Perform Robust Scaling — step_robust • extrasteps

step_robust() creates a specification of a recipe step that will perform Robust scaling.

Usage

step_robust(
  recipe,
  ...,
  role = NA,
  trained = FALSE,
  range = c(0.25, 0.75),
  res = NULL,
  columns = NULL,
  skip = FALSE,
  id = rand_id("robust")
)

Arguments

recipe: A recipe object. The step will be added to the sequence of operations for this recipe.
...: One or more selector functions to choose which variables are affected by the step. See recipes::selections() for more details. For the tidy method, these are not currently used.
role: Not used by this step since no new variables are created.
trained: A logical to indicate if the quantities for preprocessing have been estimated.
range: A numeric vector with 2 values denoting the lower and upper quantile that is used for scaling. Defaults to c(0.25, 0.75).
res: A list containing the 3 quantiles of training variables is stored here once this preprocessing step has be trained by recipes::prep().
columns: A character string of variable names that will be populated (eventually) by the terms argument.
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 existing steps (if any). For the tidy method, a tibble with columns terms (the columns that will be affected) and base.

Details

The scaling performed by this step is done using the following transformation

$$x_new = (x - Q2(x)) / (Q3(x) - Q1(x))$$

where Q2(x) is the median, Q3(x) is the upper quantile (defaults to 0.75) and Q1(x) is the lower quantile (defaults to 0.25). The upper and lower quantiles can be changed with the range argument.

Examples

library(recipes)

rec <- recipe(~., data = mtcars) %>%
  step_robust(all_predictors()) %>%
  prep()

rec %>%
  bake(new_data = NULL)
#> # A tibble: 32 × 11
#>        mpg   cyl   disp     hp     drat     wt   qsec    vs    am  gear  carb
#>      <dbl> <dbl>  <dbl>  <dbl>    <dbl>  <dbl>  <dbl> <dbl> <dbl> <dbl> <dbl>
#>  1  0.244    0   -0.177 -0.156  0.244   -0.685 -0.623     0     1     0   1  
#>  2  0.244    0   -0.177 -0.156  0.244   -0.437 -0.344     0     1     0   1  
#>  3  0.488   -0.5 -0.430 -0.359  0.185   -0.977  0.448     1     1     0  -0.5
#>  4  0.298    0    0.301 -0.156 -0.732   -0.107  0.862     1     0    -1  -0.5
#>  5 -0.0678   0.5  0.798  0.623 -0.649    0.112 -0.344     0     0    -1   0  
#>  6 -0.149    0    0.140 -0.216 -1.11     0.131  1.25      1     0    -1  -0.5
#>  7 -0.664    0.5  0.798  1.46  -0.577    0.238 -0.932     0     0    -1   1  
#>  8  0.705   -0.5 -0.242 -0.731 -0.00595 -0.131  1.14      1     0     0   0  
#>  9  0.488   -0.5 -0.271 -0.335  0.268   -0.170  2.59      1     0     0   0  
#> 10  0        0   -0.140  0      0.268    0.112  0.294     1     0     0   1  
#> # ℹ 22 more rows

tidy(rec, 1)
#> # A tibble: 33 × 4
#>    terms statistic value id          
#>    <chr> <chr>     <dbl> <chr>       
#>  1 mpg   lower      15.4 robust_afHYT
#>  2 mpg   median     19.2 robust_afHYT
#>  3 mpg   higher     22.8 robust_afHYT
#>  4 cyl   lower       4   robust_afHYT
#>  5 cyl   median      6   robust_afHYT
#>  6 cyl   higher      8   robust_afHYT
#>  7 disp  lower     121.  robust_afHYT
#>  8 disp  median    196.  robust_afHYT
#>  9 disp  higher    326   robust_afHYT
#> 10 hp    lower      96.5 robust_afHYT
#> # ℹ 23 more rows

rec <- recipe(~., data = mtcars) %>%
  step_robust(all_predictors(), range = c(0.1, 0.9)) %>%
  prep()

rec %>%
  bake(new_data = NULL)
#> # A tibble: 32 × 11
#>        mpg   cyl    disp      hp     drat      wt   qsec    vs    am  gear
#>      <dbl> <dbl>   <dbl>   <dbl>    <dbl>   <dbl>  <dbl> <dbl> <dbl> <dbl>
#>  1  0.114    0   -0.115  -0.0732  0.171   -0.337  -0.281     0     1   0  
#>  2  0.114    0   -0.115  -0.0732  0.171   -0.215  -0.155     0     1   0  
#>  3  0.229   -0.5 -0.280  -0.169   0.129   -0.480   0.202     1     1   0  
#>  4  0.140    0    0.196  -0.0732 -0.512   -0.0526  0.388     1     0  -0.5
#>  5 -0.0317   0.5  0.519   0.293  -0.453    0.0550 -0.155     0     0  -0.5
#>  6 -0.0698   0    0.0910 -0.101  -0.778    0.0645  0.563     1     0  -0.5
#>  7 -0.311    0.5  0.519   0.687  -0.403    0.117  -0.420     0     0  -0.5
#>  8  0.330   -0.5 -0.157  -0.344  -0.00416 -0.0645  0.514     1     0   0  
#>  9  0.229   -0.5 -0.176  -0.158   0.187   -0.0837  1.16      1     0   0  
#> 10  0        0   -0.0910  0       0.187    0.0550  0.132     1     0   0  
#> # ℹ 22 more rows
#> # ℹ 1 more variable: carb <dbl>

tidy(rec, 1)
#> # A tibble: 33 × 4
#>    terms statistic value id          
#>    <chr> <chr>     <dbl> <chr>       
#>  1 mpg   lower      14.3 robust_ayUXK
#>  2 mpg   median     19.2 robust_ayUXK
#>  3 mpg   higher     30.1 robust_ayUXK
#>  4 cyl   lower       4   robust_ayUXK
#>  5 cyl   median      6   robust_ayUXK
#>  6 cyl   higher      8   robust_ayUXK
#>  7 disp  lower      80.6 robust_ayUXK
#>  8 disp  median    196.  robust_ayUXK
#>  9 disp  higher    396   robust_ayUXK
#> 10 hp    lower      66   robust_ayUXK
#> # ℹ 23 more rows