Dear TMVA users,
I tried the TMVA MLP implementation to perform a regression task. If I plot the resulting distribution of the fitted parameter with
TestTree->Draw( “eta” )
the shape looks good but the x-axis range spans from 0 to about 8. Unfortunately, the “truth” distribution range is [-5, 5]. Is there any clean way to “shift back” the output so I can check the consistency of the two histograms e.g. with a KS test? I’m not sure that a simple loop would be enough, the x-range is calculated on the fly by ROOT when I call Draw.
Thanks,
Riccardo