[Master Index] [Index for export/lyngby]

lyngby_cva

(export/lyngby/lyngby_cva.m)


Function Synopsis

[U,S,V] = lyngby_cva(X, Y, arg1, arg2, arg3, arg4, arg5, ...

Help text

 lyngby_cva           - Canonical variate analysis (Canonical Ridge)

       function [U,S,V] = lyngby_cva(X, Y, ...
           'Propertyname', 'PropertyValue');

	Input:	X   First datamatrix
               Y   Second datamatrix
               Property:
                  'RidgeX'     { 0 } Canonical ridge parameter for X
                  'RidgeY'     { 0 } Canonical ridge parameter for Y
                  'Components' { 5 } Max number of canonical
                               components returned
                  'SubsetComp' { size(X,2) } Number of subset
                               component, ie, SVD components piped to
                               CVA 
                  'NormRidgeX' [ {1} (true) | 0 ] Normalization of the
                               ridge parameter with the trace of X's
                               covariance: trace(X'X)/length(X'X)
                  'NormRidgeY' [ {1} (true) | 0 ] Normalization of the
                               ridge parameter with the trace of Y's
                               covariance: trace(Y'Y)/length(Y'Y)
                  'InitSVD'    [ {1} (true) | 0 ] Perform initial SVD
                               on a rank-deficient X matrix. 

       Output: U   Canonical sequence (canonical correlation vectors -
                   sequence)
               S   Canonical values (squared canonical correlation
                   coefficient)
               V   Canonical image (canonical correlation vectors -
                   images)

       Performs canonical variate analysis, or rather canonic
       correlation analysis, on full ranked or rank-deficient
       matrices. Ridge terms can be applied forming canonical ridge
       analysis.
       If 'InitSVD' is true a initial SVD is performed on the first
       matrix (X) if the matrix is rank-deficient and then the ridge
       parameter is only applied in the SVD subspace of X. 

       Only canonic values larger than max(size(K)) * norm(K) * eps
       will be returned.

       Ref: Mardia, Multivariate Analysis

       See also: lyngby_svd, lyngby_opls

Cross-Reference Information

This function calls
This function is called by

Produced by mat2html on Wed Jul 29 15:43:40 2009
Cross-Directory links are: OFF