Plot Trees From Hierarchical Clustering

DESCRIPTION:

USAGE:

plclust(tree, hang = .1, unit = F, level = F, hmin = 0,  
        square = T, labels = <<see below>>, plot = T) 

REQUIRED ARGUMENTS:

tree

a hierarchical clustering tree, of the form returned by function hclust.

OPTIONAL ARGUMENTS:

hang

the fraction of the height of the plot that each individual node will hang below the cluster that it joins. A value of --1 will cause all individuals to start at y-value 0.

unit

logical flag: if TRUE, the heights of the merges will be ignored and instead merge i will occur at height i. Useful for spreading out the tree to see the sequence of merges.

level

logical flag: if TRUE, the plotted tree will be "leveled", where merges in different subtrees are arbitrarily assigned the same height in order to compress the vertical scale. This is useful with unit=TRUE.

hmin

minimum height at which merges will take place. This can be used to get rid of irrelevant detail at low levels.

square

logical flag: if TRUE, the tree is plotted with "U" shaped branches, if FALSE, it has "V" shaped branches.

labels

vector of character strings giving the labels for the leaves of the tree. If omitted, leaves will be labeled by number. To omit labels entirely, use labels=FALSE.

plot

logical flag: If TRUE, plotting takes place. If FALSE, no plotting is done but a value is returned.

Graphical parameters may also be supplied as arguments to this function (see par ). In addition, the high-level graphics arguments described under plot.default and the arguments to title may be supplied to this function.

VALUE:

if plot is FALSE, a list containing the coordinates of the leaves of the tree and the interior nodes of the tree:

x,y

x and y coordinates of the leaves of the tree, i.e., x[i],y[i] gives the coordinates of the leaf corresponding to the ith individual.

xn,yn

x and y coordinates of the interior nodes of the tree, i.e., xn[i],yn[i] gives the coordinates of the node representing the ith merge.

SIDE EFFECTS:

DETAILS:

Creates a plot of a clustering tree given a structure by the function hclust . Can produce a plot of a clustering tree on the current graphics device, or a list containing the coordinates of the leaves of the tree and the interior nodes of the tree.

SEE ALSO:

EXAMPLES:

# Create sample object using built-in data 
distances <- dist(votes.repub) 
plclust(hclust(distances)) 
# Another example 
hc <- hclust(dist(longley.y)) 
# plot without labels, save structure 
xy <- plclust(hc, label = FALSE) 
# allow user to point at leaf and have it identified 
identify(xy) 
sums <- apply(author.count,1,sum) 
adjusted <- sweep(author.count,1,sums,"/") 
par(mar=c(18,4,4,1)) 
plclust(hclust(dist(adjusted)),label=dimnames(author.count)[[1]]) 
title("Clustering of Books Based on Letter Frequency")