subclust: subclust

In cbsteen/clustermap: clustermap: a Versatile Toolbox in R for Clustering and Heatmap Visualisation

Description

The function identifies subclusters in a dendrogram. The effect is that later calls to draw.tree will show the subclusters in different colors. You can either provide the desired number of clusters using the argument k, or you can set k=NA to automatically estimate the number of cluster with one of two algorithms: gap (Tibshirani et al. (2001), J Roy Statist Soc B, 63: 411-423) or PART (Nilsen et al (2012), Stat Appl Genet Mol Biol, 12: 637-652).

Usage

subclust(k = NA, clust = "col", method = "gap", B = 50, min.size = 5,
  max.level = 3)

Arguments

k	Integer. Default is NA. Desired number of clusters. If default, k is estimated automatically.
clust	String. Identify clusters column-wise if set to "col" (default), or row-wise if set to "row".
method	String. Default is "gap". If k is NA, the number of clusters that automatically be estimated by one of two algorithms: gap (default) (Tibshirani et al. (2001), J Roy Statist Soc B, 63: 411-423) or PART (Nilsen et al (2012), Stat Appl Genet Mol Biol, 12: 637-652).
B	Integer. Number of permutations. Used if method set to 'PART'.
min.size	Integer. Least number of element in a single cluster. Used if method set to 'PART'.
max.level	Integer. The level in the dendogram counting from the root to stop searching for clusters.

cbsteen/clustermap documentation built on May 24, 2019, 1:35 a.m.