From 2da6ec40ba8965b75e009a1709779e0407609b9a Mon Sep 17 00:00:00 2001
From: Joerg Stelzer <joerg.stelzer@cern.ch>
Date: Mon, 27 Jun 2011 13:19:22 +0000
Subject: [PATCH] fix event requests for Category classifier
git-svn-id: http://root.cern.ch/svn/root/trunk@39989 27541ba8-7e3a-0410-8455-c3a389f83636
---
tmva/doc/LICENSE | 2 +-
tmva/doc/README | 70 +--
tmva/inc/TMVA/DataSetFactory.h | 107 ++--
tmva/inc/TMVA/DataSetInfo.h | 4 +-
tmva/inc/TMVA/MethodBase.h | 51 +-
tmva/inc/TMVA/SVKernelFunction.h | 0
tmva/inc/TMVA/SVKernelMatrix.h | 0
tmva/src/CrossEntropy.cxx | 1 -
tmva/src/DataSetFactory.cxx | 720 ++++++++++++-----------
tmva/src/DataSetInfo.cxx | 84 +--
tmva/src/DataSetManager.cxx | 10 +-
tmva/src/Factory.cxx | 4 +-
tmva/src/MethodBoost.cxx | 2 +
tmva/src/MethodCategory.cxx | 46 +-
tmva/src/MethodCompositeBase.cxx | 90 +--
tmva/src/PDEFoamVect.cxx | 8 +-
tmva/src/RegressionVariance.cxx | 2 +-
tmva/src/VariableNormalizeTransform.cxx | 1 +
tmva/test/TMVAClassificationCategory.cxx | 34 +-
tmva/test/setup.sh | 3 +
20 files changed, 660 insertions(+), 579 deletions(-)
mode change 100644 => 100755 tmva/inc/TMVA/SVKernelFunction.h
mode change 100644 => 100755 tmva/inc/TMVA/SVKernelMatrix.h
diff --git a/tmva/doc/LICENSE b/tmva/doc/LICENSE
index e555484b511..40a9b19fd4e 100644
--- a/tmva/doc/LICENSE
+++ b/tmva/doc/LICENSE
@@ -1,6 +1,6 @@
TMVA -- Toolkit for Multivariate Data Analysis
-Copyright (c) 2005-2009, Regents of CERN (Switzerland), the
+Copyright (c) 2005-20010, Regents of CERN (Switzerland), the
University of Victoria (Canada), the MPI fuer Kernphysik Heidelberg
(Germany), LAPP (France), the University of Bonn (Germany).
diff --git a/tmva/doc/README b/tmva/doc/README
index 0df2207c9d3..68c92e81a10 100644
--- a/tmva/doc/README
+++ b/tmva/doc/README
@@ -4,7 +4,7 @@
TMVA Users Guide : http://tmva.sourceforge.net/docu/TMVAUsersGuide.pdf
TMVA home page : http://tmva.sourceforge.net/
-TMVA developer page : http://sourceforge.net/projects/tmva
+TMVA download page : http://sourceforge.net/projects/tmva
TMVA mailing list : http://sourceforge.net/mailarchive/forum.php?forum_name=tmva-users
TMVA license (BSD) : http://tmva.sourceforge.net/LICENSE
@@ -15,7 +15,7 @@ System requirements:
TMVA has been tested to run on Linux, MAC/OSX and Windows platforms.
- Running TMVA requires the availability of ROOT shared libraries with ROOT_VERSION >= 5.08
+ Running TMVA requires the availability of ROOT shared libraries with ROOT_VERSION >= 5.14
(type "root-config --version" to see the version installed)
========================================================================================
@@ -26,64 +26,58 @@ Getting Started:
How to compile the code:
------------------------
/home> cd TMVA
- /home/TMVA> source setup.[c]sh // includes TMVA/lib in your lib path
- /home/TMVA> cd src
- /home/TMVA/src> make // compile and build the library ../libTMVA.1.so
+ /home/TMVA> make # compile and build the library lib/libTMVA.1.so
- How to run the code as ROOT macro: // training/testing of an academic example
+ How to run the code as ROOT macro: # training/testing of an academic example
----------------------------------
- /home/TMVA> cd macros
+ /home/TMVA> cd test
+ /home/TMVA/test> source setup.sh # setup script must be exectuted before running macros (use setup.csh for c-shell)
--- For classification:
- /home/TMVA/macros> root -l TMVAClassification.C // run all standard classifiers
- /home/TMVA/macros> root -l TMVAClassification.C\(\"LD,Likelihood\"\) // run LD and Likelihood classifiers
+ /home/TMVA/test> root -l TMVAClassification.C # run all standard classifiers (takes a while)
+ /home/TMVA/test> root -l TMVAClassification.C\(\"LD,Likelihood\"\) # run LD and Likelihood classifiers
--- For regression:
- /home/TMVA/macros> root -l TMVARegression.C // run all regression algorithms
- /home/TMVA/macros> root -l TMVARegression.C\(\"LD,KNN\"\) // run LD and k-NN regression algorithms
+ /home/TMVA/test> root -l TMVARegression.C # run all regression algorithms n(takes a while)
+ /home/TMVA/test> root -l TMVARegression.C\(\"LD,KNN\"\) # run LD and k-NN regression algorithms
--> at the end of the jobs, a GUI will pop up: try to click through all the buttons;
some of the lower buttons are method-specific, and will only work when the
corresponding classifiers/regression algorithms have been trained/tested before
(unless they are greyed out)
- How to run the code as an executable: // training/testing of an academic example
+ How to run the code as an executable:
-------------------------------------
- /home/TMVA> cd execs
- /home/TMVA/execs> make
- /home/TMVA/execs> ./TMVAClassification // run all standard classifiers
- /home/TMVA/execs> ./TMVAClassification LD Likelihood // run LD and Likelihood classifiers
+ /home/TMVA/test> make
+ /home/TMVA/test> ./TMVAClassification # run all standard classifiers
+ /home/TMVA/test> ./TMVAClassification LD Likelihood # run LD and Likelihood classifiers
- ... and similar for regression
+ ... and similarly for regression
- /home/TMVA/examples> root -l ../macros/TMVAGui.C // start the GUI
+ /home/TMVA/test> root -l TMVAGui.C # start the GUI
+
+ How to run the code as an python script using PyROOT:
+ -------------------------------------
+ /home/TMVA/test> make
+ /home/TMVA/test> python ./TMVAClassification.py --method LD,Likelihood
How to apply the TMVA methods:
-------------------------------------
- /home/TMVA> cd macros
+ /home/TMVA> cd test
--- For classification:
- /home/TMVA/macros> root -l TMVAClassificationApplication.C
- /home/TMVA/macros> root -l TMVAClassificationApplication.C\(\"LD,Likelihood\"\)
+ /home/TMVA/test> root -l TMVAClassificationApplication.C
+ /home/TMVA/test> root -l TMVAClassificationApplication.C\(\"LD,Likelihood\"\)
- ... and similar for regression
+ ... and similar for regression.
