Do the cuts for my TMVA training affect the weights I chose?

Hi everyone,

I’m using TMVA Root on several samples (Signal vs Backgrounds). In inputs, I specified a weight for every background for the training to make my sample statistics fit with the reality. But, I also applied some cuts for the training in the PrepareTrainingAndTestTree method. Then these cuts affect the statistic of my samples. My question is : Does TMVA take into account the cuts and modify the existing weights to keep the same kind of statistic before the cuts ?

Thanks !

I guess @moneta can help you.


I don’t think the cuts will affect the weights. The weights are needed to rescale the backgrounds to take into account the correct proportions of them, which depends on their relative production rates. If, afterwards you apply some selection cuts, this should not affect these weights.


1 Like