What is wrong with my Fit

[Jamshid435]

Hi Dear Experts…!!
I am working on RooFit to fit the data of di-muon invariant maas distribution. I’m trying to separate the signal from the background by fitting the sideband regions. I used two guassians pdfs for the signal region with mean, m= 91 and (sigma1, sigma2) = ( 1.5, 2.0) and a cheybechev polynomial for the sideband region with (a0, a1) = 0.5, 0.6) I’ve chosen the number of signal events as (5000, 0, 1000000) and that of the background as (3000, 0, 100000). And I get the fit as

I would like to ask, how can I improve my fitting…???

[StephanH]

Hi,

@Jamshid435

I would like to ask, how can I improve my fitting…???

• You could post the model. Maybe there’s a mistake or a typo somewhere.
• The peak is probably not exactly Gaussian. Maybe you have to allow for another signal component, e.g. a Gaussian with slightly shifted mean or different sigma.
• Check out tutorial ROOT: tutorials/roofit/rf201_composite.C File Reference to plot the different components + side band model in one go.