CombinedFit.C in RooFit

[smgrig]

Dear RooFit experts,

there is a very useful ROOT fit code:
https://root.cern.ch/doc/master/combinedFit_8C.html

Could you please provide its implementation in the RooFit?

Thanks a lot in advance,
Smbat

[jonas]

There is a very similar tutorial for RooFit:

https://root.cern.ch/doc/master/rf501__simultaneouspdf_8C.html

It’s not exactly the same model and fit (it’s a likelihood fit instead of a chi2 fit), but the concepts are the same.

Maybe that’s already good enough for you to learn from, so you don’t need a rewrite of the combinedFit tutorial in RooFit?

Cheers,
Jonas

[smgrig]

Dear Jones,
thank you for the tutorial.

In fact the combinedFit can use the likelihood fit too.
Actually I’m doing simultaneous fit of more than 100 datasets with functions having at least one common parameter, using the combinedFit method.
Now I’d like to do the same fit with the RooFit to compare performances of two methods.

Cheers, Smbat

[jonas]

Nice, that’s a good idea to use RooFit in this scenario

One of the selling points of RooFit is exactly a big speedup when fitting many datasets simultaneously, because of its extensive caching. And if you want the best performance in RooFit, be sure to give a try to the BatchMode() option for createNLL() or fitTo(), explained in the documentation:

https://root.cern.ch/doc/master/classRooAbsPdf.html#a52c4a5926a161bcb72eab46890b0590e