Skip to content
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.

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

Dumping of statistical results #19

Open
marinang opened this issue Dec 6, 2019 · 5 comments
Open

Dumping of statistical results #19

marinang opened this issue Dec 6, 2019 · 5 comments
Labels
enhancement New feature or request

Comments

@marinang
Copy link
Member

marinang commented Dec 6, 2019

After a chat yesterday at CERN with @mayou36 and Carsten, a RooFit developer, we concluded that storing the results of hepstats is crucial as they can be reused, for plotting for instance, without re-running CPU-intensive calculations done in hepstats.

The idea is to store in a dictionary:

  • best fit result
  • likelihood scan
  • significances
  • upper limits / intervals
  • null and alternative hypothesis
  • parameter ranking (not implement yet)

This dictionary can be easily serialised into yml or json file.

@marinang marinang added the enhancement New feature or request label Dec 6, 2019
@eduardo-rodrigues
Copy link
Member

Brilliant, really nice to see these ideas moving forward (a lot of this has been discussed among the community in recent times).

cc @benkrikler - folks, I would like to organise a PyHEP WG meeting early next year as a follow-up to the fitting topical meeting we had. Seems to me that fitting tools and connections to statistics tools are a hot and important topic. What would be the timescale for you folks to be available for a dedicated meeting? I would imagine a couple of short presentations and a few slides backing up an open discussion. We can continue this discussion offline as it's far more general than this repo :-D.

@jonas-eschle
Copy link
Contributor

Yes, that sounds good, I am mostly available in the new year.

Additionally, we also discussed the storage of the likelihood, most importantly of the fitting model. Since this is a larger topic, maybe overlapping with preservation (?), do you know by chance people that are also working on that?

@eduardo-rodrigues
Copy link
Member

Good point, @mayou36. I do know at least 2 relevant people, namely Lukas and Matthew from the pyhf team. I will throw an email soon to start organising a meeting … Not to be done here on GitHub.

@jonas-eschle
Copy link
Contributor

Thanks, @eduardo-rodrigues those I know and we're already in contact. In fact, it's them, Carsten (from RooFit) and me who are mostly interested in this.
But if you know others (more from analysis preservation?), let me know

@eduardo-rodrigues
Copy link
Member

I thought you would … I think we will get more names is the preparations of a meeting, especially that I will include the HSF analysis WG in there too, since the link is very relevant in this instance ...

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

3 participants