- To: mathgroup at smc.vnet.net
- Subject: [mg48022] Matrix Inversion
- From: Gregor <athlonicNOSPAM at yahoo.com>
- Date: Fri, 7 May 2004 04:29:40 -0400 (EDT)
- Sender: owner-wri-mathgroup at wolfram.com
I try to invert a badly conditioned 8x8 matrix trough the normal
I receive the bad conditioning warning message, telling me that
significant numerical errors could be made. Indeed this is the case as
M.M^-1 is not an Identity matrix at all.
When I use PseudoInverse[M] I don't get the warning (that's normal), but
I don't get a correctly inverted matrix. M.M^-1 is also not identity at all.
I did chop the resulst of M.M^-1 but I have very big numbers outside of
the diagonal in the first case and numbers of the order of 10^-2 in the
The matrix is correctly inverted in another system, but I would like to do
everything in Mathematica as my whole program is in Mathematica.
What should I do?
Prev by Date:
Re: Working with binaries
Next by Date:
simultaneous fit of multiple data sets
Previous by thread:
RE: Building List
Next by thread:
Re: Matrix Inversion