Re: Info about warning during fit

Hi,

 Minuit did not finish the minimization due to a call limit. Try to increase it by doing before fitting:

ROOT::Math::Minimizer::SetDefaultMaxFunctionCalls(n);

with n=10000 for example or larger if needed. The default depends on the number of parameters but is around 1350 in your case

Best Regards

Lorenzo
On May 26, 2010, at 11:33 AM, Alberto Pulvirenti wrote:

> Dear ROOTers,
>
> I tried to fit a 4-points histogram with a function, and the messages coming out from this were the following:
>
> Warning in <Fit>: Abnormal termination of minimization.
> FCN=196.407 FROM MIGRAD STATUS=CALL LIMIT 1354 CALLS 1355 TOTAL
> EDM=0.119482 STRATEGY= 1 ERROR MATRIX UNCERTAINTY 25.4 per cent
> EXT PARAMETER APPROXIMATE STEP FIRST
> NO. NAME VALUE ERROR SIZE DERIVATIVE
> 1 p0 4.35504e+08 5.93416e+07 1.03921e+07 -1.89308e-07
> 2 p1 4.81494e-03 1.32883e-04 -1.25183e-05 -1.00685e+05
> 3 p2 6.22473e+00 3.22669e-02 3.13859e-03 2.24320e+02
> Warning in <TBackCompFitter::GetCovarianceMatrix>: Invalid fit result
> Error in <TF1Helper>: Fitter has no covariance matrix
> Warning in <TBackCompFitter::GetCovarianceMatrix>: Invalid fit result
> Error in <TF1Helper>: Fitter has no covariance matrix
> Warning in <TBackCompFitter::GetCovarianceMatrix>: Invalid fit result
> Error in <TF1Helper>: Fitter has no covariance matrix
> Warning in <TBackCompFitter::GetCovarianceMatrix>: Invalid fit result
> Error in <TF1Helper>: Fitter has no covariance matrix
> Warning in <TBackCompFitter::GetCovarianceMatrix>: Invalid fit result
> Error in <TF1Helper>: Fitter has no covariance matrix
>
>
> The last errors were due probably to my attempt to compute the integral error, which probably fails due to this "abnormal termination" of the fit.
> How can I check what's wrong with the fit and how I can get rid of that?
>
> Thanks, cheers
>
> Alberto
>
Received on Wed May 26 2010 - 11:42:37 CEST