUE as it may affect the computations for subsequent operations.
id: A character string that is unique to this step to identify it.

Value

An updated version of recipe with the new step added to the sequence of any existing operations.

Details

When using this flexible step, use extra care to avoid data leakage in your preprocessing. Consider, for example, the transformation x = w > mean(w). When applied to new data or testing data, this transformation would use the mean of w from the new data, not the mean of w from the training data.

When an object in the user's global environment is referenced in the expression defining the new variable(s), it is a good idea to use quasiquotation (e.g. !!) to embed the value of the object in the expression (to be portable between sessions). See the examples.

If a preceding step removes a column that is selected by name in step_mutate(), the recipe will error when being estimated with prep().

Tidying

When you tidy() this step, a tibble is returned with columns terms, value , and id:

terms: character, the selectors or variables selected
value: character, expression passed to mutate()
id: character, id of this step

Case weights

The underlying operation does not allow for case weights.

Examples

rec <-
  recipe(~., data = iris) %>%
  step_mutate(
    dbl_width = Sepal.Width * 2,
    half_length = Sepal.Length / 2
  )

prepped <- prep(rec, training = iris %>% slice(1:75))

library(dplyr)

dplyr_train <-
  iris %>%
  as_tibble() %>%
  slice(1:75) %>%
  mutate(
    dbl_width = Sepal.Width * 2,
    half_length = Sepal.Length / 2
  )

rec_train <- bake(prepped, new_data = NULL)
all.equal(dplyr_train, rec_train)
#> [1] TRUE

dplyr_test <-
  iris %>%
  as_tibble() %>%
  slice(76:150) %>%
  mutate(
    dbl_width = Sepal.Width * 2,
    half_length = Sepal.Length / 2
  )
rec_test <- bake(prepped, iris %>% slice(76:150))
all.equal(dplyr_test, rec_test)
#> [1] TRUE

# Embedding objects:
const <- 1.414

qq_rec <-
  recipe(~., data = iris) %>%
  step_mutate(
    bad_approach = Sepal.Width * const,
    best_approach = Sepal.Width * !!const
  ) %>%
  prep(training = iris)

bake(qq_rec, new_data = NULL, contains("appro")) %>% slice(1:4)
#> # A tibble: 4 × 2
#>   bad_approach best_approach
#>          <dbl>         <dbl>
#> 1         4.95          4.95
#> 2         4.24          4.24
#> 3         4.52          4.52
#> 4         4.38          4.38

# The difference:
tidy(qq_rec, number = 1)
#> # A tibble: 2 × 3
#>   terms         value               id          
#>   <chr>         <chr>               <chr>       
#> 1 bad_approach  Sepal.Width * const mutate_p75TX
#> 2 best_approach Sepal.Width * 1.414 mutate_p75TX