cba (version 0.2-25)

clmplot: Plotting Logical Matrices

Description

A wrapper function to image that produces a level plot with the option to color the rows (or columns) based on a clustering and/or classification of the data, and the option to reorder the rows and columns for better presentation.

Usage

clmplot(x, col, col.bycol = FALSE, order=FALSE,
        dist.method = "binary", hclust.method = "average",
        axes = FALSE, xlab = "", ylab = "", ...)

Value

A list with the following components:

rows

the row order.

cols

the column order.

Arguments

x

an logical matrix.

col

an optional vector defining a coloring.

col.bycol

option to color by columns.

order

option to (pre)order the rows and columns.

dist.method

method to be used by dist.

hclust.method

method to be used by hclust.

axes

option to plot axes.

xlab, ylab

labels for the x and y axis.

...

further arguments to image.

Author

Christian Buchta

Details

For dummy coded data the level FALSE is assumed to carry no information and is colored white. Thus, the level TRUE can be colored according to some classification of the rows (or columns) of the data matrix. If no color specification is supplied the default color black is used.

If col is of type character it is assumed to contain color codes. Otherwise, it must be a factor and the levels are assigned colors from heat.colors.

If order is TRUE the rows and columns are ordered by hclust where the distances are computed by dist.

Note that an axis is only plotted if the number of elements (rows or columns) is less than 100.

See Also

lmplot for plotting of logical data at reduced resolutions, heatmap for ordered plotting of real-valued data, and package gclus for ordering functions.

Examples

x <- matrix(sample(c(FALSE,TRUE),100,rep=TRUE),ncol=10)
clmplot(x, order=TRUE, axes=TRUE)
clmplot(x, col=rep(c(1,2),each=5))
clmplot(x, col=rep(c("red","blue"),each=5))
clmplot(x, col=rep(c("red","blue"),each=5), col.bycol=TRUE)
if (FALSE) {
example("rockCluster")
### continue example (see rockCluster)
col <- Votes$Class                                  # color by party
levels(col) <- c("red","blue")
op <- par(mfrow=c(1,2), pty="s")
clmplot(x, order=TRUE, col=as.character(col), main="Parties")
col <- rf$cl                                        # color by cluster
levels(col) <- c("blue","red","green", "black")     # map NA to black
clmplot(x, order=TRUE, col=as.character(col), main="Clusters")
par(op)
}