reorgByCluNo {wrMisc}

Reorganize matrix according to clustering-output

Description

Reorganize input matrix as sorted by cluster numbers (and geometric mean) according to vector with cluster names, and index for sorting per cluster and per geometric mean. In case mat is an array, the 3rd dimension will be considered as 'column' with arguments useColumn ( and cluNo, if it designs a 'column' of mat).

Usage

reorgByCluNo(
  mat,
  cluNo,
  useColumn = NULL,
  meanCol = NULL,
  addInfo = TRUE,
  retList = FALSE,
  silent = FALSE,
  callFrom = NULL,
  debug = FALSE
)

Arguments

mat

(matrix or data.frame) main input

cluNo

(positive integer, length to match nrow(dat) initial cluster numbers for each line of 'mat' (obtained by separate clustering or other segmentation) or may desinn column of mat to use as cluster-numbers

useColumn

(character or integer) the columns to use from mat as main data (default will use all, exept cluCol and/or meanCol if they design columns))

meanCol

(character or integer) alternative summarizing data for intra-cluster sorting (instead of geometric mean)

addInfo

(logical) allows adding of columns 'index', 'geoMean' and 'cluNo' (or array if FALSE)

retList

(logical) return as list of matrixes (or array if FALSE)

silent

(logical) suppress messages

callFrom

(character) allow easier tracking of messages produced

debug

(logical) additional messages for debugging

Value

This function returns a list or array (as 2- or 3 dim) with possible number of occurances for each of the 3 elements in nMax. Read results vertical : out[[1]] or out[,,1] .. (multiplicative) table for 1st element of nMax; out[,,2] .. for 2nd

See Also

pairwise combinations combn, clustering kmeans

Examples

dat1 <- matrix(round(runif(24),2), ncol=3, dimnames=list(NULL,letters[1:3]))
clu <- stats::kmeans(dat1, 5)$cluster
reorgByCluNo(dat1, clu) 

dat2 <- cbind(dat1, clu=clu) 
reorgByCluNo(dat2, "clu")