biplot3D {compositions}R Documentation

Three-dimensional biplots, based on package rgl

Description

Plots variables and cases in the same plot, based on a principal component analysis.

Usage

biplot3D(x,...)
## Default S3 method:
biplot3D(x,y,var.axes=TRUE,col=c("green","red"),cex=c(2,2),
            xlabs = NULL, ylabs = NULL, expand = 1,arrow.len = 0.1,
            ...,add=FALSE)
 ## S3 method for class 'princomp':
 biplot3D(x,choices=1:3,scale=1,...,
            comp.col=par("fg"),comp.labs=paste("Comp.",1:3),
            scale.scores=lambda[choices]^(1-scale),
            scale.var=scale.comp, scale.comp=sqrt(lambda[choices]), 
            scale.disp=1/scale.comp)

Arguments

x princomp object or matrix of point locations to be drawn (typically, cases)
choices Which principal components should be used?
scale a scaling parameter like in biplot
scale.scores a vector giving the scaling applied to the scores
scale.var a vector giving the scaling applied to the variables
scale.comp a vector giving the scaling applied to the unit length of each component
scale.disp a vector giving the scaling of the display in the directions of the components
comp.col color to draw the axes of the components
comp.labs labels for the components
... further plotting parameters as defined in rgl.material
y matrix of second point/arrow-head locations (typically, variables)
var.axes logical, TRUE draws arrows and FALSE points for y
col vector/list of two elements the first giving the color/colors for the first data set and the second giving color/colors for the second data set.
cex vector/list of two elements the first giving the size for the first data set and the second giving size for the second data set.
xlabs labels to be plotted at x-locations
ylabs labels to be plotted at y-locations
expand the relative expansion of the y data set with respect to x
arrow.len The length of the arrows as defined in arrows3D
add logical, adding to existing plot or making a new one?

Details

This "biplot" is a triplot, relating data, variables and principal components. The relative scaling of the components is still experimental, meant to mimic the behavior of classical biplots.

Value

nothing

Author(s)

K.Gerald v.d. Boogaart http://www.stat.boogaart.de

See Also

gsi

Examples

data(SimulatedAmounts)
pc <- princomp(acomp(sa.lognormals5))
pc
summary(pc)
plot(pc)           #plot(pc,type="screeplot")
biplot3D(pc)

[Package compositions version 1.01-1 Index]