Hi Heiko,
with 'non extended chi^2 fit' I actually meant that I performed a
minumum chi^2 fit with model that was built using a different syntax
w.r.t. the one I have used until now...
If I have understood correctly, in RooFit when you specify the range for
the numbers of events for all the PDF in your model, i.e.:
RooRealVar argusevent("","argus",10000,0,200000); //Argus events
RooRealVar crisballevent("","crisball",10000,0,200000); //cb events
RooRealVar signalevent("","signal",10000,0,20000); //signal events
RooAbsPdf
model("","model",RooArgList(argus,cristalball,signal),RooArgList(argusevent,crisball,signalevent));
the model is extended by default, while doing something like:
RooRealVar argusfrac("","argus",0.5,0,1); //Argus fraction
RooRealVar signalfrac("","signal",0.5,0,1); //signal fraction
RooAbsPdf
model("","model",RooArgList(argus,signal,cristalball),RooArgList(argusfrac,signalfrac));
the model is non-extended.
So this time I used the latter definition to build the model and then I
performed the chi^2 fit.
Antonio
Heiko Lacker ha scritto:
> Hi Antionio,
>
> that looks now much better.
>
> What is not cleat to me is your statement
> that now it is 'non extended chi^2 fit'.
>
> Could you please explain what you mean?
>
> Cheers,
> Heiko
>
>
> On Wed, 3 May 2006, Antonio Petrella wrote:
>
>> Hi Heiko,
>> I was just finishing the update of this section....
>>
>>
>>> I have just seen that there is something new.
>>>
>>> * You say that it is now a non-extended chi^2 fit.
>>> Could you please give some technical details?
>>> Also, it would be interesting to see the fitted
>>> parameters.
>>>
>> Now the table is complete!
>>
>>> * Should we compare these results with the ones
>>> posted at April 12th?
>> No, those are wrong since there is no scaling with the errors.
>>
>>> * Could you please post on your webpage somewhere
>>> the numbers of entries in the histo for
>>> a) data
>>> b) MC
>>>
>> Sure, I added these informations right below the title.
>> I write them also here:
>>
>> a) data = 437889
>> b) MC = 1561024
>>
>>> * It might be also helpful to provide the sum of the
>>> fitted yields in order to see how it compares to the
>>> number of entries in the histo.
>>>
>> Ok, I will add this now.
>>
>> Bye,
>> Antonio
>>
>>> Cheers,
>>> Heiko
|