| SVDC & PCOMP [message #17707] |
Tue, 09 November 1999 00:00 |
Samuel Djavidnia
Messages: 1
Registered: November 1999
|
Junior Member |
|
|
Hi,
I am trying to use the SVDC (and the PCOMP) function to calculate the
Empirical Orthogonal Functions and the Principal Components of a data
set.
The data set is in the following format:
D=fltarr(15,2688)
When I use the SVDC function I end up with 3 matrices:
X=fltarr(15) Eigenvalues (Variances)
Y=(15,2688) Eigenvectors (Empirical Orthogonal Functions)
Z=(15,15) Principal Components
The problem is that the values of X which should be monotonically
decreasing are not. X represents the variance (X/tot(X), but it seems
that when using a small data set X does not decrease monotonically as it
should.
Does anybody know if there is a bug in the SVDC function ?
Also, if I transpose my dataset to DT=fltarr(2688,15) I get completely
different eigenvalues. Why ?
If I instead use the PCOMP function, I get totally different values of
the variance, eigenvectors and principal components (in this case the
variance correctly decreases monotonically).
How can I check if PCOMP is computing the correct calculations ?
Thanks in advance.
Samy
*******************************************************
Samy Djavidnia
Research Assistant
Space & Atmospheric Physics
The Blackett Laboratory
Imperial College of Science, Technology and Medicine
London SW7 2BZ
Tel: (0044)-171-594 7675 E-mail: s.djavidnia@ic.ac.uk
Fax: (0044)-171-594 7900 http://www.sp.ph.ic.ac.uk/~samy
*******************************************************
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
