Statistical uncertainty of sum of two correlated histograms

ROOT
1

Hi guys,

I have two correlated histograms AAAA, BBBB.
likes:

TH1D* AAAA = (TH1D*)file1->Get("");
TH1D* BBBB = (TH1D*)file1->Get("");

I want to calculated the statistical uncertainty for the sum of AAAA->Add(BBBB).
By default, I used the sumw2(). I know sumw2 is only correct to two independent histograms.
So, how will I do? It is better if give a corresponding example.

Thanks a lot.

2

Maybe @moneta can help

3

Hi,

In this case you will have to compute the resulting bin uncertainties yourself, using the obtained values from GetBinError() of the two histograms and the known correlation using for example the standard formula of correlated error propagation, for example from

Lorenzo

1 Like
4

@moneta, Nice. Thanks a lot.

5

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.