Information Technology Reference
In-Depth Information
The actual plotting of the CA biplot is done by either the function
drawbipl.ca
or the function
drawbipl.3dim.ca
. These functions are called by
cabipl
and are
usually not called by a user. Two functions that can be called by a user, are the functions
ca.predictivities
(Section 7.5.2) and
ca.predictions.mat
(Section 7.5.3).
7.5.2 Function
ca.predictivities
The function
ca.predictivities
is for calculating overall quality, axis predictivities
and sample predictivities associated with a CA.
Arguments
Contingency table in the form of a
p
×
q
matrix.
data
Integer-valued vector specifying the output. Default is that entire output
list is printed.
out
Value
The output of
ca.predictivities
is an R list with the following components:
The overall quality of the biplot display.
Quality
The weights for calculating the quality from the axis
predictivities.
Weights
Adequacies associated with the individual biplot axes.
Adequacies
Axis predictivities associated with the individual biplot
axes.
Axis.predictivities
Sample predictivities associated with the rows of the
input matrix.
Sample.predictivities
Predicted matrix (7.7).
X.hat
7.5.3 Function
ca.predictions.mat
The function
ca.predictions.mat
is an alternative function for calculating row
profile predictions of the input matrix. It gives identical predictions to function
cabipl
with arguments
ca.variant = "RowProfA"
or
"RowProfB"
and
Row-
Prof.scaled.markers = TRUE
, but provides also for predictions in more than three
dimensions.
Arguments
p
×
q
matrix with positive elements. Preferably a correspondence
matrix, i.e. all elements of
Pmat
are nonnegative and sum to 1.
Pmat
Integer-valued vector specifying the eigenvectors used in calculating
the predictions.
e.vects
