8000 Normalization parameter errors of ExtendedUnbinnedNLL with weighted data · Issue #606 · zfit/zfit · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Normalization parameter errors of ExtendedUnbinnedNLL with weighted data #606

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
spyrosmer opened this issue Jan 29, 2025 · 3 comments
Open

Normalization parameter errors of ExtendedUnbinnedNLL with weighted data #606

spyrosmer opened this issue Jan 29, 2025 · 3 comments

Comments

@spyrosmer
Copy link
spyrosmer commented Jan 29, 2025 •
edited
Loading

Dear Zfit experts,

I just discovered Zfit and I am very happy about its capabilities. I have a very naive question.

I am trying to create an ExtendedUnbinnedNLL made of exponential pdfs with weighted data. I am extracting the hesse errors for :

  1. The exponential parameters --> Which seem reasonable
  2. The normalization parameters (yield) --> Which is almost 0.

I am not sure why (2) happens, through other frameworks (Like RooFit etc.) I get the almost the same results for the exponential parameter and its error and the same normalization parameter but higher error.

I am using zfit version : 0.16.0

I attach the code and the log file with the results.

spyros.log

code.txt

Thank you very much for your time!
@spyrosmer spyrosmer changed the title Normalization parameter errors with weighted ExtendedUnbinnedNLL Normalization parameter errors of ExtendedUnbinnedNLL with weighted data Jan 29, 2025
@jonas-eschle jonas-eschle mentioned this issue Jan 31, 2025
Merged
8 tasks
@jonas-eschle
Copy link
Contributor

Thanks a lot for bringing this up, in fact, we're aware that there are some issues regarding the extended parameter uncertainties in weights! I'll be back with some more information to be able to fully fix this, meanwhile we're anyway adding and changing the weigth correction method (see #607 ) to also allow to disable it or to use a much simpler, yet less precise procedure.

On it, back with more later on!

@spyrosmer
Copy link
Author
spyrosmer commented Jan 31, 2025 •
edited
Loading

Hi @jonas-eschle thanks for the nice framework and the quick look! I also posted the issue on stack overflow https://stackoverflow.com/questions/79402148/normalization-parameter-errors-of-extendedunbinnednll-with-weighted-data
as suggested.

Do you think a simple σ_norm_param = sqrt(sum(weight_i^2))

would be an ok calculation for now?

Are the non extended parameter uncertainties correctly calculated anyways?

@jonas-eschle
Copy link
Contributor

Hi @spyrosmer apologies for the late reply, I thought I've written here already. This is indeed possible and a good approximation, and with the newest version (0.25.0) actually available in zfit. You can now use the weightcorr=... argument to disable the approximation, use the sum of the weights or the asymptotically correct (well, or in this case unfortunately incorrect) correction.

That being said, we're working on this of course as well, just a bit trickier

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@jonas-eschle @spyrosmer
0