Global Fit Package not calculating residuals properly
I am using the Global Fit Package to globally fit several traces to a User-Defined Function. It seems that the residuals waves are not being calculated properly.
In the attached figure, I show a trace wave0 along with the fit wave, gFit_wave0, and the residuals wave, gRes_wave0, generated by the Global Fit Package. Clearly, gRes_wave0 is not the difference between wave0 and gfit_wave0. Furthermore, when I change the fit parameters, gFit_wave0 is updated, but gRes_wave0 remains the same.
On the other hand, the residuals are being accurately calculated in the wave
root:Packages:NewGlobalFit:NewGF_ResidY
which is displayed in the Global Analysis Progress window (figure attached).
I have tried deleting the gRes_ waves and unchecking and rechecking the "Calculate Residuals" checkbox to recalculate the residuals waves, to no avail. I have tried starting a new Experiment and reloading the Global Fit Package, to no avail.
I am running Igor Pro version 9.0.5.1 Build 56551 on Windows 11 Pro version 23H2.
Thank you in advance for any insight.
Do you maybe use x waves (i.e., an XY form of data with separate waves for each dimension)? Indeed there seems to be a bug with calculating residuals from such data, which I found that may (or may not) fix this issue. Since I am not comfortable with posting these changes here, could you please contact the support directly? You may get a new version sent then.
October 13, 2024 at 08:13 pm - Permalink
In reply to Do you maybe use x waves (i… by chozo
Yes, I am using x waves for the fitted traces.
October 13, 2024 at 09:50 pm - Permalink
I see. Then it is very likely the issue I meant. Please send an email to support@wavemetrics.com referring to this forum post and the staff should know what this is about.
October 14, 2024 at 03:04 am - Permalink
Okay, thanks for letting me know about this. I will reach out to support.
October 14, 2024 at 09:11 am - Permalink
To provide an update, the issue did have to do with the use of x waves, and I reached out to support and they provided me with an updated version of the global fit procedure that resolved the issue.
October 18, 2024 at 09:48 am - Permalink