TMVA::MethodBDT

 the standard constructor for the "boosted decision trees" 

 MethodBDT (Boosted Decision Trees) options:
 know options:
 nTrees=Int_t:    number of trees in the forest to be created
 BoostType=       the boosting type for the trees in the forest (AdaBoost e.t.c..)
                  known: AdaBoost
                         Bagging
 SeparationType   the separation criterion applied in the node splitting
                  known: GiniIndex
                         MisClassificationError
                         CrossEntropy
                         SDivSqrtSPlusB
 nEventsMin:      the minimum number of events in a node (leaf criteria, stop splitting)
 nCuts:           the number of steps in the optimisation of the cut for a node
 UseYesNoLeaf     decide if the classification is done simply by the node type, or the S/B
                  (from the training) in the leaf node
 UseWeightedTrees use average classification from the trees, or have the individual trees
                  trees in the forest weighted (e.g. log(boostweight) from AdaBoost
 PruneMethod      The Pruning method: 
                  known: ExpectedError
                         CostComplexity 
                         CostComplexity2
 PruneStrength    a parameter to adjust the amount of pruning. Should be large enouth such that overtraining is avoided");

 constructor for calculating BDT-MVA using previously generatad decision trees
 the result of the previous training (the decision trees) are read in via the
 weightfile. Make sure the "theVariables" correspond to the ones used in 
 creating the "weight"-file

 define the options (their key words) that can be set in the option string 
 know options:
 nTrees=Int_t:    number of trees in the forest to be created
 BoostType=       the boosting type for the trees in the forest (AdaBoost e.t.c..)
                  known: AdaBoost
                         Bagging
 SeparationType   the separation criterion applied in the node splitting
                  known: GiniIndex
                         MisClassificationError
                         CrossEntropy
                         SDivSqrtSPlusB
 nEventsMin:      the minimum number of events in a node (leaf criteria, stop splitting)
 nCuts:           the number of steps in the optimisation of the cut for a node
 UseYesNoLeaf     decide if the classification is done simply by the node type, or the S/B
                  (from the training) in the leaf node
 UseWeightedTrees use average classification from the trees, or have the individual trees
                  trees in the forest weighted (e.g. log(boostweight) from AdaBoost
 PruneMethod      The Pruning method: 
                  known: ExpectedError
                         CostComplexity 
                         CostComplexity2
 PruneStrength    a parameter to adjust the amount of pruning. Should be large enouth such that overtraining is avoided");

 the option string is decoded, for available options see "DeclareOptions"

 common initialisation with defaults for the BDT-Method

destructor

 write all Events from the Tree into a vector of TMVA::Events, that are 
 more easily manipulated.  
 This method should never be called without existing trainingTree, as it
 the vector of events from the ROOT training tree

 default sanity checks

 prune a tree adjusting the prunestrength using the "test sample" until
 the best efficiency on the test sample is obtained. In principle the
 test sample should not be used for that but rather a third validation
 sample, or the trainng sample with "cross validation". The latter is
 planned but will come later.

 test the tree quality.. in terms of Miscalssification
 

 apply the boosting alogrithim (the algorithm is selecte via the the "option" given
 in the constructor. The return value is the boosting weight 

 the AdaBoost implementation.
 a new training sample is generated by weighting 
 events that are misclassified by the decision tree. The weight
 applied is w = (1-err)/err or more general:
            w = ((1-err)/err)^beta
 where err is the fracthin of misclassified events in the tree ( <0.5 assuming
 demanding the that previous selection was better than random guessing)
 and "beta" beeing a free parameter (standard: beta = 1) that modifies the
 boosting.

 call it Bootstrapping, re-sampling or whatever you like, in the end it is nothing
 else but applying "random Weights" to each event.

 and save the Weights

 read variable names and min/max
 NOTE: the latter values are mandatory for the normalisation 
 in the reader application !!!

 return the MVA value (range [-1;1]) that classifies the
 event.according to the majority vote from the total number of
 decision trees
 In the literature I found that people actually use the 
 weighted majority vote (using the boost weights) .. However I
 did not see any improvement in doing so :(  
 --> this is currently switched off

 here we could write some histograms created during the processing
 to the output file.

 return the relative variable importance, normalized to all
 variables together having the importance 1. The importance in
 evaluated as the total separation-gain that this variable had in
 the decision trees (weighted by the number of events)

 returns the measure for the variable importance of variable "ivar"
 which is later used in GetVariableImportance() to calculat the
 relative variable importances

 computes ranking of input variables