Dear Experts,
I have hard times finding the solution for the following problem, which is aimed to be solved with the RooFit framework:
I have two samples of events (physics and control), which have partially shared pdfs.
I aim to build something like that:
PHYSICS PDF: nTM PDF_{TM} + nSCF f PDF_{SCF}+nBBbar PDF_{BBbar}+nQQbar PDF_{QQbar}
CONTROL SAMPLE PDF: nSCF (1-f) PDF_{SCF}+mBBbar PDF_{BBbar}+mQQbar PDF_{QQbar}
nTM, nSCF, nBBbar, nQQbar, mBBbar, mQQbar — yields.
PDF_{SCF}, PDF_{BBbar}, PDF_{QQbar} are really supposed to be the same for both physics and control sample pdfs.
From the signal Monte-Carlo I will fix shapes of the following pdfs:
PDF_{TM}, PDF_{SCF}
And also value of “f”-ratio for the SCF between physics and background samples will be fixed from monte-carlo truth (no fit of f-parameter is required).
My question is: What can be the way allowing to insert somehow constant f-parameter to the above pdfs, such that nSCF yield will be the same for the physics and control samples?
So far I was able to do the fit of the following:
PHYSICS PDF: nTM PDF_{TM} + nPeak PDF_{Peak}+nQQbar PDF_{QQbar}
CONTROL SAMPLE PDF: mPeak PDF_{Peak}+mQQbar PDF_{QQbar}
using simultaneous fit with components (again parameters of PDF_{Peak} and PDF_{QQbar} are shared), however it is a big simplification of what I really aim to perform.
I need some sort of hint of PDFs redefinition but can’t figure out by myself how properly to do that.
Thank you in advance!
Best regards,
Borys