Semi-clean TMVA training from Python

#! /usr/bin/env python

import sys, time
from ROOT import TFile, TTree, TCut, TMVA

## Files
inputFile = TFile("tmva_reg_example.root")
outputFile = TFile("out.root", 'RECREATE')

## Create instance of TMVA factory
factory = TMVA.Factory("TMVARegression", outputFile, "!V:!Silent:Color:DrawProgressBar")
factory.SetVerbose(True)

## Define the input variables that shall be used for the MVA training
## note that you may also use variable expressions, such as: "3*var1/var2*abs(var3)"
## [all types of expressions that can also be parsed by TTree::Draw( "expression" )]
factory.AddVariable("var1", "Variable 1", "units", 'F')
factory.AddVariable("var2", "Variable 2", "units", 'F')

## You can add so-called "Spectator variables", which are not used in the MVA training,
## but will appear in the final "TestTree" produced by TMVA. This TestTree will contain the
## input variables, the response values of all trained MVAs, and the spectator variables
#factory.AddSpectator("spec1:=var1*2", "Spectator 1", "units", 'F')
#factory.AddSpectator("spec2:=var1*3", "Spectator 2", "units", 'F')

## Add the variable carrying the regression target
factory.AddTarget("fvalue")
## For MLP, can declare additional targets for multi-dimensional regression, e.g.
# factory.AddTarget("fvalue2")

## Set up regression tree
# for k in inputFile.GetListOfKeys():
#    print k
regTree = inputFile.Get("TreeR")
regWeight = 1.0
factory.AddRegressionTree(regTree, regWeight);
## This would set individual event weights (the variables defined in the
## expression need to exist in the original TTree)
factory.SetWeightExpression("var1", "Regression")

## Prepare for regression with no cut
## Cut *could* be e.g. TCut("abs(var1)<0.5 && abs(var2-0.5)<1")
## Tell the factory to use all remaining events in the trees after training for testing:
factory.PrepareTrainingAndTestTree(TCut(""), "nTrain_Regression=0:nTest_Regression=0:SplitMode=Random:NormMode=NumEvents:!V")
## If no numbers of events are given, half are used for training, and the other half for testing:
# factory.PrepareTrainingAndTestTree(mycut, "SplitMode=random:!V")

## Set up machine learning method
#config = "!H:!V:VarTransform=Norm:NeuronType=tanh:NCycles=20000:HiddenLayers=N+20"
config = "!H:!V:VarTransform=Norm:NeuronType=tanh:NCycles=20000:HiddenLayers=N+20,N+20"
config += ":TestRate=6:TrainingMethod=BFGS:Sampling=0.3:SamplingEpoch=0.8"
config += ":ConvergenceImprove=1e-6:ConvergenceTests=15:!UseRegulator"
factory.BookMethod(TMVA.Types.kMLP, "MLP", config)
# TODO: add CalculateError
#factory.BookMethod( TMVA.Types.kMLP, "MLPBFGS", "H:!V:NeuronType=tanh:VarTransform=N:NCycles=600:HiddenLayers=N+5:TestRate=5:TrainingMethod=BFGS:!UseRegulator" )
#factory.BookMethod( TMVA.Types.kMLP, "MLPBNN", "H:!V:NeuronType=tanh:VarTransform=N:NCycles=600:HiddenLayers=N+5:TestRate=5:TrainingMethod=BFGS:UseRegulator" ) # BFGS training with bayesian regulators
#
#factory.BookMethod(TMVA.Types.kSVM, "SVM", "Gamma=0.25:Tol=0.001:VarTransform=Norm")
#factory.BookMethod(TMVA.Types.kBDT, "BDT", "!H:!V:NTrees=100:nEventsMin=5:BoostType=AdaBoostR2:SeparationType=RegressionVariance:nCuts=20:PruneMethod=CostComplexity:PruneStrength=30")
#factory.BookMethod( TMVA.Types.kBDT, "BDTG", "!H:!V:NTrees=2000::BoostType=Grad:Shrinkage=0.1:UseBaggedGrad:GradBaggingFraction=0.5:nCuts=20:MaxDepth=3:NNodesMax=15")

## Train MVAs using the set of training events
factory.TrainAllMethods()
## Evaluate all MVAs using the set of test events
factory.TestAllMethods()
## Evaluate and compare performance of all configured MVAs
factory.EvaluateAllMethods()

## Save the output
outputFile.Close()