Likelihood distribution

Roofit and RooStats
1

Hi, all:

I have a question:
Now I have a likelihood distribution with the number of the signal events. Also the systematic error is known. Then how
could I obtain the new likelihood distribution convoluted with a Gaussian function for the systematic error? Here the Gaussian function resolution is Nsig* systematic error .
Thanks in advance!

2

Hi,

To do so you can simply multiply your constraint pdf with your original pdf, i.e.

RooAbsPdf* myPdf ; // e.g. Gaussian(x,mean,sigma)
RooAbsPdf* myConstraint // e.g. Gaussian(mean,mean_val,mean_error))

you construct

RooProdPdf myPdfWithConst(“blah”,“blah”,RooArgSet(*myPdf,*myConstraint)) ;

and then do

RooAbsReal* myLL = myPdfWithConst.createNLL(*data) ;

If you are using a ROOT release earlier than 5.24, an additional argument
’Constrain(mean)’ is needed in the call to createNLL()

Wouter

3

Dear Wouter:

Thanks for your information. However I do not quite understand it.
For my case, I obtained the likelihood value distribution. Please see the enclosed file likelihood.dat. The corresponding range of signal events (x-axis) is from 0 to 24.
The systematic error is 0.17.
Then could you tell me more about how to write it? Thanks.

4

sorry forgot to attach the file