step_lincomp: Linear Components Variable Reduction

In MachineShop: Machine Learning Models and Tools

Linear Components Variable Reduction

Description

Creates a specification of a recipe step that will compute one or more linear combinations of a set of numeric variables according to a user-specified transformation matrix.

Usage

step_lincomp(
  recipe,
  ...,
  transform,
  num_comp = 5,
  options = list(),
  center = TRUE,
  scale = TRUE,
  replace = TRUE,
  prefix = "LinComp",
  role = "predictor",
  skip = FALSE,
  id = recipes::rand_id("lincomp")
)

## S3 method for class 'step_lincomp'
tidy(x, ...)

## S3 method for class 'step_lincomp'
tunable(x, ...)

Arguments

recipe	recipe object to which the step will be added.
...	one or more selector functions to choose which variables will be used to compute the components. See selections for more details. These are not currently used by the tidy method.
transform	function whose first argument x is a matrix of variables with which to compute linear combinations and second argument step is the current step. The function should return a transformation matrix or Matrix of variable weights in its columns, or return a list with element `weights` containing the transformation matrix and possibly with other elements to be included as attributes in output from the tidy method.
num_comp	number of components to derive. The value of num_comp will be constrained to a minimum of 1 and maximum of the number of original variables when prep is run.
options	list of elements to be added to the step object for use in the transform function.
center, scale	logicals indicating whether to mean center and standard deviation scale the original variables prior to deriving components, or functions or names of functions for the centering and scaling.
replace	logical indicating whether to replace the original variables.
prefix	character string prefix added to a sequence of zero-padded integers to generate names for the resulting new variables.
role	analysis role that added step variables should be assigned. By default, they are designated as model predictors.
skip	logical indicating whether to skip the step when the recipe is baked. While all operations are baked when prep is run, some operations may not be applicable to new data (e.g. processing outcome variables). Care should be taken when using skip = TRUE as it may affect the computations for subsequent operations.
id	unique character string to identify the step.
x	step_lincomp object.

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 (selectors or variables selected), weight of each variable in the linear transformations, and name of the new variable names.

See Also

recipe, prep, bake

Examples

library(recipes)

pca_mat <- function(x, step) {
  prcomp(x)$rotation[, 1:step$num_comp, drop = FALSE]
}

rec <- recipe(rating ~ ., data = attitude)
lincomp_rec <- rec %>%
  step_lincomp(all_numeric_predictors(),
               transform = pca_mat, num_comp = 3, prefix = "PCA")

lincomp_prep <- prep(lincomp_rec, training = attitude)
lincomp_data <- bake(lincomp_prep, attitude)

pairs(lincomp_data, lower.panel = NULL)

tidy(lincomp_rec, number = 1)
tidy(lincomp_prep, number = 1)

MachineShop documentation built on Sept. 18, 2023, 5:06 p.m.