search

nicoeinsidler / pigor

Quick analysis tool for the beam lines at Atominstitut of TU Wien in Vienna, Austria.
http://www.pigor.org/
MIT License
2 stars 0 forks source link

fit improvement #23

Open uncertaintyguy opened 4 years ago

uncertaintyguy commented 4 years ago

Dear Nico,

as discussed, I am sending you an example of a bad sinusoidal fit which nonetheless has about the right frequency. Thanks a lot that you are still open to work on your code.

I am not sure what files you need. I renamed the .dat file as .txt so I can attach it. Please let know in case this didn't work or you need something else.

Armin

PS: how do I add a bug label or do only you do that?

Edit: Initially, I accidentally attached a fit which worked quite good ;)

2020_06_18_1017_100s_comp_tilt9_DC4_scan_sine 2020_06_18_1017_100s_comp_tilt9_DC4_scan_sine_fit_functions.txt 2020_06_18_1017_100s_comp_tilt9_DC4_scan_sine - Copy.txt

nicoeinsidler commented 4 years ago

Bug is reproducible, in the method where I try to find meaningful bounds an arbitrary scaling factor s exists, which describes the width of the parameters values that are fed into the fitting routine. This factor doesn't seem to have any effects on the used bounds. Don't know why and why it works on so many other data sets though. Investigating...

Ad labels: Do you see the 'labels' tab on the right hand side? If not I may have to give you additional rights for this repository.