[Fwd: RE: Rank using JAMA]

[Ronald Boisvert <boisvert@nist.gov>]

Sender: Cleve Moler <Cleve.Moler@mathworks.com>
Subject: RE: Rank using JAMA


Hello --

   MATLAB and Jama are using the same algorithm and tolerance to compute numerical rank, so they should give the same value.  How large is your matrix and what values do MATLAB and Jama give for rank?

   Matrix rank is highly sensitive to scaling and errors in the matrix and to the tolerance used in the rank calculation, so it is usually not an appropriate quantity for practical computation.  Why are you try to compute rank?

  -- Cleve Moler

-----Original Message-----
From: jama@nist.gov [mailto:jama@nist.gov] On Behalf Of Ronald Boisvert
Sent: Thursday, June 25, 2009 9:29 AM
To: Multiple recipients of list
Subject: [Fwd: Rank using JAMA]




-------- Original Message --------
Subject: 	Rank using JAMA
Date: 	Thu, 25 Jun 2009 09:23:39 -0400
From: 	siddharth singh <sid027@gmail.com>
Reply-To: 	sid027@gmail.com <sid027@gmail.com>
To: 	Boisvert, Ronald F. <ronald.boisvert@nist.gov>



Hi,
I'm doing a project on Tensegrity Designing and used JAMA package for 
the matrix analysis part.I've come across certain problems.While 
calculating rank MATLAB gives a different result for the same matrix and 
JAMA does calculate rank but gives a very big value.It seems that it is 
not sensetive to values in the matrix.If it is not then modification 
needs to be done.If I'm wrong please tell me what should I do.
e.g including a value like 0.000001 in the matrix would result in an 
erroneous result in JAMA.Please tell me the solution as soon as possible!

Regards,
Siddharth