Summary: | custom fitting fails | ||
---|---|---|---|
Product: | [Applications] LabPlot2 | Reporter: | Uwe Stöhr <uwestoehr> |
Component: | general | Assignee: | Stefan Gerlach <stefan.gerlach> |
Status: | RESOLVED WORKSFORME | ||
Severity: | normal | CC: | stefan.gerlach, yurchor |
Priority: | NOR | Keywords: | investigated, triaged |
Version: | 2.4.0 | ||
Target Milestone: | --- | ||
Platform: | Microsoft Windows | ||
OS: | Microsoft Windows | ||
Latest Commit: | Version Fixed In: | ||
Attachments: | screenshot describing the axes and fit result error |
Description
Uwe Stöhr
2017-04-23 00:39:33 UTC
Finding the fit parameter automatically is not implemented yet. We will care about this. A result like "1±nan (nan %)" usually means that the fit did not converge and the error could not be estimated. Probably the start value was not chosen well. Can you try a better start value? (I checked it and it works for me) Please supply the data if it still doe not work. > the "a" is not recognized as fit variable This works now in LP 2.5RC2. > - use for the parameter "a" as start value "1" and press recalculate > result: one gets this: a = 1±nan (nan %) This is still the case. It is clear that the fit did not converge and therefore I expect as user an information like: "The fit did not converge. Please use a different start value for the parameter 'a'." Created attachment 109058 [details]
screenshot describing the axes and fit result error
Now the behavior of LabPlot is different. Nevertheless the user does not get the info that the fit did not converge and that he should use different start values.
Moreover, the fit plot is incorrect, see the axes in the attached screenshot.
resetting bug status Hi, i uploaded a new package including the latest fixes (also fixing a parser crash on my system) at https://theorie.physik.uni-konstanz.de/gerlach/labplot/labplot-2.5.0rc3-64bit-setup.exe There are two problems in your screenshot. The axes currently do not support values bigger than 1e+15. That's already on our TODO list. The second problem is the missing Unicode support in the fit result. I do not have this problem but i am no Windows expert to help here. > There are two problems in your screenshot. The axes currently do not support values bigger than 1e+15. In my opinion that is not a bug but that case that the y-axis has the auto fit option and as you can see this uses internally the values: -4.30099e+42 3.11822e+43 for the axis min/max. To fix this, LabPlot should limit the max/min value to -1e+15/1e+15 of whatever is a sensible value. > missing Unicode support in the fit result Yes, this is only in my build and I haven't yet found out why. I see that you use Qt 5.9.2. Qt 5.9.3 is available. This bug has had its resolution changed, but accidentally has been left in NEEDSINFO status. I am thus closing this bug and setting the status as RESOLVED to reflect the resolution change. |