TMVA

TMVA version 4.1.0 is included in this root release. The most important new feature is the support for simulataneous classification of multiple output classes for several multi-variate methods. A lot of effort went into consolidation of the software, i.e. method performance and robustness, and framework stability. The changes with respect to ROOT 5.27 / TMVA 4.0.7 are in detail:

Framework

Multi-class support. The support of multiple output classes (i.e., more than a single background and signal class) has been enabled for these methods: MLP (NN), BDTG, FDA.

The multiclass functionality can be enabled with the Factory option "AnalysisType=multiclass". Training data is specified with an additional classname, e.g. via factory->AddTree(tree,"classname");. After the training a genetic algorithm is invoked to determine the best cuts for selecting a specific class, based on the figure of merit: purity*efficiency. TMVA comes with two examples in $ROOTSYS/tmva/test: TMVAMulticlass.C and TMVAMulticlassApplication.C
New TMVA event vector building. The code for splitting the input data into training and test samples for all classes and the mixing of those samples to one training and one test sample has been rewritten completely. The new code is more performant and has a clearer structure. This fixes several bugs which have been reported by some users of TMVA.
Code and performance test framework: A unit test framework for daily software and method performance validation has been implemented.

Methods

BDT Automatic parameter optimisation for building the tree architecture: The optimisation procedure uses the performance of the trained classifier on the "test sample" for finding the set of optimal parameters. Two different methods to traverse the parameter space are available (scanning, genetic algorithm). Currently parameter optimization is implemented only for these three parameters that influence the tree architectur: the maximum depth of a tree, MaxDepth, the minimum number of events in each node, NodeMinEvents, and the number of tress, NTrees.
Optimization can is invoked by calling factory->OptimizeAllMethods(); prior to the call factory->TrainAllMethods();.
Automated and configurable parameter optimization is soon to be enabled for all methods (for those parameters where optimization is applicable).
BDT node splitting: While Decision Trees typically have only univariate splits, in TMVA one can now also opt for multivariate splits that use a "Fisher Discriminant" (option: UseFisherCuts), built from all observables that show correlations larger than some threshold (MinLinCorrForFisher). The training will then test at each split a cut on this fisher discriminant in addition to all univariate cuts on the variables (or only on those variables that have not been used in the Fisher discriminant, option UseExcusiveVars). No obvious improvement betwen very simple decision trees after boosting has been observed so far, but only a limited number of studies has been performed concerning potiential benenfit of these simple multivariate splits.

Bug fixes

A problem in the BDTG has been fixed, leading to a much improved regression performance.
A problem in the TMVA::Reader has been fixed.
With the new test framework and the coverity checks of ROOT a number of bugs were discovered and fixed. They mainly concerned memory leaks, and did not affect the performance.