MathGroup Archive 2008

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

Search the Archive

Re: Trying hard to find ' Findfit ' of Mathematics in C++

  • To: mathgroup at smc.vnet.net
  • Subject: [mg84887] Re: Trying hard to find ' Findfit ' of Mathematics in C++
  • From: nano bagonghi <nano.bagonghiCUT at CUTgmail.com>
  • Date: Fri, 18 Jan 2008 05:44:45 -0500 (EST)
  • References: <fmmkv3$f57$1@smc.vnet.net>

Mayneord wrote:

> Well, I knew how to work with ' Findfit ' of Mathematica but i would
> like to know how to do the same in C++.  Also I came to know that
> there is no direct function in C++  similar to 'Findfit' .. Please is
> there somebody who can suggest me how to approach this problem in C++
> or is there already a third party source code available ?

It is quite obvious that there cannot be such a function directly
in C++ (or in any other of the main programming languages): it
is a too hi-level function and can be implemented in hundreds of
ways.

Mathematica's FindFit has a lot of options (for example, different
norms can be used in the process of error minimization),
maybe you have a more restricted problem in mind.

Anyway, the basic algorithm used by FindFit is
a least squares fitting.
You can found some information here
http://mathworld.wolfram.com/LeastSquaresFitting.html
or in other thousands of web pages
(and of course in any book about numerical analysis).

If you search with google, you'll find surely a lot
of c++ code to perform least squares fitting, however,
in order to use it properly and obtain meaningful results,
I suggest that you must at least understand how it works.

bye,
g.


  • Prev by Date: Re: Simplifying integer expressions with Mod and BitAnd
  • Next by Date: Re: What is the algorithm Mathematica uses to detrmine the automatic
  • Previous by thread: Trying hard to find ' Findfit ' of Mathematics in C++
  • Next by thread: Simplify/FullSimplify Infinite Loop