I’ve been working on a extension of the root macro /root/tutorials/mlp/mlpHiggs.C (with ROOT’s TMultiLayerPerceptron) to provide separation between my signal and background and use the output (a *.cxx file) later on when doing a particular measurement. Problem: for each NN training I find a different measurement. Is there any way to cure this? Perhaps play with the seed and set it to a fixed value??
Any help is greatly appreciated.
P.S:Attached my macro here.
SetI_SM_mlpTop_type3_tauQCD.C (21.5 KB)