TMVA Users Guide TMVAUsers
User Manual: Pdf
Open the PDF directly: View PDF
Page Count: 158 [warning: Documents this large are best viewed by clicking the View PDF Link!]
- Introduction
- TMVA Quick Start
- Using TMVA
- The TMVA Factory
- Specifying training and test data
- Negative event weights
- Defining input variables, spectators and targets
- Preparing the training and test data
- Booking MVA methods
- Help option for MVA booking
- Training the MVA methods
- Testing the MVA methods
- Evaluating the MVA methods
- Classification performance evaluation
- Regression performance evaluation
- Overtraining
- Other representations of MVA outputs for classification: probabilities and probability integral transformation (Rarity)
- ROOT macros to plot training, testing and evaluation results
- The TMVA Reader
- An alternative to the Reader: standalone C++ response classes
- The TMVA Factory
- Data Preprocessing
- Probability Density Functions – the PDF Class
- Optimisation and Fitting
- Boosting and Bagging
- The TMVA Methods
- Rectangular cut optimisation
- Projective likelihood estimator (PDE approach)
- Multidimensional likelihood estimator (PDE range-search approach)
- Likelihood estimator using self-adapting phase-space binning (PDE-Foam)
- k-Nearest Neighbour (k-NN) Classifier
- H-Matrix discriminant
- Fisher discriminants (linear discriminant analysis)
- Linear discriminant analysis (LD)
- Function discriminant analysis (FDA)
- Artificial Neural Networks (nonlinear discriminant analysis)
- Deep Neural Networks
- Support Vector Machine (SVM)
- Boosted Decision and Regression Trees
- Predictive learning via rule ensembles (RuleFit)
- The PyMVA Methods
- Combining MVA Methods
- Which MVA method should I use for my problem?
- TMVA implementation status summary for classification and regression
- Conclusions and Plans
- Acknowledgements
- More Classifier Booking Examples
- Bibliography
- Index