Mathematica 9 is now available
Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2007
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 2007

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

Search the Archive

Re: Curve-fitting/data analysis question...

  • To: mathgroup at smc.vnet.net
  • Subject: [mg73351] Re: Curve-fitting/data analysis question...
  • From: "Robert Dodier" <robert.dodier at gmail.com>
  • Date: Tue, 13 Feb 2007 07:01:23 -0500 (EST)
  • References: <eqmcfh$bdv$1@smc.vnet.net>

On Feb 10, 11:18 pm, "sherifffruitfly" <sherifffruit... at gmail.com>
wrote:

> I've got some data that is approximately constant, c, over an
> interval, and then drops off like -(a exp(b x)).
>
> How can I determine when the constant part ends, and the negative
> exponential begins in Mathematica? I can eyeball it on a graph, but I'd much
> prefer to have a principled Mathematica routine settle the question. Once
> that's settled, it should be straightforward to determine the
> exponential's parameters via some curve-fitting routine or other.

See Seber & Wild, "Nonlinear Regression", Wiley, 1989 (iirc).
They have a chapter on problems of this kind.
I'm sure there are more recent refs, but I don't know any.

I suspect that an automated procedure cannot do much better than
you can do it by eye. If, however, you need to solve a collection of
such problems and you need to automate the process, then an
automated method can save you some time (without affecting the
results much).

FWIW
Robert Dodier



  • Prev by Date: Re: a function containg MeijerG (limit behavior)
  • Next by Date: Re: RandomArray from user defined distribution?
  • Previous by thread: Re: Curve-fitting/data analysis question...
  • Next by thread: free boundary condition