Re: [ROOT] Fitting in TGraphError

From: Rene Brun (Rene.Brun@cern.ch)
Date: Wed May 19 2004 - 09:49:11 MEST


Hi Alberto,

Alberto Pulvirenti wrote:
> 
> Thanks for the answer, I read the documentation and I fount answer to
> all question, except one:
> 
> I have understood that the LeastSquareLinearFit() method does a straight
> line fit through the data.
> But, it does not take into account the errors, because a common TGraph
> has not errors in its points.
> 
> In the case of a TGrapErrors object, does it perform the linear fit
> weighting all points with its error, or not?

Yes. Please read carefully the documentation of TGraph::Fit at:
http://root.cern.ch/root/htmldoc/TGraph.html#TGraph:Fit


Rene Brun

> 
> Cheers,
>         Alberto
> 
> On Thu, 2004-05-06 at 15:09, Rene Brun wrote:
> > Hi Alberto,
> >
> > see the documentation of TGraph and TGraph::Fit, etc at:
> >
> > http://root.cern.ch/root/htmldoc/TGraph.html
> > http://root.cern.ch/root/htmldoc/TGraph.html#TGraph:Fit
> > http://root.cern.ch/root/htmldoc/TGraph.html#TGraph:LeastSquareLinearFit
> >
> > Rene Brun
> >
> > Alberto Pulvirenti wrote:
> > >
> > > Sorry, I forgot to put a topic in the email, and maybe many of you could
> > > have deleted it as spam.
> > > Here it is again.
> > >
> > > Cheers,
> > >         Alberto
> > >
> > > On Wed, 2004-05-05 at 15:38, Alberto Pulvirenti wrote:
> > > > Dear ROOTers,
> > > >
> > > > I have seen that it is possible to fit a TGraphErrors object with a
> > > > given TF1.
> > > >
> > > > My question is: does this fit take into account the errors which I pass
> > > > to this object, or not?
> > > > Maybe, must I specify something in order to accomplish this task?
> > > >
> > > > Another question: I saw in the class definition a method called
> > > >
> > > > virtual void LeastSquareLinearFit(Int_t n, Double_t& a0, Double_t& a1,
> > > > Int_t& ifail, Int_t first, Int_t last)
> > > >
> > > > Does it fit (according to parameters names) to which of the following?
> > > > 1) a0 + a1*x
> > > > 2) a0*x + a1
> > > >
> > > > What does 'ifail' mean?
> > > >
> > > > Thanks for all.
> >



This archive was generated by hypermail 2b29 : Sun Jan 02 2005 - 05:50:08 MET