R/normdist.R

In mkampert/rCOSA: Clustering Objects on Subsets of Attributes (COSA)

#' Normalize the distance
#'
#'
#' Normalize the dissimilarities sucht that the squared norm of the Dissimilarities equals N, the number of objects.
#'
#'
#' @param D A matrix of distances of class dist, or a vector or matrix containing distances.
#'
#'
#' @return
#' Returns a normalised version of D of similar class and mode.


normdist <- function(D) {
  fac <- sum(D**2)
  n <- dimdata(D)
  D <- D * ((n / fac)**.5)
  return(D)
}