Error on 2 intersecting linear fits


I am trying to find the error on the on the intersection coordinates of two linear fits. I am able to access the gradient and constants of both fits along with their corresponding errors. How do i access the correlation coefficient between the error on the gradient and the error on the constant?

Thanks, Claire

Hi ,

to get the correlation coefficient you need to get the covariance matrix from the fit from the TVirtualFitter.

After fitting you can do :

TVirtualFitter * fitter = TVirtualFitter::GetFitter();

   // for i and j 
 double corr_ij = fitter->GetCovarianceMatrixElement(i,j) / sqrt( fitter->GetCovarianceMatrixElement(i,i) *
                                                   fitter->GetCovarianceMatrixElement(j,j) ) ;