Use of machine learning algorithms when distribution of features is parametrized by nuisance parameters or parameters of interest.

Based on scikit-learn and RooFit/RooStats.

Notes on physics example: ttbar_14tev_mx700_alljes.root // input to training ttbar_14tev_jes1.root // input to training ttbar_14tev_jes1_eval.root // file with outputs using 1 input; mx as param ttbar_14tev_alljes_eval.root // file with outputs using 2 inputs; jes & mx params