Re: normalisation of fit parameters

From: Lorenzo Moneta <Lorenzo.Moneta_at_cern.ch>
Date: Wed, 3 Feb 2010 22:03:55 +0100

Hi Karla,

 I would need eventually your root file to run your code. I see however an error at line 121,

> cout << " N in peak = "<<myfit->GetParameter(1)<<" +/- "<<myfit->GetParError(0)<<endl;

should be: myfit->GetParameter(0)

I see also that you compute your-self the t distribution. For your information, in ROOT we have a function providing it, ROOT::Math::tdistribution_pdf(double x, double r)

see http://project-mathlibs.web.cern.ch/project-mathlibs/sw/html/group__PdfFunc.html#gf4aad664d7d8d39845a21cea93ade08e

Regards

 Lorenzo
On Feb 3, 2010, at 9:30 PM, Karla Beatriz Cantun Avila wrote:

> 
> Thank you Lorenzo!
> 
> Attached to this mail you may found my macro,
> I am using a TSudent function to fit signal, and a polinomial of second order to fit background.
> 
> Again thanks for your time!
> 
> Cheers,
> karla
> 
> On Wed, 3 Feb 2010, Lorenzo Moneta wrote:
> 
>> Dear Karla,
>> 
>> The proper normalization depends how you have defined exactly your function. I would need your complete example in order to help you,
>> 
>> Best Regards
>> 
>> Lorenzo
>> On Feb 3, 2010, at 6:34 PM, Karla Beatriz Cantun Avila wrote:
>> 

>>>
>>> Dear roottalkers,
>>>
>>> I am fitting a histogram with the sum of two functions (one per signal and one per background), when I got the values of parameters, the parameter of the total fit which corresponds to number of particles in the peak of the signal is printed out like:
>>>
>>> N 2.568e+06 103447
>>>
>>> but this is not the number of particles in peak.
>>> Dou you know how may I apply the proper normalization in order to get the right numbers? and how to print out them on the statistical pad?
>>>
>>> Thanks in advance,
>>>
>>> karla
>>>
>>>

>> 
>> 
> <FitTest.C>

Received on Wed Feb 03 2010 - 22:04:31 CET