polar.clust: Convert a hclust object to a polar dendrogram

In backlin/polarClust: Polar representation of clustering trees

Description

Any hclust object can be converted to a polar dendrogram using this function and then plotted with the standard graphics functions.

Usage

polar.clust(x, labels = sprintf("#%i", 1:length(x$order)), col,
  padding = 0.01, hang = 0.1 * max(x$height))

Arguments

x	hclust object.
labels	Chraracter vector of leaf labels.
col	Vector of leaf colours.
padding	Spacing between branches.
hang	Branch lengths of the leaves.

Value

An object of class polar.clust

Author(s)

Christofer Bäcklin

See Also

plot.polar.clust

Examples

my.data <- iris[-5] + .5*matrix(rnorm(nrow(iris)*(ncol(iris)-1)), nrow(iris))
cl <- hclust(dist(my.data))
x <- polar.clust(cl)

plot(x)

# Color according to class
labels(x) <- sprintf("%s %i", iris$Species, 1:150)
colors(x) <- iris$Species
plot(x, expand=1.2, label.size=.3, las=2)

# Color according to branch
colors(x) <- c(brewer.pal(6, "Set1"))
labels(x) <- NULL # Clear the current labels
labels(x) <- paste("Branch", 1:6)
plot(x, las=1)

# Manual compilation of the plot
s <- c(.4,2.2)*pi
colors(x) <- NULL
plot(x, type="n", labels=FALSE, axes=FALSE, expand=1.1)
branches(x, k=6, sector=s, fill=TRUE, lwd=10, lightness=.3)
lines(x, sector=s)
points(x, pch=19, sector=s)
legend("topright", paste("Branch", 1:6), fill=brewer.pal(6, "Set1"), bg="white")

backlin/polarClust documentation built on May 11, 2019, 5:23 p.m.