Likelihood with a numerically evaluated pdf

Roofit and RooStats
1

Hello everyone,

I have a function F(x) = a * f1(x) + b * f2(x), where “a” and “b” are parameters and “f1(x)” and “f2(x)” are functions, which can hardly be expressed analytically, but it’s possible to evaluate these functions numerically for a set of argument values x = {x1,x2,x3,…} (e.g. for values, which correspond to the data points – data = {data1(x1), data2(x2), data3(x3),…}). How can I define (if it’s possible in principle) my pdf in this case?

Thanks a lot.
Maxim

2

Hi Maxim,

can you express these pds as RooHistPdfs?
What are the values of f(x) between, say, x1 and x2?

Cheers,
Danilo