Is it possible to apply a Gaussian transformation to the target variable for TMVA regression?
I book my bdt regression using
factory->BookMethod( TMVA::Types::kBDT, “BDT”, “!H:!V:VarTransform=G,D,G,D:NTrees=100:BoostType=AdaBoostR2:SeparationType=RegressionVariance:nCuts=20:PruneMethod=CostComplexity:PruneStrength=30” );
This seems to apply the desired Gaussian,Decorrelate,Gaussian,Decorrelate sequence for my input variables, but the (single) regression target variable seems not to be transformed. I would like to apply just the simple gaussian transformation to it.
I’m using TMVA 4.1.2, and the release notes imply that this is possible
"Variable transformations can now be applied to a user-defined subset of variables (and regression targets)."
Though I couldn’t find documentation so far on how to do it.