| Title: | Mathematics at DEC |
| Moderator: | RUSURE::EDP |
| Created: | Mon Feb 03 1986 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2083 |
| Total number of notes: | 14613 |
Hello. I am working on a curve fitting program in C from a magazine
article. The problem is the article left out the algorithms of the
three routines that solve overdetermined systems of equations. These
three norms are the L1-norm, the least squares norm, and the minimax
norm. Does anyone know what books I can find these algorithms in?
Marie
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1255.1 | Some algorithms available from ACM | ALLVAX::JROTH | It's a bush recording... | Thu Jun 21 1990 22:04 | 25 |
One general reference is "Matrix Computations" by Golub and Van Loan,
published by Johns Hopkins Press. For least squares, see "Solving
Least Squares Problems" by Lawson and Hanson, which has FORTRAN
routine listings.
Algorithms for these are available on-line (in FORTRAN...) from
netlib - in particular see ACM algorithms 544 (weighted least squares
solutions), 551/552 for L1 solutions of over determined equations
and 495 for the Chebychev (L_infinity or minimax) norm.
You can obtain these by sending a message to netlib (at Oak Ridge
National Labs) of the following form:
to: DECWRL::"[email protected]"
subj: send 544 from toms
one line message: send 544 from toms
You can also ask for "send index from toms" or just "send index" to find
the other goodies.
Hope this helps... you'll find that the least squares case is by far
the easiest to implement and I'd recommend trying that first.
- Jim
| |||||
| 1255.2 | Thank you | OLDTMR::TUCKER | Wed Jun 27 1990 12:48 | 6 | |
Thanks Jim, what you told me was very helpful.
Marjorie Marie
| |||||