Is there a robust scaler method in the recipes package in the R programming language? As a result of my research, I could not find this method.
1 Answers
I'm assuming you are referring to the RobustScaler from scikit-learn. You are correct that there isn't a similar step in the recipes package.
It is implemented in the extrasteps package which you can install with
# install.packages("devtools")
devtools::install_github("EmilHvitfeldt/extrasteps")
Then you can use the step_robust() which will do what you are expecting.
library(recipes)
library(extrasteps)
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
#> # … with 22 more rows
tidy(rec, 1)
#> # A tibble: 33 × 4
#> terms statistic value id
#> <chr> <chr> <dbl> <chr>
#> 1 mpg lower 15.4 robust_hS9q6
#> 2 mpg median 19.2 robust_hS9q6
#> 3 mpg higher 22.8 robust_hS9q6
#> 4 cyl lower 4 robust_hS9q6
#> 5 cyl median 6 robust_hS9q6
#> 6 cyl higher 8 robust_hS9q6
#> 7 disp lower 121. robust_hS9q6
#> 8 disp median 196. robust_hS9q6
#> 9 disp higher 326 robust_hS9q6
#> 10 hp lower 96.5 robust_hS9q6
#> # … with 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
#> # … with 22 more rows, and 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_MygTA
#> 2 mpg median 19.2 robust_MygTA
#> 3 mpg higher 30.1 robust_MygTA
#> 4 cyl lower 4 robust_MygTA
#> 5 cyl median 6 robust_MygTA
#> 6 cyl higher 8 robust_MygTA
#> 7 disp lower 80.6 robust_MygTA
#> 8 disp median 196. robust_MygTA
#> 9 disp higher 396 robust_MygTA
#> 10 hp lower 66 robust_MygTA
#> # … with 23 more rows
Related