swh:1:snp:af87cd67498ef4fe47c76ed3e7caffe5b61facaf
Tip revision: b09d279a7caa57d7edcdf27e43631419b84ac278 authored by Unknown Author on 06 June 2007, 10:03:49 UTC
This commit was manufactured by cvs2svn to create tag 'v5-15-08-caf'.
This commit was manufactured by cvs2svn to create tag 'v5-15-08-caf'.
Tip revision: b09d279
README
Dear TMVA User,
this README contains three sections
+ Overview
+ How to run the example
+ The TMVA GUI
Overview
========
This directory contains two files that illustrate the usage of the
TMVA package to perform a MultiVariate Analysis. It also contains a
GUI for easy access to the training result.
An MV analysis is executed in two steps, the training of the MVA
methods, and the application of the best suited MVA method to your
dataset. These two steps are shown in the files TMVAnalysis.C and
TMVApplication.C, respectively.
The example uses some toy data (tmva_example.root) that comes with
the sourceforge distribution or is available on the ROOT web. The
example files contains a signal and a background tree with four
variables.
TMVAnalysis.C
-------------
TMVAnalysis.C explains the usage of the TMVA::Factory class, which
is the framework for the whole training process. The following steps
are taken:
1st: The factory is created (a name for the project is specified,
which will help with the bookkeeping of different training cycles)
2nd: The input data trees are made available to the factory ( one
can also use text files as input, also shown in this example ). The
sizes of the training and the testing data samples are also specified.
3rd: The variables that are selected to be used in the MVA are
declared to the factory
4th: The factory needs to transfer the input data into the local
data structure used for the training and testing
5th: Methods are than booked with the factory. For each method
number of options can be set that steer the training behavior. The
options are described in greater detail in the method classes and
can be looked at in the CVS web-viewer
6th: The training of all methods on the training data sample is run
by the factory. The training results for each method are stored in
the directory weights/
7th: The testing of all methods on the testing data sample is run by
the factory. The test results are stored in a root file (filename
specified in the Factory constructor)
8th: The performance of all methods is analyzed by the factory. The
performance results are stored in the root file (see last step)
TMVApplication.C
----------------
TMVApplication.C explains the usage of the TMVA::Reader class, which
is to be used to evaluate your data with the trained MVA
methods. The following steps are taken:
1st: The reader is created.
2nd: A set of local variables is created and declared to the reader.
3rd: The method(s) - which means the name of the weight files
created in the training process - are booked with the reader.
4th: Your event loop:
- The values of the local variables are set (either by assignment
of, as in the example, by setting the BranchAddresses of your
datatree)
- the reader is asked to evaluate the MVA, the result of which is
used to separate signal from background
How to run the example
======================
First, TMVAnalysis.C needs to be run to perform the training. You
can edit TMVAnalysis.C (boolean flags at the head of the file) to
enable (disable) the methods you would (not) like to use. Then run
> root TMVAnalysis.C
Add the end a file TMVA.root is written, and a GUI opens. Once
TMVA.root exists, the GUI can also be started via
> root TMVAGui.C
You can also run
> root TMVApplication.C
which evaluates the methods on the signal data from the toy
tmva_example.root and writes the result to TMVApp.root. However,
TMVApplication is more of a pedagogical example for the usage of
TMVA inside your own analysis framework.
The TMVA GUI
============
The GUI provides easy access to a large number of macros that plot
various distributions and correlations of the input data, the MVA
method output distributions and the performance comparison plot
(background rejection versus efficiency). It can be used once the
training is run, and is started with
> root TMVAGui.C
or
> root TMVA\(\"MyTMVA.root\"\)
in case you had directed your training output to a different root
file (2nd argument of the TMVA::Factory constructor)