Ability to do all-at-once fitting from the Curve Fitting dialog

Share icon

Wed, 06/09/2010 - 06:03 pm

It'd be fantastic to be able to run all-at-once fits from Analysis > Curve Fitting... .

Thanks.
Analysis->Curvefitting does to all-at-once fits, if you have an all-at-once function:

Function mystraightline(w, yy, xx):fitfunc
Wave w,yy,xx
yy=xx*w[0] + w[1]
End


However, the Curvefitting dialogue does not handle fits that have more than one independent variable, or structure fits.

June 9, 2010 at 10:56 pm - Permalink

Sorry, I terribly mis-stated the problem. The dialog doesn't seem to support the "_default_" coefficient wave option on the Coefficients tab. I always get "Igor can't tell how many coefficients the fit function requires. Select a wave from the Coefficient Wave menu." It seems like Igor doesn't process the comment info in the "//CurveFitDialog/" format when the function is all-at-once.

You are absolutely right -- the Curve Fitting dialog can run all-at-once fits as long as you select a pre-existing coefficient wave. You will not see the "human-readable" name of each coefficient, though, which is a major bummer when you have a complicated function.

June 9, 2010 at 11:28 pm - Permalink

nplumb wrote:
Sorry, I terribly mis-stated the problem. The dialog doesn't seem to support the "_default_" coefficient wave option on the Coefficients tab. I always get "Igor can't tell how many coefficients the fit function requires. Select a wave from the Coefficient Wave menu." It seems like Igor doesn't process the comment info in the "//CurveFitDialog/" format when the function is all-at-once.

You are absolutely right -- the Curve Fitting dialog can run all-at-once fits as long as you select a pre-existing coefficient wave. You will not see the "human-readable" name of each coefficient, though, which is a major bummer when you have a complicated function.


Yes, the dialog actually parses the code to figure out how many fit coefficients it has. The parsing code doesn't understand the format of an all-at-once fit function.

It really should fall back on the comments, though. I will add that to my to-do list.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

June 10, 2010 at 04:50 pm - Permalink