Errors of fit parameters in ROOT

ROOT
1

Dear ROOT community,

I’ve been thinking about this for a while, and when using ROOT for my scientific work and fitting with ROOT, I’d like to be prepared for questions: “How are the errors of the fit parameters” calculated.

Some digging through the ROOT code files did not clarify this question to me, hence I would like to ask the community to explain (or maybe refer to a document) how the fit parameter errors are calculated from ROOT.

Cheers,
Robert

2

Can you clarify which “fitting” you are talking about ? TH1::Fit ? roofit ?

3

A sorry, I didn’t specify that.

I’m using TH1::Fit on a TGraphErrors; the x-errors are zero, y-errors are present.

Cheers

4

It’s recommended to call TH1::Fit with the “E” option if you want reliable errors on the fit function’s parameters. That option will fit using MINOS, and some info can be found here: http://cmd.inp.nsk.su/old/cmd2/manuals/cernlib/minuit/node14.html