StatDA (version 1.7.11)

boxes: Boxes

Description

The function boxes computes boxes of multivariate data. If add=TRUE the boxes are plotted in the current plot otherwise nothing is plotted.

Usage

boxes(x, xA = 1, yA = 2, zA = 3, labels = dimnames(x)[[1]], locations = NULL,
nrow = NULL, ncol = NULL, key.loc = NULL, key.labels = dimnames(x)[[2]],
key.xpd = TRUE, xlim = NULL, ylim = NULL, flip.labels = NULL, len = 1,
leglen = 1, axes = FALSE, frame.plot = axes, main = NULL, sub = NULL,
xlab = "", ylab = "", cex = 0.8, lwd = 0.25, lty = par("lty"), xpd = FALSE,
mar = pmin(par("mar"), 1.1 + c(2 * axes + (xlab != ""), 2 * axes + (ylab != ""),
 1, 0)), add = FALSE, plot = TRUE, ...)

Value

No return value, creates a plot.

Arguments

x

multivariate data in form of matrix or data frame

xA

assignment of clusters to the coordinates of the boxes

yA

assignment of clusters to the coordinates of the boxes

zA

assignment of clusters to the coordinates of the boxes

labels

vector of character strings for labeling the plots

locations

locations for the boxes on the plot (e.g. X/Y coordinates)

nrow

integers giving the number of rows ands columns to use when 'locations' is 'NULL'. By default, 'nrow == ncol', a square will be used.

ncol

integers giving the number of rows and columns to use when 'locations' is 'NULL'. By default, 'nrow == ncol', a square will be used.

key.loc

vector with x and y coordinates of the unit key.

key.labels

vector of character strings for labeling the segments of the unit key. If omitted, the second component of 'dimnames(x)' ist used, if available.

key.xpd

clipping switch for the unit key (drawing and labeling), see 'par("xpd")'.

xlim

vector with the range of x coordinates to plot

ylim

vector with the range of y coordinates to plot

flip.labels

logical indicating if the label locations should flip up and down from diagram to diagram. Defaults to a somewhat smart heuristic.

len

multiplicative values for the space used in the plot window

leglen

multiplicative values for the space of the labels on the legend

axes

logical flag: if 'TRUE' axes are added to the plot

frame.plot

logical flag: if 'TRUE', the plot region ist framed

main

a main title for the plot

sub

a sub title for the plot

xlab

a label for the x axis

ylab

a label for the y axis

cex

character expansion factor for the labels

lwd

line width used for drawing

lty

line type used for drawing

xpd

logical or NA indicationg if clipping should be done, see 'par(xpd=.)'

mar

argument to 'par(mar=*)', rypically choosing smaller margings than by default

add

logical, if 'TRUE' add boxes to current plot

plot

logical, if 'FALSE', nothing is plotted

...

further arguments, passed to the first call of 'plot()'

Author

Peter Filzmoser <> http://cstat.tuwien.ac.at/filz/

Details

This type of graphical approach for multivariate data is only applicable where the data can be grouped into three clusters. This means that before the plot can be made the data undergo a hierarchical cluster to get the size of each cluster. The distance measure for the hierarchicla cluster is complete linkage. Each cluster represents one side of the boxes.

References

C. Reimann, P. Filzmoser, R.G. Garrett, and R. Dutter: Statistical Data Analysis Explained. Applied Environmental Statistics with R. John Wiley and Sons, Chichester, 2008.

See Also

plot.default,box

Examples

#plots the background and the boxes for the elements
data(ohorizon)
X=ohorizon[,"XCOO"]
Y=ohorizon[,"YCOO"]
el=log10(ohorizon[,c("Co","Cu","Ni","Rb","Bi","Na","Sr")])
data(kola.background)

sel <- c(3,8,22, 29, 32, 35, 43, 69, 73 ,93,109,129,130,134,168,181,183,205,211,
      218,237,242,276,292,297,298,345,346,352,372,373,386,408,419,427,441,446,490,
      516,535,551,556,558,564,577,584,601,612,617)

x=el[sel,]
xwid=diff(range(X))/12e4
ywid=diff(range(Y))/12e4
plot(X,Y,frame.plot=FALSE,xaxt="n",yaxt="n",xlab="",ylab="",type="n",
   xlim=c(360000,max(X)))
plotbg(map.col=c("gray","gray","gray","gray"),add.plot=TRUE)

boxes(x,locations=cbind(X[sel],Y[sel]),len=20000,key.loc=c(800000,7830000),leglen=25000,
     cex=0.75, add=TRUE, labels=NULL, lwd=1.1)