Re: non-linear fitting of data
- To: mathgroup at yoda.physics.unc.edu
- Subject: Re: non-linear fitting of data
- From: sfeir at mines.u-nancy.fr
- Date: Tue, 2 Jun 92 19:13:45 +0200
>
> Univ of South Carolina, 29-MAY-1992
>
> The Fit function as described in the MMA book only does linear fitting,
>meaning that it fits only to a linear combination of user-supplied
>functions. I want to do nonlinear fitting, meaning that I want to give MMA
>a function of x and also a function of adjustable parameters, and have MMA
>find the best values for those parameters, not just some multiplicative
>factor for the function. Does anyone know how I can do that with Fit, or
>if there is a nice package I can use?
>
>I tried the following:
>
>Fit[ wire5k, {1/(c (1 + ((x-d)/c)^2))}, {x,c,d} ]
>
>The idea is that I want x,c and d to all vary. The red error msg I got
>was:
>
>Fit::arg: Fit has received {x, c, d} as argument.
>
> Scott
>
>
Yes!
You should look-up Mathematica Journal Vol2-Number 2 : NONLINEAR FITTING
LOOPING & RECURSION BY NANCY BLACHMAN. Its a nice package and it does
exactly what you want.
Good luck
A.A. Sfeir