Add a static Coefficent Wave to a fit function

Share icon

Hallo,

I would like to add a static coefficient wave to a user defined fit function, so the coefficient values would be prepopulated when selecting the function the fit dialog. I read already that a dynamically created coefficient wave is not possible. But maybe a static one?

Regards, Lars
You can always make a coefficient wave, fill it with initial guesses, and then select it on the Coefficients tab of the Curve Fit dialog. I'm not sure what you mean by "static" and "dynamic".

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

August 28, 2012 at 12:05 pm - Permalink

Hello,

sorry for beeing not clear. With dynamic I was referring to a function calculating the coefficient start values from within the fit function it self. This had been already discussed here (I yust don't find the link right now).

I know that I could create a wave file containing the coefficients before entering the fit dialog and select it there. I was wondering if it was possible to define the start coefficient wave in the fit function it self so I would reduce the working steps when fitting via the dialog.

Regards, Lars

August 29, 2012 at 08:44 am - Permalink

No, not yet. One of my dreams is to figure out a nice syntax for specifying an "autoguess" function that could be called by FuncFit to provide automatic initial guesses. Development of such a function can be pretty tricky, though.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

August 29, 2012 at 09:30 am - Permalink

That would be really awesome (and badly needed ;) ). An other nice thing would be function that is called after the fit finished to do additional adjustment. In the simplest case something like setting the style of the fit result text box.

Regards, Lars

August 30, 2012 at 02:14 am - Permalink

You can, of course, choose what's included in the text box before the fit. But if you mean that you want to post-process the textbox in order to change text styles and such, that would require a post-processing step.

A post-processing function would be nice, too. It would make it easy for folks to transform things like the width of a Gaussian to FWHM.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

August 30, 2012 at 09:04 am - Permalink