lyngby_sop_main - Strother OPLS
function [U,S,V] = lyngby_sop_main(X, R,
'PropertyName', 'PropertyValue')
Input: X The datamatrix
R The run structure
Property: Paradigm The paradigm
Components { 5 } Number of canonical components
maintained
Output: U Canonical sequences
S Canonical values
V Canonical images
Orthonormalized partial least square the "Strother way":
grouping is based on the intrarun scanindex. This algorithm is
based on a singular value decomposition of the inner product
between two matrices svd(X'*Y) where Y is constructed as a
design matrix with identicator variables based on the intrarun
scan index. The lyngby_sdesign function is used to construct Y
from R.
The optional 'Paradigm' property is only use to get a more
prober sign on the canonical vectors.
See also LYNGBY, LYNGBY_OPLS, LYNGBY_CVA, LYNGBY_SVD,
LYNGBY_SDESIGN, LYNGBY_UI_MAIN, LYNGBY_UI_SOP_INIT.
$Id: lyngby_sop_main.m,v 1.11 2002/12/17 23:32:00 fnielsen Exp $