+ ... and similar for executables.
The directory structure:
------------------------
- src/ : the TMVA source code
+ inc/ : the TMVA class headers
+ src/ : the TMVA class source
lib/ : here you'll find the TMVA library (libTMVA.1.so) after compilation
- (copy it to you preferred library directory or include
- this directory in your LD_LIBRARY_PATH as it is done
- by: source setup.[c]sh
- macros/ : example code of how to use the TMVA library with a ROOT macro
- uses input data from a Toy Monte Carlo;
- also: handy root macros which read and display the
- results produced by TMVAClassification and TMVARegression
- execs/ : same example code as in 'macros', but for using the TMVA library in an executable
- execs/data : the Toy Monte Carlo data
- python/ : example code of how to use the TMVA library with a python script;
- requires availability of PyROOT
- development/ : for use by developers only
+ test/ : example code for analysis macros and executables, the GUI and analysis scripts.
========================================================================================
@@ -158,11 +152,11 @@ Please report any problems and/or suggestions for improvements to the authors.
========================================================================================
-Copyright 漏 (2005-2009):
+Copyright 漏 (2005-2010):
------------------------
- Andreas Hoecker, Peter Speckmayer, J枚rg Stelzer (all: CERN, Switzerland),
- Jan Therhaag, Eckhard von Toerne (both: U. Bonn, Germany),
+ Andreas Hoecker, Peter Speckmayer, J枚rg Stelzer (CERN, Switzerland),
+ Jan Therhaag, Eckhard von Toerne (U. Bonn, Germany),
Helge Voss (MPI-KP Heidelberg, Germany),
Contributed to TMVA have, please see: http://tmva.sourceforge.net/#authors
diff --git a/tmva/inc/TMVA/DataSetFactory.h b/tmva/inc/TMVA/DataSetFactory.h
index 04856343b55..bc92730269f 100644
--- a/tmva/inc/TMVA/DataSetFactory.h
+++ b/tmva/inc/TMVA/DataSetFactory.h
@@ -73,24 +73,13 @@
#endif
namespace TMVA {
-
+
class DataSet;
class DataSetInfo;
class DataInputHandler;
class TreeInfo;
class MsgLogger;
- typedef std::vector< Event* > EventVector;
- typedef std::vector< EventVector > EventVectorOfClasses;
- typedef std::map<Types::ETreeType, EventVectorOfClasses > EventVectorOfClassesOfTreeType;
- typedef std::map<Types::ETreeType, EventVector > EventVectorOfTreeType;
-
- typedef std::vector< Double_t > ValuePerClass;
- typedef std::map<Types::ETreeType, ValuePerClass > ValuePerClassOfTreeType;
-
- typedef std::vector< Int_t > NumberPerClass;
- typedef std::map<Types::ETreeType, NumberPerClass > NumberPerClassOfTreeType;
-
// =============== maybe move these elswhere (e.g. into the tools )
// =============== functors =======================
@@ -139,7 +128,7 @@ namespace TMVA {
template <typename F>
- class null_t
+ class null_t
{
private:
// returns argF
@@ -157,7 +146,7 @@ namespace TMVA {
}
-
+
template <typename F, typename G, typename H>
class compose_binary_t : public std::binary_function<typename G::argument_type,
typename H::argument_type,
@@ -171,7 +160,7 @@ namespace TMVA {
compose_binary_t(const F& _f, const G& _g, const H& _h) : f(_f), g(_g), h(_h)
{
}
-
+
typename F::result_type operator()(const typename G::argument_type& argG,
const typename H::argument_type& argH) const
{
@@ -218,12 +207,49 @@ namespace TMVA {
class DataSetFactory {
+ typedef std::vector< Event* > EventVector;
+ typedef std::vector< EventVector > EventVectorOfClasses;
+ typedef std::map<Types::ETreeType, EventVectorOfClasses > EventVectorOfClassesOfTreeType;
+ typedef std::map<Types::ETreeType, EventVector > EventVectorOfTreeType;
+
+ typedef std::vector< Double_t > ValuePerClass;
+ typedef std::map<Types::ETreeType, ValuePerClass > ValuePerClassOfTreeType;
+
+ class EventStats {
+ public:
+ Int_t nTrainingEventsRequested;
+ Int_t nTestingEventsRequested;
+ Int_t nInitialEvents;
+ Int_t nEvBeforeCut;
+ Int_t nEvAfterCut;
+ Float_t nWeEvBeforeCut;
+ Float_t nWeEvAfterCut;
+ Double_t nNegWeights;
+ Float_t* varAvLength;
+ EventStats():
+ nTrainingEventsRequested(0),
+ nTestingEventsRequested(0),
+ nInitialEvents(0),
+ nEvBeforeCut(0),
+ nEvAfterCut(0),
+ nWeEvBeforeCut(0),
+ nWeEvAfterCut(0),
+ nNegWeights(0),
+ varAvLength(0)
+ {}
+ ~EventStats() { delete[] varAvLength; }
+ Float_t cutScaling() const { return Float_t(nEvAfterCut)/nEvBeforeCut; }
+ };
+
+ typedef std::vector< int > NumberPerClass;
+ typedef std::vector< EventStats > EvtStatsPerClass;
+
public:
// singleton class
- static DataSetFactory& Instance() { if (!fgInstance) fgInstance = new DataSetFactory(); return *fgInstance; }
+ static DataSetFactory& Instance() { if (!fgInstance) fgInstance = new DataSetFactory(); return *fgInstance; }
static void destroyInstance() { if (fgInstance) { delete fgInstance; fgInstance=0; } }
DataSet* CreateDataSet( DataSetInfo &, DataInputHandler& );
@@ -231,34 +257,37 @@ namespace TMVA {
protected:
~DataSetFactory();
-
+
DataSetFactory();
static DataSetFactory *fgInstance;
DataSet* BuildInitialDataSet( DataSetInfo&, TMVA::DataInputHandler& );
DataSet* BuildDynamicDataSet( DataSetInfo& );
-
+
// ---------- new versions
- void BuildEventVector ( DataSetInfo& dsi,
- DataInputHandler& dataInput,
- EventVectorOfClassesOfTreeType& tmpEventVector);
-
