TEfficiency::Fit option S not working?

[gerstein]

Hi,

I am trying to get a TFitResultPtr from fitting a TEfficiency object.
I am calling
e11_5051->Fit(f2,“S”);
but it returns zero.
Other options (i.e. I or V) seem to have an effect, it’s just S that is not working.

any idea why this is happening?

Tried it in several versions of root, latest is 6.06/02

Thanks,
Yuri

[couet]

Can you post a small example reproducing your problem ?

[gerstein]

Well, it’s just this. I have a TEfficiency object e11_5051 ( and as you can guess from the name, many like it). It looks fine, the errors are ok and the fit is correct. It’s just that when I call

e11_5051->Fit(f2,“S”);

It returns zero.
I can make the fit different with “I” option, so some parameters get through.
Just not “S”.

I can send a root file and the .C file that is my function f2, but the problem should be reproducible in one line. Would you like me to do that, nonetheless?

-y

[couet]

Ok, I told the expert in that area. He will have a look and let you know if he needs more details.

[gerstein]

OK, I made a standalone example, inspired by the description of the TEfficiency class:

TEfficiency* pEff = new TEfficiency("eff","my efficiency;x;#epsilon",20,0,10);
TRandom3 rand3;
double x;
for(int i=0; i<100000; ++i){ x = rand3.Uniform(10); pEff->Fill(rand3.Rndm()<TMath::Gaus(x,5,1),x);}
pEff->Draw();
TF1 * f1 = new TF1("f1","[0]*exp(-0.5*(x-[1])*(x-[1])/[2]/[2])",0,10);
f1->SetParameter(0,1);
f1->SetParameter(1,5);
f1->SetParameter(2,2);
pEff->Fit(f1);
pEff->Fit(f1,"I");
TFitResultPtr rr = pEff->Fit(f1,"S");
rr.Get();

[moneta]

Hi,

Yes, option S is not yet implemented for TEfficieny. Since it is trivial to add, I will do it now.
If you need, as workaround you can use the TBinomialEfficiencyFitter class which is used inside TEfficiency

Thank you for reporting this issue

Best Regards

Lorenzo

[gerstein]

Thanks, Lorenzo.

Can you give me a quick line of how to instantiate TBinomialEfficiencyFitter class for my purposes?

[gerstein]

oh, never mind. I guess you just want me not to use TEfficiency and use TBinomialEfficiencyFitter directly. I see examples of that on the web pages.

[moneta]

@gerstein

TEfficiency* pEff = new TEfficiency(“eff”,“my efficiency;x;#epsilon”,20,0,10);TRandom3 rand3;double x;for(int i=0; i<100000; ++i){ x = rand3.Uniform(10); pEff->Fill(rand3.Rndm()<TMath::Gaus(x,5,1),x);}pEff->Draw();TF1 * f1 = new TF1(“f1”,“[0]exp(-0.5(x-[1])*(x-[1])/[2]/[2])”,0,10);f1->SetParameter(0,1);f1->SetParameter(1,5);f1->SetParameter(2,2);pEff->Fit(f1);pEff->Fit(f1,“I”);TFitResultPtr rr = pEff->Fit(f1,“S”);rr.Get();

Hi,

The fix for TEfficiency is available in the master.
For using TBinomialEfficiencyFitter, just see the implementation of TEfficiency::Fit
https://root.cern.ch/doc/master/TEfficiency_8cxx_source.html#l02336

Cheers

Lorenzo