[SciPy-User] curve_fit - fitting a sum of 'functions'

[Yury V. Zaytsev]

On Wed, 2011-05-25 at 07:36 +0000, Meesters, Christian wrote: 
> My problem now is with the 'guess' parameter: If handing over
> parameters as estimated by eye, all works fine, but more general
> approaches failed so far. 

I don't know if it would seem to make any sense to you, but you could
heavily filter your function with something like Savitzky–Golay
smoothing filter (it's basically just a weighted sum, so no numerical
heavy-lifting is involved) and use the filtered version to derive your
guesses, if the noise is really the problem...

-- 
Sincerely yours,
Yury V. Zaytsev