Information Technology Reference
In-Depth Information
Defaults to 0.05. Determines the
x
-coordinates of a grid
overlaying the biplot space.
x.grid
Defaults to 0.05. Determines the
y
-coordinates of a grid
overlaying the biplot space.
y.grid
Defaults to 16. Plotting symbol for labelling a grid point.
plot.symbol
Defaults to 0.05. Size of plotting symbol for labelling a
grid point.
plot.symbol.size
A
k
-vector specifying the colouring of each grid point
according to the canonical mean with the shortest
distance to it, where
k
denotes the number of canonical
means.
colours.pred.regions
4.7.6 Function
CVA.predictivities
This is a function for calculating the various measures of fit discussed in Section 4.6. It has
the five arguments discussed above for
CVAbipl
:
X
,
G
,
weightedCVA
,
X.new.samples
and
X.new.vars
. The various measures of fit are provided for dimensions 1, 2,
...
,
p
,
and the output (value) of
CVA.predictivities
is a list with the following named
components:
OverAllMeans
Means of the original data matrix.
XBar.groups
Group means of the uncentred data matrix.
XcentBar.groups
Group means of the column-centred data
matrix.
canon.group.means
Canonical group means.
CVA.quality.canvar
Overall quality with respect to the
canonical variables (4.16).
CVA.quality.origvar
Overall quality with respect to the original
variables (4.17).
Axis.Adequacies
Adequacies associated with the variables
(biplot axes) defined in (4.18).
Axis.Predictivities
Predictivities associated with the original
variables (biplot axes) defined in (4.19).
Class.Predictivities
The predictivities for each group
calculated according to (4.20).
WithinGroup.axis.Predictivities
The within-group axis predictivity for
each variable (column) of (
I
-
H
)
X
,
calculated according to (4.22).
WithinGroup.Sample.Predictivities
The within-group sample predictivities
calculated according to (4.23).
The within-group sample predictivities for
new samples calculated according to
Section 4.6.1.
new.sample.pred
Axis predictivities for new variables
calculated according to (4.26).
Axis.predictivities.new.vars