- DataSet* MixEvents ( DataSetInfo& dsi,
- EventVectorOfClassesOfTreeType& tmpEventVector,
- NumberPerClassOfTreeType& nTrainTestEvents,
- const TString& splitMode,
- const TString& mixMode,
- const TString& normMode,
- UInt_t splitSeed);
-
- void RenormEvents ( DataSetInfo& dsi,
- EventVectorOfClassesOfTreeType& tmpEventVector,
- const TString& normMode );
-
- void InitOptions ( DataSetInfo& dsi,
- NumberPerClassOfTreeType& nTrainTestEvents,
- TString& normMode, UInt_t& splitSeed, TString& splitMode, TString& mixMode );
-
+ void BuildEventVector ( DataSetInfo& dsi,
+ DataInputHandler& dataInput,
+ EventVectorOfClassesOfTreeType& eventsmap,
+ EvtStatsPerClass& eventCounts);
+
+ DataSet* MixEvents ( DataSetInfo& dsi,
+ EventVectorOfClassesOfTreeType& eventsmap,
+ EvtStatsPerClass& eventCounts,
+ const TString& splitMode,
+ const TString& mixMode,
+ const TString& normMode,
+ UInt_t splitSeed);
+
+ void RenormEvents ( DataSetInfo& dsi,
+ EventVectorOfClassesOfTreeType& eventsmap,
+ const EvtStatsPerClass& eventCounts,
+ const TString& normMode );
+
+ void InitOptions ( DataSetInfo& dsi,
+ EvtStatsPerClass& eventsmap,
+ TString& normMode, UInt_t& splitSeed,
+ TString& splitMode, TString& mixMode );
+
// ------------------------
@@ -282,7 +311,7 @@ namespace TMVA {
Bool_t fVerbose; //! Verbosity
TString fVerboseLevel; //! VerboseLevel
- // the event
+ // the event
mutable TTree* fCurrentTree; //! the tree, events are currently read from
mutable UInt_t fCurrentEvtIdx; //! the current event (to avoid reading of the same event)
diff --git a/tmva/inc/TMVA/DataSetInfo.h b/tmva/inc/TMVA/DataSetInfo.h
index 776d8993565..2d4038c3480 100644
--- a/tmva/inc/TMVA/DataSetInfo.h
+++ b/tmva/inc/TMVA/DataSetInfo.h
@@ -14,7 +14,7 @@
* Peter Speckmayer <speckmay@mail.cern.ch> - CERN, Switzerland *
* Joerg Stelzer <Joerg.Stelzer@cern.ch> - DESY, Germany *
* *
- * Copyright (c) 2008: *
+ * Copyright (c) 2008-2011: *
* CERN, Switzerland *
* MPI-K Heidelberg, Germany *
* DESY Hamburg, Germany *
@@ -157,7 +157,7 @@ namespace TMVA {
std::vector<TString> GetListOfVariables() const;
- // correlation matrix
+ // correlation matrix
const TMatrixD* CorrelationMatrix ( const TString& className ) const;
void SetCorrelationMatrix ( const TString& className, TMatrixD* matrix );
void PrintCorrelationMatrix( const TString& className );
diff --git a/tmva/inc/TMVA/MethodBase.h b/tmva/inc/TMVA/MethodBase.h
index 58eb36cf9de..5c43369b066 100644
--- a/tmva/inc/TMVA/MethodBase.h
+++ b/tmva/inc/TMVA/MethodBase.h
@@ -341,14 +341,14 @@ namespace TMVA {
TString GetTrainingTMVAVersionString() const;
TString GetTrainingROOTVersionString() const;
- TransformationHandler& GetTransformationHandler(Bool_t takeReroutedIfAvailable=true)
- {
- if(fTransformationPointer && takeReroutedIfAvailable) return *fTransformationPointer; else return fTransformation;
- }
- const TransformationHandler& GetTransformationHandler(Bool_t takeReroutedIfAvailable=true) const
- {
- if(fTransformationPointer && takeReroutedIfAvailable) return *fTransformationPointer; else return fTransformation;
- }
+ TransformationHandler& GetTransformationHandler(Bool_t takeReroutedIfAvailable=true)
+ {
+ if(fTransformationPointer && takeReroutedIfAvailable) return *fTransformationPointer; else return fTransformation;
+ }
+ const TransformationHandler& GetTransformationHandler(Bool_t takeReroutedIfAvailable=true) const
+ {
+ if(fTransformationPointer && takeReroutedIfAvailable) return *fTransformationPointer; else return fTransformation;
+ }
void RerouteTransformationHandler (TransformationHandler* fTargetTransformation) { fTransformationPointer=fTargetTransformation; }
@@ -413,7 +413,7 @@ namespace TMVA {
// ---------- protected event and tree accessors -----------------------------
- // names of input variables (if the original names are expressions, they are
+ // names of input variables (if the original names are expressions, they are
// transformed into regexps)
const TString& GetInternalVarName( Int_t ivar ) const { return (*fInputVars)[ivar]; }
const TString& GetOriginalVarName( Int_t ivar ) const { return DataInfo().GetVariableInfo(ivar).GetExpression(); }
@@ -435,18 +435,18 @@ namespace TMVA {
// some basic statistical analysis
void Statistics( Types::ETreeType treeType, const TString& theVarName,
- Double_t&, Double_t&, Double_t&,
+ Double_t&, Double_t&, Double_t&,
Double_t&, Double_t&, Double_t& );
- // if TRUE, write weights only to text files
+ // if TRUE, write weights only to text files
Bool_t TxtWeightsOnly() const { return kTRUE; }
protected:
-
+
// access to event information that needs method-specific information
-
- Float_t GetTWeight( const Event* ev ) const {
- return (fIgnoreNegWeightsInTraining && (ev->GetWeight() < 0)) ? 0. : ev->GetWeight();
+
+ Float_t GetTWeight( const Event* ev ) const {
+ return (fIgnoreNegWeightsInTraining && (ev->GetWeight() < 0)) ? 0. : ev->GetWeight();
}
Bool_t IsConstructedFromWeightFile() const { return fConstructedFromWeightFile; }
@@ -479,13 +479,13 @@ namespace TMVA {
// PDFs for classifier response (required to compute signal probability and Rarity)
void CreateMVAPdfs();
- // for root finder
+ // for root finder
static Double_t IGetEffForRoot( Double_t ); // interface
