Re: non-linear fitting data to a theoretical curve

[Muthiah Annamalai]

On 6/22/07, address@hidden <address@hidden> wrote:
> Hello,
> i am searching for a fitting command in octave to fit data points to
> a theoretical curve, which is an exponential decay with a bit
> complicated exponent. How can one do that with octave?
> Thanks for any help in advance ...

If you want to go for a non-linear fit, use 'leasqr'.
If you are looking for a polynomial type curve-fit use 'polyfit'
See their help functions at Octave command line.

You could also use polyfit() function if you take a logarithm of
the curve so the exponential becomes a polynomial and then
fit it. With the coefficients, convert it back into the place in
the exponential.

With a little bit of looking in the archives
http://dir.gmane.org/gmane.comp.gnu.octave.general you can
get examples on both.

-Muthu