plot.clusterGroupBound {hdi}

Plot output of hierarchical testing of groups of variables

Description

The plot() method for "clusterGroupBound" objects plots the outcome of applying a lower bound on the l1-norm on groups of variables in a hierarchical clustering tree.

Usage

## S3 method for class 'clusterGroupBound'
plot(x, cexfactor = 1, yaxis = "members",
     xlab = "", col = NULL, pch = 20, ...)

Arguments

x

an object of class "clusterGroupBound", as resulting from clusterGroupBound().

cexfactor

numeric expansion factor for the size of the node symbols.

yaxis

a string; for the default "members", the hierarchical tree is shown as function of cluster size on the y-axis, whereas the node sizes are proportional to the lower l1-norm of the respective groups of variables. If yaxis takes any different value, then this is reversed and the tree is shown against the lower l1-norm on the y-axis, while node sizes are now proportional to the number of elements in each cluster.

xlab

label used for the x-axis; by default none.

col

the colour of the symbols for the nodes.

pch

the plot symbol (see points) of the symbols for the nodes.

...

optional additional arguments passed to plot.default.

Value

Nothing is returned

Author(s)

Nicolai Meinshausen [e-mail]

See Also

Use clusterGroupBound() to test all groups in a hierarchical clustering tree. Use groupBound() to compute the lower bound for selected groups of variables.

Examples

## Create a regression problem with correlated design (n = 10, p = 3):
## a block of size 2 and a block of size 1, within-block correlation is 0.99

set.seed(29)
p   <- 3
n   <- 10

Sigma <- diag(p)
Sigma[1,2] <- Sigma[2,1] <- 0.99

x <- matrix(rnorm(n * p), nrow = n) %*% chol(Sigma)

## Create response with active variable 1
beta    <- rep(0, p)
beta[1] <- 5

y  <- as.numeric(x %*% beta + rnorm(n))

## Compute the lower bound for all groups in a hierarchical clustering tree
cgb5 <- clusterGroupBound(x, y, nsplit = 4) ## use larger value for nsplit!

## Plot the tree with y-axis proportional to the (log) of the number of
## group members and node sizes proportional to the lower l1-norm bound.
plot(cgb5)

## Show the lower bound on the y-axis and node sizes proportional to
## number of group members
plot(cgb5, yaxis = "")