Home | News | Documentation | Download

Calculate the sigma between results of the same mesurement when parameters are correlated

Hello, in this topic Calculate the sigma between results of the same mesurement - #12 by moneta @moneta showed me a method to calculate the number of sigma of agreement of parameters measured by three differents instruments (by calculating the X^2 and p-value).

In the experiment I measured 3 parameters (let’s call them x,y,z) by three instruments and I calculated by Root, the number of sigma of agreement. But, in my experiment, x,y and z are correlated, instead in this method, the X^2 is calculated individually for each paramether.
Is there, in ROOT, an easy way to calculate the error of this approssimation (i.e. neglecting the correlation between the parameter), or an easy way to calculate the number of sigma considering also the correlation between the three parameters??
Thank you

Hi,

We don’t have in ROOT a function to compute the Chi2 of correlated measuresuments. It is however not difficult to compute given the covariance matrix of the measurements.
See for example Generalized least squares - Wikipedia

Lorenzo

Hi @moneta, thank you for your reply, I thought it was possible in ROOT !

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