Hi Martina,
The status="RESET" may happen when you change the conditions of the fit,
eg setting limits, parameters, clearing, a call to Minos with no improvements
and returning to the Migrad settings.
The status is set by TMinuit::mnrset
http://root.cern.ch/root/htmldoc/TMinuit.html#TMinuit:mnrset
I will need a concrete/small application to tell you why it happens.
Rene Brun
Schaefer Martina wrote:
>
> Hello,
>
> I'm doing an unbinned fit in a tree.
> Sometimes I'm getting STATUS=RESET and not STATUS=CONVERGED, can anybody
> tell me, what STATUS=RESET is meaning, I didn't found it on the root or
> minuit pages on the web.
>
> Thanks,
> Martina
>
> (As you can see in the following lines, I'm doing a loop - for some i I
> got convergence, for others RESET. Everytime I'm doing first a binned
> fit with the h_help to get initial parameters. When I get "RESET" as
> output, root doesn't try to fit, it just takes the parametres of the
> binned fit,...)
>
> TF1 *f2 = new TF1("f2","[0]*(3/8*(1+x*x)+[1]*x)",-0.95,0.95);
> TF1 *f2u = new TF1("f2u","[0]*(3/8*(1+x*x)+[1]*x)",-0.95,0.95);
> Double_t PAR[2];
> TH1D *h_help = new TH1D("h_help","h_help",25,-1.1,1.1);
>
> //t_cos is a TBranch of a tree
> for(int i=1;i<=5;i++){
> f2->SetLineColor(1);
> j=2*i-5;//defined befor as Int_t
> k=j-2;//same
> //M and delta are defined above
> sprintf(sss,"t_mass <= %f+(%i*%f) && t_mass > %f+(%i*%f) &&
> t_cos >-1.1 && t_cos <= 1.1 ",M,j,delta,M,k,delta);
> sprintf(ss,"t_mass <= %f+(%i*%f) && t_mass > %f+(%i*%f)
> ",M,j,delta,M,k,delta);
> t_theta->Draw("t_cos >> h_help",sss);
> h_help->Fit("f2","R");
> f2->GetParameters(&PAR[0]);
> cout<<"histo : "<<PAR[0]<<" "<<PAR[1]<<endl<<endl;
> f2u->SetParLimits(0,1,1);
> f2u->SetParameters(&PAR[0]);
> t_theta->UnbinnedFit("f2u","t_cos",ss);
> f2u->GetParameters(&PAR[0]);
> f2u->SetLineColor(i+1);
> f2u->Draw("same");
> cout<<"unbinned : "<<PAR[0]<<" "<<PAR[1]<<endl<<endl;
> }
>
> output:
>
> FCN=39.7616 FROM MIGRAD STATUS=CONVERGED 35 CALLS 36
> TOTAL
> EDM=4.75254e-17 STRATEGY= 1 ERROR MATRIX
> ACCURATE
> EXT PARAMETER STEP FIRST
> NO. NAME VALUE ERROR SIZE DERIVATIVE
> 1 p0 1.99509e+01 1.14194e+00 3.55991e-03 9.97977e-12
> 2 p1 -3.37357e-02 5.17265e-02 1.61256e-04 -1.88480e-07
> histo : 19.9509 -0.0337357
>
> FCN=0 FROM MIGrad STATUS=RESET 0 CALLS 0 TOTAL
> EDM= unknown STRATEGY= 1 NO ERROR MATRIX
> EXT PARAMETER CURRENT GUESS PHYSICAL LIMITS
> NO. NAME VALUE ERROR NEGATIVE POSITIVE
> 1 p0 1.99509e+01 fixed
> 2 p1 -3.37357e-02 constant
> unbinned : 19.9509 -0.0337357
>
> --
> ----
> Martina Schaefer SMartina@gmx.de
> "Le risque est comme le sel, il en faut, mais pas trop!"
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:17 MET