Home »
Public Forums »
archive »
Re: Curve Fits and Partial Derivatives
| Re: Curve Fits and Partial Derivatives [message #6262] |
Tue, 28 May 1996 00:00 |
rivers
Messages: 228
Registered: March 1991
|
Senior Member |
|
|
In article <4oefja$qbl@sinfo.ll.iac.es>, bsams@ll.iac.es (Bruce Sams) writes:
> Hello,
>
> I am a new user of IDL, and I am having some less than satisfying
> results with the CURVEFIT routine. I want to fit a user
> defined function to some 1-d data. The function is simple: a power-law.
> I am able to get CURVEFIT to run, bu the initial guesses I give to
> the program have to be SOOOOO close to the final "fitted" values
> (within about 5% or so) that the routine is effectively useless.
> A simple program (Kaleidagraph) on my Macintosh can fit practically
> *any* curve with a power law and NO initial guesses. So what's up?
> Does anyone have a .pro file they could e-mail me where the
> call to fit with CURVEFIT and the user defined function are
> shown?
An easy mistake to make with CURVEFIT is to get the analytical derivatives
wrong. If you are using derivatives calculated in your procedure, try leting
CURVEFIT do the derivatives numerically.
____________________________________________________________
Mark Rivers (312) 702-2279 (office)
CARS (312) 702-9951 (secretary)
Univ. of Chicago (312) 702-5454 (FAX)
5640 S. Ellis Ave. (708) 922-0499 (home)
Chicago, IL 60637 rivers@cars3.uchicago.edu (Internet)
|
|
|
|
Current Time: Wed Oct 08 17:43:50 PDT 2025
Total time taken to generate the page: 0.00279 seconds
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
