auto_cut: Discretizes a Vector

In flashlight: Shed Light on Black Box Machine Learning Models

Discretizes a Vector

Description

This function takes a vector x and returns a list with information on disretized version of x. The construction of level names can be controlled by passing ... arguments to formatC().

Usage

auto_cut(
  x,
  breaks = NULL,
  n_bins = 27L,
  cut_type = c("equal", "quantile"),
  x_name = "value",
  level_name = "level",
  ...
)

Arguments

x	A vector.
breaks	An optional vector of breaks. Only relevant for numeric x.
n_bins	If x is numeric and no breaks are provided, this is the maximum number of bins allowed or to be created (approximately).
cut_type	For the default type "equal", bins of equal width are created by pretty(). Choose "quantile" to create quantile bins.
x_name	Column name with the values of x in the output.
level_name	Column name with the bin labels of x in the output.
...	Further arguments passed to cut3().

Value

A list with the following elements:

• data: A data.frame with colums x_name and level_name each with the same length as x. The column x_name has values in output bin_means while the column level_name has values in bin_labels.

• breaks: A vector of increasing and unique breaks used to cut a numeric x with too many distinct levels. NULL otherwise.

• bin_means: The midpoints of subsequent breaks, or if there are no breaks in the output, factor levels or distinct values of x.

• bin_labels: Break labels of the form "(low, high]" if there are breaks in the output, otherwise the same as bin_means. Same order as bin_means.

Examples

auto_cut(1:10, n_bins = 3)
auto_cut(c(NA, 1:10), n_bins = 3)
auto_cut(1:10, breaks = 3:4, n_bins = 3)
auto_cut(1:10, n_bins = 3, cut_type = "quantile")
auto_cut(LETTERS[4:1], n_bins = 2)
auto_cut(factor(LETTERS[1:4], LETTERS[4:1]), n_bins = 2)
auto_cut(990:1100, n_bins = 3, big.mark = "'", format = "fg")
auto_cut(c(0.0001, 0.0002, 0.0003, 0.005), n_bins = 3, format = "fg")

flashlight documentation built on May 31, 2023, 6:19 p.m.