Double_t GetEffForRoot ( Double_t ); // implementation
// used for file parsing
Bool_t GetLine( std::istream& fin, char * buf );
-
+
// fill test tree with classification or regression results
virtual void AddClassifierOutput ( Types::ETreeType type );
virtual void AddClassifierOutputProb( Types::ETreeType type );
@@ -494,12 +494,13 @@ namespace TMVA {
private:
- void AddInfoItem( void* gi, const TString& name, const TString& value) const;
+ void AddInfoItem( void* gi, const TString& name,
+ const TString& value) const;
- static void CreateVariableTransforms(const TString& trafoDefinition,
- TMVA::DataSetInfo& dataInfo,
- TMVA::TransformationHandler& transformationHandler,
- TMVA::MsgLogger& log );
+ static void CreateVariableTransforms(const TString& trafoDefinition,
+ TMVA::DataSetInfo& dataInfo,
+ TMVA::TransformationHandler& transformationHandler,
+ TMVA::MsgLogger& log );
// ========== class members ==================================================
@@ -523,7 +524,7 @@ namespace TMVA {
private:
// MethodCuts redefines some of the evaluation variables and histograms -> must access private members
- friend class MethodCuts;
+ friend class MethodCuts;
Bool_t fDisableWriting; //! set to true in order to suppress writing to XML
@@ -537,11 +538,11 @@ namespace TMVA {
// naming and versioning
TString fJobName; // name of job -> user defined, appears in weight files
TString fMethodName; // name of the method (set in derived class)
- Types::EMVA fMethodType; // type of method (set in derived class)
+ Types::EMVA fMethodType; // type of method (set in derived class)
TString fTestvar; // variable used in evaluation, etc (mostly the MVA)
UInt_t fTMVATrainingVersion; // TMVA version used for training
UInt_t fROOTTrainingVersion; // ROOT version used for training
- Bool_t fConstructedFromWeightFile; // is it obtained from weight file?
+ Bool_t fConstructedFromWeightFile; // is it obtained from weight file?
// Directory structure: fMethodBaseDir/fBaseDir
// where the first directory name is defined by the method type
diff --git a/tmva/inc/TMVA/SVKernelFunction.h b/tmva/inc/TMVA/SVKernelFunction.h
old mode 100644
new mode 100755
diff --git a/tmva/inc/TMVA/SVKernelMatrix.h b/tmva/inc/TMVA/SVKernelMatrix.h
old mode 100644
new mode 100755
diff --git a/tmva/src/CrossEntropy.cxx b/tmva/src/CrossEntropy.cxx
index e7885cdc684..e886ff381a3 100644
--- a/tmva/src/CrossEntropy.cxx
+++ b/tmva/src/CrossEntropy.cxx
@@ -32,7 +32,6 @@
// -p log (p) - (1-p)log(1-p); p=purity
//_______________________________________________________________________
-//#include <math.h>
#include "TMath.h"
#include "TMVA/CrossEntropy.h"
diff --git a/tmva/src/DataSetFactory.cxx b/tmva/src/DataSetFactory.cxx
index 74b629c8d7a..68cdc6f5f8f 100644
--- a/tmva/src/DataSetFactory.cxx
+++ b/tmva/src/DataSetFactory.cxx
@@ -1,30 +1,30 @@
// @(#)root/tmva $Id$
// Author: Andreas Hoecker, Peter Speckmayer, Joerg Stelzer, Eckhard von Toerne, Helge Voss
-/**********************************************************************************
- * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
- * Package: TMVA *
- * Class : DataSetFactory *
- * Web : http://tmva.sourceforge.net *
- * *
- * Description: *
- * Implementation (see header for description) *
- * *
- * Authors (alphabetical): *
- * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
- * Peter Speckmayer <Peter.Speckmayer@cern.ch> - CERN, Switzerland *
- * Joerg Stelzer <Joerg.Stelzer@cern.ch> - CERN, Switzerland *
- * Eckhard von Toerne <evt@physik.uni-bonn.de> - U. of Bonn, Germany *
- * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
- * *
- * Copyright (c) 2009: *
- * CERN, Switzerland *
- * MPI-K Heidelberg, Germany *
- * U. of Bonn, Germany *
- * Redistribution and use in source and binary forms, with or without *
- * modification, are permitted according to the terms listed in LICENSE *
- * (http://tmva.sourceforge.net/LICENSE) *
- **********************************************************************************/
+/*****************************************************************************
+ * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
+ * Package: TMVA *
+ * Class : DataSetFactory *
+ * Web : http://tmva.sourceforge.net *
+ * *
+ * Description: *
+ * Implementation (see header for description) *
+ * *
+ * Authors (alphabetical): *
+ * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
+ * Peter Speckmayer <Peter.Speckmayer@cern.ch> - CERN, Switzerland *
+ * Joerg Stelzer <Joerg.Stelzer@cern.ch> - MSU, USA *
+ * Eckhard von Toerne <evt@physik.uni-bonn.de> - U. of Bonn, Germany *
+ * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
+ * *
+ * Copyright (c) 2009: *
+ * CERN, Switzerland *
+ * MPI-K Heidelberg, Germany *
+ * U. of Bonn, Germany *
+ * Redistribution and use in source and binary forms, with or without *
+ * modification, are permitted according to the terms listed in LICENSE *
+ * (http://tmva.sourceforge.net/LICENSE) *
+ *****************************************************************************/
#include <assert.h>
@@ -85,7 +85,7 @@ TMVA::DataSetFactory* TMVA::DataSetFactory::fgInstance = 0;
namespace TMVA {
// calculate the largest common divider
// this function is not happy if numbers are negative!
