Computational Details
General Formulas
Canonical discriminant analysis is equivalent to canonical correlation
analysis between the quantitative variables and a set of dummy variables
coded from the class variable. In the following notation the dummy
variables will be denoted by y and the quantitative variables by x.
The total sample covariance matrix for the x and y variables is
When c is the number of groups, nt is the number of observations in
group t, and St is the sample covariance matrix for the x variables
in group t, the within-class pooled covariance matrix for the x
variables is
The canonical correlations, , are the square roots of the
eigenvalues, , of the following matrix. The corresponding
eigenvectors are vi.
-
Sp-1/2SxySyy-1SyxSp-1/2
Let V be the matrix with the eigenvectors vi that correspond to
nonzero eigenvalues as columns. The raw canonical coefficients are
calculated as follows
-
R = Sp-1/2V
The pooled within-class standardized canonical coefficients are
-
P = diag(Sp)1/2R
And the total sample standardized canonical coefficients are
-
T = diag(Sxx)1/2R
Let Xc be the matrix with the centered x variables as columns. The
canonical scores may be calculated by any of the following
-
Xc R
-
Xc diag(Sp)-1/2P
-
Xc diag(Sxx)-1/2T
For the Multivariate tests based on E-1H
-
E = (n-1)(Syy - SyxSxx-1Sxy)
-
H = (n-1)SyxSxx-1Sxy
where n is the total number of observations.
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.