Home > hfrc > matlab > utility > EOF.m



EOF - computes EOF of a matrix.


function [L, EOFs, EC, error, norms] = EOF( U, n, norm, varargin )


 EOF - computes EOF of a matrix.

 Usage: [L, EOFs, EC, error, norms] = EOF( M, num, norm, ... )

 M is the matrix on which to perform the EOF.  num is the number of EOFs to
 return.  If num='all', then all EOFs are returned.  This is the default.

 If norm is true, then all time series are normalized by their standard
 deviation before EOFs are computed.  Default is false.  In this case,
 the fifth output argument will be the standard deviations of each
 column.  Note that the norms are not reapplied to the resulting EOFs,
 so you will need to do this to get back the final data.

 ... are extra arguments to be given to the svds function.  These will
 be ignored in the case that all EOFs are to be returned, in which case
 the svd function is used instead. Use these with care.

 Data is not detrended before handling.  Use the detrend function to fix

 L are the eigenvalues of the covariance matrix ( ie. they are normalized
 by 1/(m-1), where m is the number of rows ).  EC are the expansion
 coefficients (PCs in other terminology) and error is the reconstruction
 error (L2-norm).


     $Id: EOF.m,v 1.4 2004-09-06 02:03:45 dmk Exp $    

 Copyright (C) 2001 David M. Kaplan
 Licence: GPL (Gnu Public License)



This function calls: This function is called by:
Generated on Mon 18-Dec-2006 07:26:46 by m2html © 2003