TMVA Configuration Options Reference Reference version: TMVA-v3.9.5 TMVA-version @ ROOT

Reference for configuration options defined in the option string of a classifier booking, and for the definition of data sets used for training and testing (Factory). Table fields: Option: The option identifier in the option string (given, e.g., in "factory->BookMethod(...)" call). Array: Can the option be set individually for each input variable via the "[i]" tag, where "i" is the ith variable? Default value: Value used if option is not explicitly set in the configuration option string. Predefined values: Options can be categories of predefined values among which the user must choose. Description: Info about the option. Colour codes: Greenish rows: Options shared by all classifiers (through common base class). Bluish rows: Specific classifier options. Yellowish rows: Configuration options for minimiser (fitter) classes. Redish rows: Options for other configurable classes.

Available classifiers/classes: [Classifier::BDT] [Classifier::CFMlpANN] [Classifier::Cuts] [Classifier::FDA] [Classifier::Fisher] [Classifier::HMatrix] [Classifier::KNN] [Classifier::Likelihood] [Classifier::MLP] [Classifier::PDERS] [Classifier::RuleFit] [Classifier::SVM] [Classifier::TMlpANN] [Fitter_GA] [Fitter_MC] [Fitter_Minuit] [Fitter_SA] [Factory::PrepareForTrainingAndTesting] [Factory]

Configuration options for classifier architecture and tuning : Configuration options reference for classifier: BDT Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) NTrees No 400 − Number of trees in the forest BoostType No AdaBoost AdaBoost, Bagging Boosting type for the trees in the forest AdaBoostBeta No 1 − Parameter for AdaBoost algorithm UseRandomisedTrees No False − Choose at each node splitting a random set of variables UseNvars No 4 − Number of variables used if randomised Tree option is chosen UseWeightedTrees No True − Use weighted trees or simple average in classification from the forest SeparationType No GiniIndex − Separation criterion for node splitting UseYesNoLeaf No True − Use Sig or Bkg categories, or the purity=S/(S+B) as classification of the leaf node NodePurityLimit No 0.5 − In boosting/pruning, nodes with purity > NodePurityLimit are signal; background otherwise. nEventsMin No 37 − Minimum number of events required in a leaf node (default: max(20, N_train/(Nvar^2)/10) ) nCuts No 20 − Number of steps during node cut optimisation PruneStrength No 4.5 − Pruning strength PruneMethod No CostComplexity − Method used for pruning (removal) of statistically insignificant branches PruneBeforeBoost No False − Flag to prune the tree before applying boosting algorithm NoNegWeightsInTraining No False − Ignore negative event weights in the training process Configuration options for classifier architecture and tuning : Configuration options reference for classifier: CFMlpANN Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No True − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) NCycles No 500 − Number of training cycles HiddenLayers No N+1,N − Specification of hidden layer architecture (N stands for number of variables; any integers may also be used) Configuration options for classifier architecture and tuning : Configuration options reference for classifier: Cuts Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) FitMethod No SA GA, SA, MC, MCEvents, MINUIT, EventScan Minimization Method (GA, SA, and MC are the primary methods to be used; the others have been introduced for testing purposes and are depreciated) EffMethod No EffSel EffSel, EffPDF Selection Method CutRangeMin Yes -1 − Minimum of allowed cut range (set per variable) CutRangeMax Yes -1 − Maximum of allowed cut range (set per variable) VarProp Yes NotEnforced NotEnforced, FMax, FMin, FSmart, FVerySmart Categorisation of cuts Configuration options for classifier architecture and tuning : Configuration options reference for classifier: FDA Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) Formula No (0)+(1)*x0+(2)*x1+(3)*x2+(4)*x3 − The discrimination formula ParRanges No (-1,1);(-10,10);(-10,10);(-10,10);(-10,10) − Parameter ranges FitMethod No MINUIT MC, GA, SA, MINUIT Optimisation Method Converger No None None, MINUIT FitMethod uses Converger to improve result Configuration options for classifier architecture and tuning : Configuration options reference for classifier: Fisher Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 50 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 