Perform supervised lumping on a nominal variable

Lumps the levels of a nominal variable so as to preserve as much mutual information as possible between the lumped variable and a supplied outcome. The outcome may be discrete (a factor) or continuous (numeric); see vignette("supervised") for the distinction.

Usage

lump_nominal_supervised(
  data,
  outcome,
  threshold,
  adj_matrix = NULL,
  verbose = FALSE,
  level_namer = default_level_namer,
  outcome_mode = c("auto", "discrete", "continuous")
)

Arguments

data

Factor or character vector of the categorical data.

outcome

Factor or character vector. Variable to be used as a source of information about data.

threshold

The minimum number of samples each lumped level should contain.

adj_matrix

Adjancency matrix of the preference graph. Default: a complete graph, allowing all lumpings.

verbose

Logical value dictating if values should be printed. Default: FALSE.

level_namer

Function that takes a character vector of the original levels in a lump and returns the name of the new lumped level. Default: concatenating the original levels with a "+" in between.

outcome_mode

Whether to treat the outcome as discrete or continuous. Default: inferred based on the type of outcome.

Value

A factor vector with the lumped levels.

See also

maximum_mutual_information_nominal_supervised() for the underlying algorithm that this function wraps.

lump_hierarchical_supervised() for a version of this function that can take advantage of hierarchical structure in the data to speed up the execution time.

Author

Daan Koning

Examples

data    <-        c("NL", "NL", "DE", "DE", "FR", "FR", "FR", "BE")
outcome <- factor(c(  1,    0,    1,    1,    0,    0,    1,    1))
lump_nominal_supervised(data, outcome, threshold = 3)
#> [1] FR+NL FR+NL BE+DE BE+DE FR+NL FR+NL FR+NL BE+DE
#> Levels: BE+DE FR+NL