- Int_t LargestCommonDivider(Int_t a, Int_t b)
+ Int_t LargestCommonDivider(Int_t a, Int_t b)
{
if (a<b) {Int_t tmp = a; a=b; b=tmp; } // achieve a>=b
if (b==0) return a;
@@ -107,7 +107,7 @@ TMVA::DataSetFactory::DataSetFactory() :
}
//_______________________________________________________________________
-TMVA::DataSetFactory::~DataSetFactory()
+TMVA::DataSetFactory::~DataSetFactory()
{
// destructor
std::vector<TTreeFormula*>::const_iterator formIt;
@@ -122,7 +122,8 @@ TMVA::DataSetFactory::~DataSetFactory()
}
//_______________________________________________________________________
-TMVA::DataSet* TMVA::DataSetFactory::CreateDataSet( TMVA::DataSetInfo& dsi, TMVA::DataInputHandler& dataInput )
+TMVA::DataSet* TMVA::DataSetFactory::CreateDataSet( TMVA::DataSetInfo& dsi,
+ TMVA::DataInputHandler& dataInput )
{
// steering the creation of a new dataset
@@ -131,7 +132,7 @@ TMVA::DataSet* TMVA::DataSetFactory::CreateDataSet( TMVA::DataSetInfo& dsi, TMVA
if (ds->GetNEvents() > 1) {
CalcMinMax(ds,dsi);
-
+
// from the the final dataset build the correlation matrix
for (UInt_t cl = 0; cl< dsi.GetNClasses(); cl++) {
const TString className = dsi.GetClassInfo(cl)->GetName();
@@ -144,17 +145,18 @@ TMVA::DataSet* TMVA::DataSetFactory::CreateDataSet( TMVA::DataSetInfo& dsi, TMVA
}
//_______________________________________________________________________
-TMVA::DataSet* TMVA::DataSetFactory::BuildDynamicDataSet( TMVA::DataSetInfo& dsi )
+TMVA::DataSet* TMVA::DataSetFactory::BuildDynamicDataSet( TMVA::DataSetInfo& dsi )
{
Log() << kDEBUG << "Build DataSet consisting of one Event with dynamically changing variables" << Endl;
DataSet* ds = new DataSet(dsi);
- // create a DataSet with one Event which uses dynamic variables (pointers to variables)
+ // create a DataSet with one Event which uses dynamic variables
+ // (pointers to variables)
if(dsi.GetNClasses()==0){
dsi.AddClass( "data" );
dsi.GetClassInfo( "data" )->SetNumber(0);
}
-
+
std::vector<Float_t*>* evdyn = new std::vector<Float_t*>(0);
std::vector<VariableInfo>& varinfos = dsi.GetVariableInfos();
@@ -166,7 +168,7 @@ TMVA::DataSet* TMVA::DataSetFactory::BuildDynamicDataSet( TMVA::DataSetInfo& dsi
for (;it!=itEnd;++it) {
Float_t* external=(Float_t*)(*it).GetExternalLink();
if (external==0)
- Log() << kDEBUG << "The link to the external variable is NULL while I am trying to build a dynamic data set. In this case fTmpEvent from MethodBase HAS TO BE USED in the method to get useful values in variables." << Endl;
+ Log() << kDEBUG << "The link to the external variable is NULL while I am trying to build a dynamic data set. In this case fTmpEvent from MethodBase HAS TO BE USED in the method to get useful values in variables." << Endl;
evdyn->push_back (external);
}
@@ -186,11 +188,14 @@ TMVA::DataSet* TMVA::DataSetFactory::BuildDynamicDataSet( TMVA::DataSetInfo& dsi
}
//_______________________________________________________________________
-TMVA::DataSet* TMVA::DataSetFactory::BuildInitialDataSet( DataSetInfo& dsi, DataInputHandler& dataInput )
+TMVA::DataSet*
+TMVA::DataSetFactory::BuildInitialDataSet( DataSetInfo& dsi,
+ DataInputHandler& dataInput )
{
- // if no entries, than create a DataSet with one Event which uses dynamic variables (pointers to variables)
+ // if no entries, than create a DataSet with one Event which uses
+ // dynamic variables (pointers to variables)
if (dataInput.GetEntries()==0) return BuildDynamicDataSet( dsi );
- // ------------------------------------------------------------------------------------
+ // -------------------------------------------------------------------------
// register the classes in the datasetinfo-object
// information comes from the trees in the dataInputHandler-object
@@ -200,32 +205,31 @@ TMVA::DataSet* TMVA::DataSetFactory::BuildInitialDataSet( DataSetInfo& dsi, Data
}
delete classList;
+ EvtStatsPerClass eventCounts(dsi.GetNClasses());
TString normMode;
TString splitMode;
TString mixMode;
- UInt_t splitSeed;
+ UInt_t splitSeed;
+ InitOptions( dsi, eventCounts, normMode, splitSeed, splitMode , mixMode );
- // ======= build event-vector tentative new ordering =================================
-
- TMVA::EventVectorOfClassesOfTreeType tmpEventVector;
- TMVA::NumberPerClassOfTreeType nTrainTestEvents;
+ // ======= build event-vector from input, apply preselection ===============
+ EventVectorOfClassesOfTreeType tmpEventVector;
+ BuildEventVector( dsi, dataInput, tmpEventVector, eventCounts );
- InitOptions ( dsi, nTrainTestEvents, normMode, splitSeed, splitMode , mixMode );
- BuildEventVector( dsi, dataInput, tmpEventVector );
-
- DataSet* ds = MixEvents( dsi, tmpEventVector, nTrainTestEvents, splitMode, mixMode, normMode, splitSeed);
+ DataSet* ds = MixEvents( dsi, tmpEventVector, eventCounts,
+ splitMode, mixMode, normMode, splitSeed);
const Bool_t showCollectedOutput = kFALSE;
if (showCollectedOutput) {
Int_t maxL = dsi.GetClassNameMaxLength();
Log() << kINFO << "Collected:" << Endl;
for (UInt_t cl = 0; cl < dsi.GetNClasses(); cl++) {
