TMinuit Error Parameters

Hi,
This may be a obvious question, but can someone explain how the Errors are calculated below. For example (gausn+pol1) is the example I have given below and in particular I would like to know how the error for the area (p0) parameter is obtained.

FCN=30.7707 FROM MIGRAD STATUS=CONVERGED 879 CALLS 880 TOTAL
EDM=3.45088e-08 STRATEGY= 1 ERROR MATRIX ACCURATE
EXT PARAMETER STEP FIRST
NO. NAME VALUE ERROR SIZE DERIVATIVE
1 p0 1.48627e+04 7.36927e+02 1.35092e-06 -1.95835e-01
2 p1 1.76204e+03 1.40553e+00 6.22047e-07 -6.11073e-01
3 p2 5.77287e+01 1.96630e+00 7.41768e-07 6.57390e-01
4 p3 1.92024e+02 1.87543e+01 1.86184e-03 -5.99276e-05
5 p4 -5.58073e-02 1.11152e-02 1.06205e-06 -9.14588e-02

Thanks,

In a simple case, like in a chi2 fit, errors in Minuit are calculated from the second derivatives of the chi2 function. This is valid if the chi2 function can be well approximated as a parabola around the minimum.

If you want to know more, see the following document

project-mathlibs.web.cern.ch/pro … nerror.pdf

Best Regards

Lorenzo