dist.clust: Distance that incorporates weights from a clustering (hclust...

Description Usage Arguments

Description

Distance that incorporates weights from a clustering (hclust or pam)

Usage

1
2
dist.clust(x, clust.dist = "euclidean", clust.method = "complete",
  clust.k = 2, clust.weight = 0.5, clust.alt = FALSE)

Arguments

x

numeric matrix

clust.dist

character, determines what preliminary distance function is applied prior to clustering

clust.method

character, determines agglomeration method in hclust (e.g. complete, single, average) or "pam"

clust.k

integer, determines depth of clustering weights

clust.weight

numeric, determines weighting of cluster distances relative to neighbor rank distance. Set to 0 to obtain pure neighbor rank distance (clustering is ignored). Set to Inf to obtain pure cluster-based distance.

clust.alt

boolean. Set TRUE to obtain an alternative clustering. Default is FALSE which returns distances in which patterns from a usual distance are reinforced by clustering.


tkonopka/MultiPattern documentation built on May 31, 2019, 3:45 p.m.