- Log() << kINFO << " "
- << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ Log() << kINFO << " "
+ << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
<< " training entries: " << ds->GetNClassEvents( 0, cl ) << Endl;
- Log() << kINFO << " "
- << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
- << " testing entries: " << ds->GetNClassEvents( 1, cl ) << Endl;
+ Log() << kINFO << " "
+ << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ << " testing entries: " << ds->GetNClassEvents( 1, cl ) << Endl;
}
Log() << kINFO << " " << Endl;
}
@@ -234,19 +238,18 @@ TMVA::DataSet* TMVA::DataSetFactory::BuildInitialDataSet( DataSetInfo& dsi, Data
}
//_______________________________________________________________________
-Bool_t TMVA::DataSetFactory::CheckTTreeFormula( TTreeFormula* ttf, const TString& expression, Bool_t& hasDollar )
-{
+Bool_t TMVA::DataSetFactory::CheckTTreeFormula( TTreeFormula* ttf,
+ const TString& expression,
+ Bool_t& hasDollar )
+{
// checks a TTreeFormula for problems
Bool_t worked = kTRUE;
-
+
if( ttf->GetNdim() <= 0 )
- Log() << kFATAL << "Expression " << expression.Data() << " could not be resolved to a valid formula. " << Endl;
- // if( ttf->GetNcodes() == 0 ){
- // Log() << kWARNING << "Expression: " << expression.Data() << " does not appear to depend on any TTree variable --> please check spelling" << Endl;
- // worked = kFALSE;
- // }
+ Log() << kFATAL << "Expression " << expression.Data()
+ << " could not be resolved to a valid formula. " << Endl;
if( ttf->GetNdata() == 0 ){
- Log() << kWARNING << "Expression: " << expression.Data()
+ Log() << kWARNING << "Expression: " << expression.Data()
<< " does not provide data for this event. "
<< "This event is not taken into account. --> please check if you use as a variable "
<< "an entry of an array which is not filled for some events "
@@ -262,7 +265,7 @@ Bool_t TMVA::DataSetFactory::CheckTTreeFormula( TTreeFormula* ttf, const TString
//_______________________________________________________________________
void TMVA::DataSetFactory::ChangeToNewTree( TreeInfo& tinfo, const DataSetInfo & dsi )
-{
+{
// While the data gets copied into the local training and testing
// trees, the input tree can change (for intance when changing from
// signal to background tree, or using TChains as input) The
@@ -347,7 +350,7 @@ void TMVA::DataSetFactory::ChangeToNewTree( TreeInfo& tinfo, const DataSetInfo &
if (dsi.GetClassInfo(clIdx)->GetName() != tinfo.GetClassName() ) { // if the tree is of another class
fWeightFormula.push_back( 0 );
- continue;
+ continue;
}
ttf = 0;
@@ -482,7 +485,7 @@ void TMVA::DataSetFactory::CalcMinMax( DataSet* ds, TMVA::DataSetInfo& dsi )
TMatrixD* TMVA::DataSetFactory::CalcCorrelationMatrix( DataSet* ds, const UInt_t classNumber )
{
// computes correlation matrix for variables "theVars" in tree;
- // "theType" defines the required event "type"
+ // "theType" defines the required event "type"
// ("type" variable must be present in tree)
// first compute variance-covariance
@@ -498,7 +501,7 @@ TMatrixD* TMVA::DataSetFactory::CalcCorrelationMatrix( DataSet* ds, const UInt_t
if (d > 0) (*mat)(ivar, jvar) /= sqrt(d);
else {
Log() << kWARNING << "<GetCorrelationMatrix> Zero variances for variables "
- << "(" << ivar << ", " << jvar << ") = " << d
+ << "(" << ivar << ", " << jvar << ") = " << d
<< Endl;
(*mat)(ivar, jvar) = 0;
}
@@ -523,7 +526,7 @@ TMatrixD* TMVA::DataSetFactory::CalcCovarianceMatrix( DataSet * ds, const UInt_t
// init matrices
TVectorD vec(nvar);
- TMatrixD mat2(nvar, nvar);
+ TMatrixD mat2(nvar, nvar);
for (ivar=0; ivar<nvar; ivar++) {
vec(ivar) = 0;
for (jvar=0; jvar<nvar; jvar++) mat2(ivar, jvar) = 0;
@@ -538,13 +541,13 @@ TMatrixD* TMVA::DataSetFactory::CalcCovarianceMatrix( DataSet * ds, const UInt_t
Double_t weight = ev->GetWeight();
ic += weight; // count used events
-
+
for (ivar=0; ivar<nvar; ivar++) {
-
+
Double_t xi = ev->GetValue(ivar);
vec(ivar) += xi*weight;
mat2(ivar, ivar) += (xi*xi*weight);
-
+
for (jvar=ivar+1; jvar<nvar; jvar++) {
Double_t xj = ev->GetValue(jvar);
mat2(ivar, jvar) += (xi*xj*weight);
@@ -570,13 +573,16 @@ TMatrixD* TMVA::DataSetFactory::CalcCovarianceMatrix( DataSet * ds, const UInt_t
// --------------------------------------- new versions
//_______________________________________________________________________
-void TMVA::DataSetFactory::InitOptions( TMVA::DataSetInfo& dsi,
- TMVA::NumberPerClassOfTreeType& nTrainTestEvents,
- TString& normMode, UInt_t& splitSeed,
- TString& splitMode,
- TString& mixMode )
+void
+TMVA::DataSetFactory::InitOptions( TMVA::DataSetInfo& dsi,
+ EvtStatsPerClass& nEventRequests,
+ TString& normMode,
+ UInt_t& splitSeed,
+ TString& splitMode,
+ TString& mixMode )
{
// the dataset splitting
+
Configurable splitSpecs( dsi.GetSplitOptions() );
splitSpecs.SetConfigName("DataSetFactory");
splitSpecs.SetConfigDescription( "Configuration options given in the \"PrepareForTrainingAndTesting\" call; these options define the creation of the data sets used for training and expert validation by TMVA" );
