step_poly() creates a specification of a recipe step that will create new
columns that are basis expansions of variables using orthogonal polynomials.
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.
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.
A logical to indicate if the quantities for preprocessing have been estimated.
A list of stats::poly() objects created once the step has
been trained.
The polynomial degree (an integer).
A list of options for stats::poly() which should not include
x, degree, or simple. Note that the option raw = TRUE will produce
the regular polynomial values (not orthogonalized).
A logical to keep the original variables in the
output. Defaults to FALSE.
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.
step_poly() can create new features from a single variable that enable
fitting routines to model this variable in a nonlinear manner. The extent of
the possible nonlinearity is determined by the degree argument of
stats::poly(). The original variables are removed from the data by default,
but can be retained by setting keep_original_cols = TRUE and new columns
are added. The naming convention for the new variables is varname_poly_1
and so on.
The orthogonal polynomial expansion is used by default because it yields
variables that are uncorrelated and doesn't produce large values which would
otherwise be a problem for large values of degree. Orthogonal polynomial
expansion pick up the same signal as their uncorrelated counterpart.
When you tidy() this step, a tibble is returned with
columns terms, degree , and id:
character, the selectors or variables selected
integer, the polynomial degree
character, id of this step
This step has 1 tuning parameters:
degree: Polynomial Degree (type: integer, default: 2)
The underlying operation does not allow for case weights.
Other individual transformation steps:
step_BoxCox(),
step_YeoJohnson(),
step_bs(),
step_harmonic(),
step_hyperbolic(),
step_inverse(),
step_invlogit(),
step_log(),
step_logit(),
step_mutate(),
step_ns(),
step_percentile(),
step_relu(),
step_sqrt()
data(biomass, package = "modeldata")
biomass_tr <- biomass[biomass$dataset == "Training", ]
biomass_te <- biomass[biomass$dataset == "Testing", ]
rec <- recipe(
HHV ~ carbon + hydrogen + oxygen + nitrogen + sulfur,
data = biomass_tr
)
quadratic <- rec |>
step_poly(carbon, hydrogen)
quadratic <- prep(quadratic, training = biomass_tr)
expanded <- bake(quadratic, biomass_te)
expanded
#> # A tibble: 80 × 8
#> oxygen nitrogen sulfur HHV carbon_poly_1 carbon_poly_2 hydrogen_poly_1
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 47.2 0.3 0.22 18.3 -0.00903 -0.0124 0.00826
#> 2 48.1 2.85 0.34 17.6 -0.0230 0.00403 0.00160
#> 3 49.1 2.4 0.3 17.2 -0.0255 0.00734 0.00160
#> 4 37.3 1.8 0.5 18.9 -0.00880 -0.0126 0.0251
#> 5 42.8 0.2 0 20.5 0.00183 -0.0226 0.0337
#> 6 41.7 0.7 0.2 18.5 -0.0183 -0.00195 0.00160
#> 7 54.1 1.19 0.51 15.1 -0.0424 0.0331 -0.00897
#> 8 33.8 0.95 0.2 16.2 -0.0282 0.0111 -0.0313
#> 9 31.1 0.14 4.9 11.1 -0.0863 0.125 -0.0415
#> 10 23.7 4.63 1.05 10.8 -0.0926 0.142 -0.0662
#> # ℹ 70 more rows
#> # ℹ 1 more variable: hydrogen_poly_2 <dbl>
tidy(quadratic, number = 1)
#> # A tibble: 2 × 3
#> terms degree id
#> <chr> <int> <chr>
#> 1 carbon 2 poly_R8bgI
#> 2 hydrogen 2 poly_R8bgI