[Date Index] [Thread Index] [Author Index]

Re: NonLinearRegress and constraint parameters

• To: mathgroup at smc.vnet.net
• Subject: [mg57200] Re: [mg57148] NonLinearRegress and constraint parameters
• From: Chris Chiasson <chris.chiasson at gmail.com>
• Date: Fri, 20 May 2005 04:44:04 -0400 (EDT)
• References: <200505190708.DAA13073@smc.vnet.net>
• Reply-to: Chris Chiasson <chris.chiasson at gmail.com>
• Sender: owner-wri-mathgroup at wolfram.com

http://forums.wolfram.com/mathgroup/archive/2003/Sep/msg00184.html

On 5/19/05, Frank Küster <frank at kuesterei.ch> wrote:
> Hi,
> 
> I have been using Mathematica for a short while, and one application is
> curve fitting to a set of measured data.  However, I have one problem
> with this.  If I want to describe my data by a function model[x_] with
> parameters a, b, c,.., how can I constrain the parameters to certain
> ranges?
> 
> I came across a case where the function model[x_] is not defined (or
> yields complex values) with certain combinations of parameters a,b,c,..
> However, much more important would be cases where the function is well
> defined (and might fit the data well), but the resulting values for the
> parameters are unphysical.  I'd like NonLinearRegress to keep the
> parameters in the "allowed" region (and rather give a bad or no fit,
> indicating my model is wrong).  How can I accomplish this?
> 
> Thanks in advance, Frank
> --
> Frank Küster
> Inst. f. Biochemie der Univ. Zürich
> Debian Developer
> 
> 


-- 
Chris Chiasson
http://chrischiasson.com/
1 (810) 265-3161

• References:
  • NonLinearRegress and constraint parameters
    • From: frank@kuesterei.ch (Frank Küster)