Mathematica 9 is now available
Services & Resources / Wolfram Forums / MathGroup Archive
-----

MathGroup Archive 2011

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

Search the Archive

Re: curve fitting question

  • To: mathgroup at smc.vnet.net
  • Subject: [mg115440] Re: curve fitting question
  • From: michael partensky <partensky at gmail.com>
  • Date: Tue, 11 Jan 2011 00:34:42 -0500 (EST)

You can Solve it for  y,

Solve[y == a + b (Exp[40*(x + c*y)/d] - 1) - (x + c*y)/e, y]

 to  find y[x]:

y = 1/40 ((40 a)/(1 + c/e) - (40 b)/(1 + c/e) - (
    40 x)/((1 + c/e) e) - (
    d ProductLog[-((
       40 b c e E^((40 a c e)/(d (c + e)) - (40 b c e)/(d (c + e)) + (
         40 x)/d - (40 c x)/(d (c + e))))/(d (c + e)))])/c.
(please double-check if this solution is unique for your range of
parameters)

This y[x] can be used with FindFit, NonlinearModelFit, etc to find the
optimal parameter values.
Best
Michael Partenskii

On Mon, Jan 10, 2011 at 2:34 AM, Michael B. Heaney <mheaney at alum.mit.edu>wrote:

> Hi,
>
> I have data in (x,y) pairs. I have an equation:
>
>  y= a +b (Exp[40*(x + c*y)/d] - 1) - (x +c*y)/e
>
> where a,b,c,d, and e are fitting parameters.
>
> Note that this equationcannot be put in the form y=f(x).
>
> How do I fit this equation to my x,y data?
>
> Thanks,
>
> Michael
>
>
>



  • Prev by Date: Re: Number of zeros finite or infinite?
  • Next by Date: Problems "Import"ing with Mathematica running Unbuntu
  • Previous by thread: Re: curve fitting question
  • Next by thread: Re: curve fitting question