This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: [off-topic?] least square problem solution ...
- From: Brian Gough <bjg at network-theory dot co dot uk>
- To: Jose Miguel Buenaposada Biencinto <jmbuena at dia dot fi dot upm dot es>
- Cc: "gsl-discuss at sources dot redhat dot com" <gsl-discuss at sources dot redhat dot com>
- Date: Tue, 22 Jan 2002 19:44:55 +0000 (GMT)
- Subject: Re: [off-topic?] least square problem solution ...
- References: <3C4C68D2.BBB2BDF3@dia.fi.upm.es>
Jose Miguel Buenaposada Biencinto writes:
> I think this is a little bit off topic but I don't know any other
> list to ask for help ...
> I have the solution of a least square problem: a =
> inverse(transpose(M)*M)*transpose(M)*b where a is kx1, M is Nxk and
> b is Nx1. I know (due to the nature of the problem I have) that
> some of the rows in M are not very significant and I'd like to be
> able to choose the best set of m rows of M to solve the system and
> to compute the error you make by doing so. Is there a direct
> procedure to do so?
I don't know, but if it can be done I'm sure it involves the SVD.
Brian