Create a Forum Topic by Logging In or Creating an Account

Empty W_sigma

Wed, 04/21/2021 - 03:15 am

Heyho!

Does anyone know what causes the W_sigma wave of a fit to be empty? I am using a relatively complex procedure in order to perform a batchcurve fit of various sets of data. The fitfunction involves solving a set of differential equations using integrateODE. The fit itself works and the chisquare is also meaningful, but the W_sigma wave is empty apart from the last two rows containing unreasonably large numbers.

Thanks for your help already in advance!

Cheers,

Angelika.

chozo

Hi Angelika, it is unfortunately a bit difficult to imagine what you are doing and what might go wrong from your short description. Would it be possible for you to post a minimal experiment file with a description of the steps you are doing? Sometimes it even helps to identify the problem by trying to create a minimal version of the task or explaining it here in detail (AKA rubber duck debugging).

April 21, 2021 at 08:21 am - Permalink

johnweeks

Sounds like something went wrong with the fit. My guess would perhaps be that it didn't converge within the allowed 40 iterations. If this is in Igor code, you should use the V_fitError and V_fitQuitReason variables to get information on the status of a fit. For iteration limit, you need V_fitQuitReason.

Slow convergence and large coefficient errors tend to be an indication of "identifiability" problems. If you have a sufficiently recent version of Igor, you can read our description of that problem: DisplayHelpTopic "Identifiability Problems"

April 21, 2021 at 11:18 am - Permalink