Currently I am working on a fit of the invariant mass distribution of D^0. For this I currently use an exponential background, gaussian signal and template function for reflections which I obtain from Monte Carlo (MC). This template function is an interpolation of a invariant mass histogram of reflections. I scale this template function based on the ratio between the amplitude of the gaussian signal from MC and the gaussian signal I get from fitting on the real data. The template function is a linear interpolation (which i wrapped in a RooAbsPdf derived class). In mathematical form that would make the model:
M(x) = A_S \left[ G(x; \mu, \sigma) + \dfrac{T(x)}{ A_S^{ (MC) } }\right] + A_B E(x; \lambda)
Where G=\exp[-(x-\mu)^2/(2 \sigma^2)] and E=\exp(\lambda x), also A_S^{(MC)} is fixed in this model and determined from another fit on the MC signal. There might be other methods on how to “scale” this template function, currently I use the ratio between the gaussian signal in the experimental data and the gaussian signal in the monte carlo data. Now I have trouble with how to implement such a model with RooFit and would like to have some advice on how to do this.
Hi,
Thanks for the post and welcome to the ROOT Forum!
If I understand correctly, you are trying to perform a fit of the function above, M(x). Within it, you are trying to scale the template function T(x) with the ratio of the gaussian signal amplitude in the data divided by the one in the simulation.
If the above is correct, and bear with me if I misunderstood, would the task be addressed by a RooSimultaneous of M’(x), data gaussian fit and simulation gaussian fit, where the functional form of M’ is like the one of M, but you use the ratio of the two aforementioned amplitudes to scale T?
Best,
Danilo