step_novel() creates a specification of a recipe step that will assign a
previously unseen factor level to "new".
step_novel(
recipe,
...,
role = NA,
trained = FALSE,
new_level = "new",
objects = NULL,
skip = FALSE,
id = rand_id("novel")
)A recipe object. The step will be added to the sequence of operations for this recipe.
One or more selector functions to choose variables for this step.
See selections() for more details.
Not used by this step since no new variables are created.
A logical to indicate if the quantities for preprocessing have been estimated.
A single character value that will be assigned to new factor levels.
A list of objects that contain the information on factor
levels that will be determined by prep().
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.
A character string that is unique to this step to identify it.
An updated version of recipe with the new step added to the
sequence of any existing operations.
The selected variables are adjusted to have a new level (given by
new_level) that is placed in the last position. During preparation there
will be no data points associated with this new level since all of the data
have been seen.
Note that if the original columns are character, they will be converted to factors by this step.
Missing values will remain missing.
If new_level is already in the data given to prep(), an error is thrown.
When fitting a model that can deal with new factor levels, consider using
workflows::add_recipe() with allow_novel_levels = TRUE set in
hardhat::default_recipe_blueprint(). This will allow your model to handle
new levels at prediction time, instead of throwing warnings or errors.
When you tidy() this step, a tibble is returned with
columns terms, value , and id:
character, the selectors or variables selected
character, the factor levels that are used for the new value
character, id of this step
The underlying operation does not allow for case weights.
Other dummy variable and encoding steps:
step_bin2factor(),
step_count(),
step_date(),
step_dummy(),
step_dummy_extract(),
step_dummy_multi_choice(),
step_factor2string(),
step_holiday(),
step_indicate_na(),
step_integer(),
step_num2factor(),
step_ordinalscore(),
step_other(),
step_regex(),
step_relevel(),
step_string2factor(),
step_time(),
step_unknown(),
step_unorder()
data(Sacramento, package = "modeldata")
sacr_tr <- Sacramento[1:800, ]
sacr_te <- Sacramento[801:806, ]
# Without converting the predictor to a character, the new level would be converted
# to `NA`.
sacr_te$city <- as.character(sacr_te$city)
sacr_te$city[3] <- "beeptown"
sacr_te$city[4] <- "boopville"
sacr_te$city <- as.factor(sacr_te$city)
rec <- recipe(~ city + zip, data = sacr_tr)
rec <- rec |>
step_novel(city, zip)
rec <- prep(rec, training = sacr_tr)
processed <- bake(rec, sacr_te)
tibble(old = sacr_te$city, new = processed$city)
#> # A tibble: 6 × 2
#> old new
#> <fct> <fct>
#> 1 SACRAMENTO SACRAMENTO
#> 2 AUBURN AUBURN
#> 3 beeptown new
#> 4 boopville new
#> 5 SACRAMENTO SACRAMENTO
#> 6 ROSEVILLE ROSEVILLE
tidy(rec, number = 1)
#> # A tibble: 2 × 3
#> terms value id
#> <chr> <chr> <chr>
#> 1 city new novel_B2lWN
#> 2 zip new novel_B2lWN