help wanted: SVD of sparse matrices

Article View: sci.math.num-analysis

Article #909

Back to sci.math.num-analysis Previous Next

#909

From: Andrew S. L. Bainbridge-Smith
Date: Tue, 15 Dec 1992 23:33

24 lines
1366 bytes

Hi,

I'm trying to solve a problem involving the hessian matrix
(second order partical derivatives) of a least squares type
cost function.  In particular I wish to find the spectral
decomposition of this matrix. Currently I have been using 
some fortran code (linpack/eispack) that finds the singular
value decomposition (SVD).  The problem is that some times
the hessian matrix can be sparse and the fortran code never
returns -- ever-thing just hangs up.  So I'm asking for a 
little bit of your time to email me (I don't read this group
often + I'll be away for a few weeks) any suggestions,
helpful advice, solutions etc. I'm not to clue-y on this
(SVD) stuff, but I'll take any sort of help that's offered.

Thanks heaps in advance,  and I'll post a summary.

Andrew.

----
Andrew Bainbridge-Smith
Dept. Electrical & Electronic Eng.
Uni. of Tasmania (Australia)
email: andrew@tasman.cc.utas.edu.au

Technical Details

Message-ID: <andrew.724462395@tasman>

Path: arc1-ca.pugleaf.net!sparky!uunet!munnari.oz.au!newsroom.utas.edu.au!tasman.cc.utas.edu.au!andrew

Show Full Headers

Complete Article Headers

Newsgroups: sci.math.num-analysis
Path: arc1-ca.pugleaf.net!sparky!uunet!munnari.oz.au!newsroom.utas.edu.au!tasman.cc.utas.edu.au!andrew
From: andrew@tasman.cc.utas.edu.au (Andrew S. L. Bainbridge-Smith)
Subject: help wanted: SVD of sparse matrices
Message-ID: <andrew.724462395@tasman>
Keywords: SVD, sparse matrices
Sender: news@newsroom.utas.edu.au
Organization: University of Tasmania, Australia.
Date: Tue, 15 Dec 1992 23:33:15 GMT
Lines: 24