What is inside the Unbinned fit?

ROOT
1

Hi
I have a question regarding what really goes on inside the TTree:UnbinnedFit. But first I’ll describe what it is that I want to do.

I have a rather small data sample and I would like to fit to a threedimensional angular distribution with two fit parameters to the data set. The distribution is a user defined function f3 defined as:
Double_t W(Double_t *x, Double_t *fitpar);
where x is the data and fitpar the parameters. The fit is performed using
TTree->UnbinnedFit(“f3”,“thetagp:thetap:phip”,cut10);
Everything works (apart from a rather large error for one of the parameters but improve that by varying the parameters manually instead.)

The questions are: Is the unbinned fit a log likelihood fit to the data?
Is the UnbinnedFit related to RooFit? Is the RooFit documentation relevant in this case?

Best Regards,
Agnes Lundborg

2

[quote]The questions are: Is the unbinned fit a log likelihood fit to the data?
[/quote]

Yes

[quote]Is the UnbinnedFit related to RooFit? Is the RooFit documentation relevant in this case?
[/quote]
No

Which version of ROOT are you using?
Is your function normalized?

Rene

3

Hi Rene
I am using Root version 4.03-02.
The fit function is not normalized to one, but it is constrained not to grow indefinately with the fit parameters. I noticed your warning there. :slight_smile:
Agnes