@@ -595,10 +601,10 @@ void TMVA::DataSetFactory::InitOptions( TMVA::DataSetInfo& dsi,
splitSpecs.AddPreDefVal(TString("Random"));
splitSpecs.AddPreDefVal(TString("Alternate"));
splitSpecs.AddPreDefVal(TString("Block"));
-
+
splitSeed = 100;
splitSpecs.DeclareOptionRef( splitSeed, "SplitSeed",
- "Seed for random event shuffling" );
+ "Seed for random event shuffling" );
normMode = "NumEvents"; // the weight normalisation modes
splitSpecs.DeclareOptionRef( normMode, "NormMode",
@@ -609,21 +615,14 @@ void TMVA::DataSetFactory::InitOptions( TMVA::DataSetInfo& dsi,
// the number of events
- // initialization
- nTrainTestEvents.insert( TMVA::NumberPerClassOfTreeType::value_type( Types::kTraining, TMVA::NumberPerClass( dsi.GetNClasses() ) ) );
- nTrainTestEvents.insert( TMVA::NumberPerClassOfTreeType::value_type( Types::kTesting, TMVA::NumberPerClass( dsi.GetNClasses() ) ) );
-
// fill in the numbers
for (UInt_t cl = 0; cl < dsi.GetNClasses(); cl++) {
- nTrainTestEvents[Types::kTraining].at(cl) = 0;
- nTrainTestEvents[Types::kTesting].at(cl) = 0;
-
TString clName = dsi.GetClassInfo(cl)->GetName();
TString titleTrain = TString().Format("Number of training events of class %s (default: 0 = all)",clName.Data()).Data();
TString titleTest = TString().Format("Number of test events of class %s (default: 0 = all)",clName.Data()).Data();
- splitSpecs.DeclareOptionRef( nTrainTestEvents[Types::kTraining].at(cl) , TString("nTrain_")+clName, titleTrain );
- splitSpecs.DeclareOptionRef( nTrainTestEvents[Types::kTesting].at(cl) , TString("nTest_")+clName , titleTest );
+ splitSpecs.DeclareOptionRef( nEventRequests.at(cl).nTrainingEventsRequested, TString("nTrain_")+clName, titleTrain );
+ splitSpecs.DeclareOptionRef( nEventRequests.at(cl).nTestingEventsRequested , TString("nTest_")+clName , titleTest );
}
splitSpecs.DeclareOptionRef( fVerbose, "V", "Verbosity (default: true)" );
@@ -637,7 +636,7 @@ void TMVA::DataSetFactory::InitOptions( TMVA::DataSetInfo& dsi,
splitSpecs.CheckForUnusedOptions();
// output logging verbosity
- if (Verbose()) fLogger->SetMinType( kVERBOSE );
+ if (Verbose()) fLogger->SetMinType( kVERBOSE );
if (fVerboseLevel.CompareTo("Debug") ==0) fLogger->SetMinType( kDEBUG );
if (fVerboseLevel.CompareTo("Verbose") ==0) fLogger->SetMinType( kVERBOSE );
if (fVerboseLevel.CompareTo("Info") ==0) fLogger->SetMinType( kINFO );
@@ -647,72 +646,58 @@ void TMVA::DataSetFactory::InitOptions( TMVA::DataSetInfo& dsi,
// adjust mixmode if same as splitmode option has been set
Log() << kINFO << "Splitmode is: \"" << splitMode << "\" the mixmode is: \"" << mixMode << "\"" << Endl;
if (mixMode=="SAMEASSPLITMODE") mixMode = splitMode;
- else if (mixMode!=splitMode)
+ else if (mixMode!=splitMode)
Log() << kINFO << "DataSet splitmode="<<splitMode
<<" differs from mixmode="<<mixMode<<Endl;
}
//_______________________________________________________________________
-void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
- TMVA::DataInputHandler& dataInput,
- TMVA::EventVectorOfClassesOfTreeType& tmpEventVector )
+void
+TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
+ TMVA::DataInputHandler& dataInput,
+ EventVectorOfClassesOfTreeType& eventsmap,
+ EvtStatsPerClass& eventCounts)
{
// build empty event vectors
// distributes events between kTraining/kTesting/kMaxTreeType
-
- tmpEventVector.insert( std::make_pair(Types::kTraining ,TMVA::EventVectorOfClasses(dsi.GetNClasses() ) ) );
- tmpEventVector.insert( std::make_pair(Types::kTesting ,TMVA::EventVectorOfClasses(dsi.GetNClasses() ) ) );
- tmpEventVector.insert( std::make_pair(Types::kMaxTreeType,TMVA::EventVectorOfClasses(dsi.GetNClasses() ) ) );
+ const UInt_t nclasses = dsi.GetNClasses();
+
+ eventsmap[ Types::kTraining ] = EventVectorOfClasses(nclasses);
+ eventsmap[ Types::kTesting ] = EventVectorOfClasses(nclasses);
+ eventsmap[ Types::kMaxTreeType ] = EventVectorOfClasses(nclasses);
// create the type, weight and boostweight branches
- const UInt_t nvars = dsi.GetNVariables();
- const UInt_t ntgts = dsi.GetNTargets();
- const UInt_t nvis = dsi.GetNSpectators();
- // std::vector<Float_t> fmlEval(nvars+ntgts+1+1+nvis); // +1+1 for results of evaluation of cut and weight ttreeformula
-
- // number of signal and background events passing cuts
- std::vector< Int_t > nInitialEvents( dsi.GetNClasses() );
- std::vector< Int_t > nEvBeforeCut( dsi.GetNClasses() );
- std::vector< Int_t > nEvAfterCut( dsi.GetNClasses() );
- std::vector< Float_t > nWeEvBeforeCut( dsi.GetNClasses() );
- std::vector< Float_t > nWeEvAfterCut( dsi.GetNClasses() );
- std::vector< Double_t > nNegWeights( dsi.GetNClasses() );
- std::vector< Float_t* > varAvLength( dsi.GetNClasses() );
+ const UInt_t nvars = dsi.GetNVariables();
+ const UInt_t ntgts = dsi.GetNTargets();
+ const UInt_t nvis = dsi.GetNSpectators();
+
+ for (size_t i=0; i<nclasses; i++) {
+ eventCounts[i].varAvLength = new Float_t[nvars];
+ for (UInt_t ivar=0; ivar<nvars; ivar++)
+ eventCounts[i].varAvLength[ivar] = 0;
+ }
Bool_t haveArrayVariable = kFALSE;
