Generate Dummy Codes for a Factor

Generate a matrix of dummy codes, also known as class indicators, for a factor or class vector.

Usage

dummy(
  x,
  method = c("treatment", "sum", "helmert", "poly", "full"),
  base = 1,
  levels = NULL
)

Arguments

x

factor or vector of classes.

method

character string specifying the contrast method. One of "treatment", "sum", "helmert", "poly", or "full". Abbreviations are accepted.

base

integer or character string specifying the baseline group. Only used for method = "treatment" (see Details).

levels

optional character vector specifying the possible levels of x. If NULL, levels are inferred by factor(x).

Value

a matrix with dummy codes. The number of rows equals length(x). For method = "full", the number of columns equals the number of levels. Otherwise, the number of columns equals the number of levels minus one.

The returned matrix has an attribute "base" containing the baseline level for treatment coding, and NA otherwise.

Details

The argument method controls the contrast coding. The option "full" returns one indicator column for each level of x. This full-rank coding is usually redundant for lm() and related modelling functions.

The base argument is only used by method = "treatment". The other contrast types have no freely choosable baseline: "sum" implicitly uses the last level as reference, "helmert" contrasts each level against the preceding ones, and "poly" uses orthogonal polynomials.

Column names reflect the semantics of the coding: level names for "treatment" (without the baseline), "full" (all levels), "sum" (without the last level) and "helmert" (without the first level); "poly" keeps the standard degree labels (.L, .Q, ...).

See also

model.frame(), contrasts(), contr.treatment(), contr.sum(), contr.helmert(), contr.poly()

Other data.recode: asBinary(), combLevels(), mReplace(), nf(), recodeX(), revCode(), stringsAsFactors()

Examples

x <- c("red", "blue", "green", "blue", "green", "red", "red", "blue")
dummy(x)
#>   green red
#> 1     0   1
#> 2     0   0
#> 3     1   0
#> 4     0   0
#> 5     1   0
#> 6     0   1
#> 7     0   1
#> 8     0   0
#> attr(,"base")
#> [1] "blue"
dummy(x, base = 2)
#>   blue red
#> 1    0   1
#> 2    1   0
#> 3    0   0
#> 4    1   0
#> 5    0   0
#> 6    0   1
#> 7    0   1
#> 8    1   0
#> attr(,"base")
#> [1] "green"
dummy(x, method = "sum")
#>   blue green
#> 1   -1    -1
#> 2    1     0
#> 3    0     1
#> 4    1     0
#> 5    0     1
#> 6   -1    -1
#> 7   -1    -1
#> 8    1     0
#> attr(,"base")
#> [1] NA

y <- c("Max", "Max", "Max", "Max", "Max", "Bill", "Bill", "Bill")
dummy(y)
#>   Max
#> 1   1
#> 2   1
#> 3   1
#> 4   1
#> 5   1
#> 6   0
#> 7   0
#> 8   0
#> attr(,"base")
#> [1] "Bill"
dummy(y, base = "Max")
#>   Bill
#> 1    0
#> 2    0
#> 3    0
#> 4    0
#> 5    0
#> 6    1
#> 7    1
#> 8    1
#> attr(,"base")
#> [1] "Max"
dummy(y, base = "Max", method = "full")
#>   Bill Max
#> 1    0   1
#> 2    0   1
#> 3    0   1
#> 4    0   1
#> 5    0   1
#> 6    1   0
#> 7    1   0
#> 8    1   0
#> attr(,"base")
#> [1] NA

# Revert full dummy coding
m <- dummy(y, method = "full")
apply(m, 1, function(z) colnames(m)[z == 1])
#>      1      2      3      4      5      6      7      8 
#>  "Max"  "Max"  "Max"  "Max"  "Max" "Bill" "Bill" "Bill" 

# Revert treatment dummy coding
m <- dummy(y)
apply(
  m,
  1,
  function(z) ifelse(sum(z) == 0, attr(m, "base"), colnames(m)[z == 1])
)
#>      1      2      3      4      5      6      7      8 
#>  "Max"  "Max"  "Max"  "Max"  "Max" "Bill" "Bill" "Bill"