1 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No True − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) Method No Fisher Fisher, Mahalanobis Discrimination method Configuration options for classifier architecture and tuning : Configuration options reference for classifier: HMatrix Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No True − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) Configuration options for classifier architecture and tuning : Configuration options reference for classifier: KNN Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) nkNN No 400 − Number of k-nearest neighbors TreeOptDepth No 6 − Binary tree optimisation depth ScaleFrac No 0.8 − Fraction of events used for scaling UseKernel No False − Use polynomial kernel weight Trim No False − Use equal number of signal and background events Configuration options for classifier architecture and tuning : Configuration options reference for classifier: Likelihood Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) NSmooth No 1 − Number of smoothing iterations for the input histograms NSmoothSig Yes -1 − Number of smoothing iterations for the input histograms NSmoothBkg Yes -1 − Number of smoothing iterations for the input histograms NAvEvtPerBin No 50 − Average number of events per PDF bin NAvEvtPerBinSig Yes -1 − Average num of events per PDF bin and variable (signal) NAvEvtPerBinBkg Yes -1 − Average num of events per PDF bin and variable (background) TransformOutput No False − Transform likelihood output by inverse sigmoid function PDFInterpol Yes KDE Spline0, Spline1, Spline2, Spline3, Spline5, KDE Method of interpolating reference histograms (e.g. Spline2 or KDE) KDEtype No Gauss Gauss KDE kernel type (1=Gauss) KDEiter No Nonadaptive Nonadaptive, Adaptive Number of iterations (1=non-adaptive, 2=adaptive) KDEFineFactor No 1 − Fine tuning factor for Adaptive KDE: Factor to multyply the width of the kernel KDEborder No None None, Renorm, Mirror Border effects treatment (1=no treatment , 2=kernel renormalization, 3=sample mirroring) Configuration options for classifier architecture and tuning : Configuration options reference for classifier: MLP Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) NCycles No 200 − Number of training cycles HiddenLayers No N+1,N − Specification of hidden layer architecture (N stands for number of variables; any integers may also be used) NeuronType No tanh linear, sigmoid, tanh, radial Neuron activation function type NeuronInputType No sum sum, sqsum, abssum Neuron input function type TrainingMethod No BP BP, GA Train with Back-Propagation (BP - default) or Genetic Algorithm (GA - slower and worse) LearningRate No 0.02 − ANN learning rate parameter DecayRate No 0.01 − Decay rate for learning parameter TestRate No 5 − Test for overtraining performed at each #th epochs BPMode No sequential sequential, batch Back-propagation learning mode: sequential or batch BatchSize No -1 − Batch size: number of events/batch, only set if in Batch Mode, -1 for BatchSize=number_of_events Configuration options for classifier architecture and tuning : Configuration options reference for classifier: PDERS Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No PCA None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) VolumeRangeMode No Adaptive Unscaled, MinMax, RMS, Adaptive, kNN Method to determine volume size KernelEstimator No Gauss Box, Sphere, Teepee, Gauss, Sinc3, Sinc5, Sinc7, Sinc9, Sinc11, Lanczos2, Lanczos3, Lanczos5, Lanczos8, Trim Kernel estimation function DeltaFrac No 3 − nEventsMin/Max for minmax and rms volume range NEventsMin No 400 − nEventsMin for adaptive volume range NEventsMax No 600 − nEventsMax for adaptive volume range MaxVIterations No 150 − MaxVIterations for adaptive volume range InitialScale No 0.99 − InitialScale for adaptive volume range GaussSigma No 0.3 − Width (wrt volume size) of Gaussian kernel estimator NormTree No False − Normalize binary search tree Configuration options for classifier architecture and tuning : Configuration options reference for classifier: RuleFit Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) GDTau No -1 − Gradient-directed path: default fit cut-off GDTauPrec No 0.01 − Gradient-directed path: precision of tau GDStep No 0.01 − Gradient-directed path: step size GDNSteps No 10000 − Gradient-directed path: number of steps GDErrScale No 1.02 − Stop scan when error>scale*errmin GDPathEveFrac No 0.5 − Fraction of events used for the path search