randIndFx {wrMisc}

Distance of categorical data (Jaccard, Rand and adjusted Rand index)

Description

randIndFx calculates distance of categorical data (as Rand Index, Adjusted Rand Index or Jaccard Index). Note: uses/requires package flexclust Methods so far available (via flexclust): "ARI" .. adjusted Rand Index, "RI" .. Rand index, "J" .. Jaccard, "FM" .. Fowlkes-Mallows.

Usage

randIndFx(
  ma,
  method = "ARI",
  adjSense = TRUE,
  silent = FALSE,
  debug = FALSE,
  callFrom = NULL
)

Arguments

ma

(matrix) main input for distance calulation

method

(character) name of distance method (eg "ARI","RI","J","FM")

adjSense

(logical) allows introducing correlation/anticorrelation (interprete neg distance results as anti)

silent

(logical) suppress messages

debug

(logical) additional messages for debugging

callFrom

(character) allow easier tracking of messages produced

Value

This function returns a distance matrix

See Also

comPart in randIndex

Examples

set.seed(2016); tab2 <- matrix(sample(1:2, size=42, replace=TRUE), ncol=7)
if(requireNamespace("flexclust")) { flexclust::comPart(tab2[1,], tab2[2,])
  flexclust::comPart(tab2[1,], tab2[3,])
  flexclust::comPart(tab2[1,], tab2[4,]) }
## via randIndFx():
  randIndFx(tab2, adjSense=FALSE)
  cor(t(tab2))
  randIndFx(tab2, adjSense=TRUE)