Principal component analysis biplots - Understanding Biplots - page 115

Character string to provide a title for a one- or
two-dimensional biplot. Default is "" .

Title

Character string to provide a title for a three-dimensional
biplot. Defaults to c(" ", " ", "x", "y", "z") .

Titles.3d

Logical value requesting or suppressing Tukey medians to be
added to a biplot. Default

Tukey.median

= FALSE.

Optional arguments passed to the plot function controlling the
appearance of the biplot.

...

Value

Matrix with each row containing the details of the point to be
plotted.

Z

A list with each component a matrix containing the details of a
biplot axis to be plotted.

Z.axes

Matrix consisting of the eigenvectors as columns.

V

Vectors used for the biplot scaffolding.

Eigenvectors

All the eigenvalues.

e.vals

The overall quality of the display.

PCA.quality

Adequacies associated with the variables.

Adequacy

The axis predictivities.

Predictivity

Predictions of the means or of the samples as specified.

Predictions

Scaffolding coordinates of the origin. origin.pos

3.7.2 Function PCAbipl.zoom

This is a function for zooming into an interactively selected area of a one- or two-
dimensional PCA biplot. In addition to the arguments of PCAbipl , this function has the
argument zoomval for controlling the amount of zooming. Specify zoomval = NULL
for no zooming, a value less than unity for zooming in and a value larger than unity for
zooming out. When PCAbipl.zoom is called, a graph window with the original PCA
biplot is opened. The cursor, when moved over the biplot, changes to a cross. Position
the cross at the bottom left-hand corner of that part of the biplot in need of zooming and
select. A second graph window is opened with the zoomed biplot.

3.7.3 Function PCAbipl.density

This is a function for constructing two-dimensional PCA biplots on top of a density plot
of a two-dimensional PCA approximation of the input matrix. The R function kde2d
described by Venables and Ripley (2002) and available in package MASS is used to
perform two-dimensional kernel density estimation with an axis-aligned bivariate normal
kernel, evaluated on a square grid. PCAbipl.density has the following arguments not
found in PCAbipl :