Re: how to fit a histogram with fixed parameters

From: Xianguo (Hsien-Kuo) LU <hsienkuo.lu_at_gmail.com>
Date: Tue, 20 Jun 2006 18:07:29 +0800


Hi Rene,

Here is the code:

test()
{
  h=new TH1F("h","",100,0,1);
  for(int i=0;i<10000;i++){
    Double_t x1=gRandom->Gaus(0.3,0.05);     h->Fill(x1);
  }
  h->Draw();

  TF1 *f=new TF1("f","gaus",0,1);
  //f->FixParameter(1,0.3);
  f->SetParLimits(1,0.3,0.3);
  h->Fit(f);
}

When FixParameter is used, the result is:   FCN=32.5374 FROM MIGRAD STATUS=CONVERGED 37 CALLS 38 TOTAL

                     EDM=1.80376e-09    STRATEGY= 1      ERROR MATRIX
ACCURATE
  EXT PARAMETER                                   STEP         FIRST
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
   1  Constant     8.10627e+02   1.00062e+01   2.29596e-02  -6.79730e-06
   2  Mean         3.00915e-01     fixed
   3  Sigma        4.90648e-02   3.54994e-04   5.52081e-06  -2.63700e-02

And about the SetParLimits:
    FCN=38.4282 FROM MIGRAD    STATUS=CONVERGED      37 CALLS          38
TOTAL
                     EDM=2.83056e-07    STRATEGY= 1      ERROR MATRIX
ACCURATE
  EXT PARAMETER                                   STEP         FIRST
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
   1  Constant     8.00170e+02   9.92677e+00   2.45814e-02  -8.89375e-05
   2  Mean         2.99435e-01     fixed
   3  Sigma        4.96739e-02   3.64274e-04   6.04274e-06  -3.11864e-01

In both cases, you can see the value of Mean is changed. How can I really fix it?
I am using 5.10/00.

Thanks!
Cheers!
Xianguo

On 6/12/06, Rene Brun <Rene.Brun_at_cern.ch> wrote:
>
> Could yous end the SHORTEST POSSIBLE RUNNING script reproducing the
> problem?
>
> Rene Brun
>
> Xianguo (Hsien-Kuo) LU wrote:
> > Hi, rooters!
> >
> > I want to fit a histogram with fixed parameters.
> > I did the following:
> >
> > f=new TF1("f","gaus",0,1);
> >
> > f->SetParLimits(1,1.1156,1.1156)
> > or
> > f->FixParameter(1,1.1156)
> >
> > but after fitting, the parameter changed.
> > Can somebody tell me why and how to solve it?
> >
> > Thanks in advance!
> >
> > Cheers!
> > Xianguo LU
> >
> >
> >
> > --
> > Take flight into the sky, beyond the moon, beyond my mind.
> >
> > Xianguo LU (Hsien-Kuo LU)
> > Physics Department, Peking University
> > Beijing, China
> > Tel: 0086-10-62753888(o)
>
>

-- 
Take flight into the sky, beyond the moon, beyond my mind.

Xianguo LU
Physics Department, Peking University
Beijing, China
Tel: 0086-10-62753888(o)
Received on Tue Jun 20 2006 - 12:07:40 MEST

This archive was generated by hypermail 2.2.0 : Mon Jan 01 2007 - 16:31:59 MET