GDValidEveFrac No 0.5 − Fraction of events used for the validation fEventsMin No 0.01 − Minimum fraction of events in a splittable node fEventsMax No 0.5 − Maximum fraction of events in a splittable node nTrees No 20 − Number of trees in forest. ForestType No AdaBoost AdaBoost, Random Method to use for forest generation RuleMinDist No 0.001 − Minimum distance between rules MinImp No 0.001 − Minimum rule importance accepted Model No ModRuleLinear ModRule, ModRuleLinear, ModLinear Model to be used RuleFitModule No RFTMVA RFTMVA, RFFriedman Which RuleFit module to use RFWorkDir No ./rulefit − Friedmans RuleFit module: working dir RFNrules No 2000 − Friedmans RuleFit module: maximum number of rules RFNendnodes No 4 − Friedmans RuleFit module: average number of end nodes Configuration options for classifier architecture and tuning : Configuration options reference for classifier: SVM Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No True − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) C No 1 − C parameter Tol No 0.001 − Tolerance parameter MaxIter No 1000 − Maximum number of training loops Sigma No 2 − Kernel parameter: sigma Order No 3 − Polynomial Kernel parameter: polynomial order Theta No 1 − Sigmoid Kernel parameter: theta Kappa No 1 − Sigmoid Kernel parameter: kappa Kernel No Gauss Linear, Gauss, Polynomial, Sigmoid Uses kernel function Configuration options for classifier architecture and tuning : Configuration options reference for classifier: TMlpANN Option Array Default value Predefined values Description D No False − Use-decorrelated-variables flag (depreciated) Normalise No False − Normalise input variables VarTransform No None None, Decorrelate, PCA, GaussDecorr Variable transformation method VarTransformType No Signal Signal, Background Use signal or background events to derive for variable transformation (the transformation is applied on both types of, course) NbinsMVAPdf No 60 − Number of bins used for the PDFs of classifier outputs NsmoothMVAPdf No 2 − Number of smoothing iterations for classifier PDFs V No False − Verbose mode VerboseLevel No Default Default, Debug, Verbose, Info, Warning, Error, Fatal Verbosity level H No True − Print classifier-specific help message CreateMVAPdfs No False − Create PDFs for classifier outputs (signal and background) TxtWeightFilesOnly No True − If True: write all training results (weights) as text files (False: some are written in ROOT format) NCycles No 200 − Number of training cycles HiddenLayers No N+1,N − Specification of hidden layer architecture (N stands for number of variables; any integers may also be used) ValidationFraction No 0.3 − Fraction of events in training tree used for cross validation LearningMethod No BFGS Stochastic, Batch, SteepestDescent, RibierePolak, FletcherReeves, BFGS Learning method Confugration options for setup and tuning of specific fitter : Configuration options reference for fitting method: Genetic Algorithm (GA) Option Array Default value Predefined values Description PopSize No 100 − Population size for GA Steps No 20 − Number of steps for convergence Cycles No 3 − Independent cycles of GA fitting SC_steps No 10 − Spread control, steps SC_rate No 5 − Spread control, rate: factor is changed depending on the rate SC_factor No 0.95 − Spread control, factor ConvCrit No 0.001 − Convergence criteria SaveBestGen No 0 − Saves the best n results from each generation; these are included in the last cycle SaveBestCycle No 10 − Saves the best n results from each cycle; these are included in the last cycle Trim No True − Trim the population to PopSize after assessing the fitness of each individual Seed No 100 − Set seed of random generator (0 gives random seeds) Confugration options for setup and tuning of specific fitter : Configuration options reference for fitting method: Monte Carlo sampling (MC) Option Array Default value Predefined values Description SampleSize No 100000 − Number of Monte Carlo events in toy sample Sigma No 0.1 − If > 0: new points are generated according to Gauss around best value and with Sigma in units of interval length Seed No 100 − Seed for the random generator (0 takes random seeds) Confugration options for setup and tuning of specific fitter : Configuration options reference for fitting method: TMinuit (MT) Option Array Default value Predefined values Description ErrorLevel No 1 − TMinuit: error level: 0.5=logL fit, 1=chi-squared fit PrintLevel No -1 − TMinuit: output level: -1=least, 0, +1=all garbage FitStrategy