Bool_t *varIsArray = new Bool_t[nvars];
- for (size_t i=0; i<varAvLength.size(); i++) {
- varAvLength[i] = new Float_t[nvars];
- for (UInt_t ivar=0; ivar<nvars; ivar++) {
- //varIsArray[ivar] = kFALSE;
- varAvLength[i][ivar] = 0;
- }
- }
+ // if we work with chains we need to remember the current tree if
+ // the chain jumps to a new tree we have to reset the formulas
+ for (UInt_t cl=0; cl<nclasses; cl++) {
- // if we work with chains we need to remember the current tree
- // if the chain jumps to a new tree we have to reset the formulas
- for (UInt_t cl=0; cl<dsi.GetNClasses(); cl++) {
-
- Log() << kINFO << "Create training and testing trees -- looping over class \""
+ Log() << kINFO << "Create training and testing trees -- looping over class \""
<< dsi.GetClassInfo(cl)->GetName() << "\" ..." << Endl;
+ EventStats& classEventCounts = eventCounts[cl];
+
// info output for weights
- const TString tmpWeight = dsi.GetClassInfo(cl)->GetWeight();
- if (tmpWeight!="") {
- Log() << kINFO << "Weight expression for class \"" << dsi.GetClassInfo(cl)->GetName() << "\": \""
- << tmpWeight << "\"" << Endl;
- }
- else {
- Log() << kINFO << "No weight expression defined for class \"" << dsi.GetClassInfo(cl)->GetName()
- << "\"" << Endl;
- }
-
+ Log() << kINFO << "Weight expression for class \'" << dsi.GetClassInfo(cl)->GetName() << "\': \""
+ << dsi.GetClassInfo(cl)->GetWeight() << "\"" << Endl;
+
// used for chains only
TString currentFileName("");
-
+
std::vector<TreeInfo>::const_iterator treeIt(dataInput.begin(dsi.GetClassInfo(cl)->GetName()));
for (;treeIt!=dataInput.end(dsi.GetClassInfo(cl)->GetName()); treeIt++) {
@@ -721,18 +706,23 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
std::vector<Float_t> tgts(ntgts);
std::vector<Float_t> vis(nvis);
TreeInfo currentInfo = *treeIt;
-
+
+ Log() << kDEBUG << "Building event vectors " << currentInfo.GetTreeType() << Endl;
+
+ EventVector& event_v = eventsmap[currentInfo.GetTreeType()].at(cl);
+
Bool_t isChain = (TString("TChain") == currentInfo.GetTree()->ClassName());
currentInfo.GetTree()->LoadTree(0);
ChangeToNewTree( currentInfo, dsi );
// count number of events in tree before cut
- nInitialEvents.at(cl) += currentInfo.GetTree()->GetEntries();
-
+ classEventCounts.nInitialEvents += currentInfo.GetTree()->GetEntries();
+
// loop over events in ntuple
- for (Long64_t evtIdx = 0; evtIdx < currentInfo.GetTree()->GetEntries(); evtIdx++) {
+ const UInt_t nEvts = currentInfo.GetTree()->GetEntries();
+ for (Long64_t evtIdx = 0; evtIdx < nEvts; evtIdx++) {
currentInfo.GetTree()->LoadTree(evtIdx);
-
+
// may need to reload tree in case of chains
if (isChain) {
if (currentInfo.GetTree()->GetTree()->GetDirectory()->GetFile()->GetName() != currentFileName) {
@@ -743,27 +733,27 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
currentInfo.GetTree()->GetEntry(evtIdx);
Int_t sizeOfArrays = 1;
Int_t prevArrExpr = 0;
-
+
// ======= evaluate all formulas =================
// first we check if some of the formulas are arrays
for (UInt_t ivar=0; ivar<nvars; ivar++) {
Int_t ndata = fInputFormulas[ivar]->GetNdata();
- varAvLength[cl][ivar] += ndata;
+ classEventCounts.varAvLength[ivar] += ndata;
if (ndata == 1) continue;
haveArrayVariable = kTRUE;
varIsArray[ivar] = kTRUE;
if (sizeOfArrays == 1) {
sizeOfArrays = ndata;
prevArrExpr = ivar;
- }
+ }
else if (sizeOfArrays!=ndata) {
Log() << kERROR << "ERROR while preparing training and testing trees:" << Endl;
Log() << " multiple array-type expressions of different length were encountered" << Endl;
- Log() << " location of error: event " << evtIdx
+ Log() << " location of error: event " << evtIdx
<< " in tree " << currentInfo.GetTree()->GetName()
<< " of file " << currentInfo.GetTree()->GetCurrentFile()->GetName() << Endl;
- Log() << " expression " << fInputFormulas[ivar]->GetTitle() << " has "
+ Log() << " expression " << fInputFormulas[ivar]->GetTitle() << " has "
<< ndata << " entries, while" << Endl;
Log() << " expression " << fInputFormulas[prevArrExpr]->GetTitle() << " has "
<< fInputFormulas[prevArrExpr]->GetNdata() << " entries" << Endl;
@@ -782,26 +772,26 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
formula = fCutFormulas[cl];
if (formula) {
Int_t ndata = formula->GetNdata();
- cutVal = (ndata==1 ?
+ cutVal = (ndata==1 ?
formula->EvalInstance(0) :
formula->EvalInstance(idata));
if (TMath::IsNaN(cutVal)) {
containsNaN = kTRUE;
- Log() << kWARNING << "Cut expression resolves to infinite value (NaN): "
+ Log() << kWARNING << "Cut expression resolves to infinite value (NaN): "
<< formula->GetTitle() << Endl;
}
}
-
+
// the input variable
for (UInt_t ivar=0; ivar<nvars; ivar++) {
formula = fInputFormulas[ivar];
- Int_t ndata = formula->GetNdata();
- vars[ivar] = (ndata == 1 ?
- formula->EvalInstance(0) :
+ Int_t ndata = formula->GetNdata();
+ vars[ivar] = (ndata == 1 ?
+ formula->EvalInstance(0) :
formula->EvalInstance(idata));
if (TMath::IsNaN(vars[ivar])) {
containsNaN = kTRUE;
- Log() << kWARNING << "Input expression resolves to infinite value (NaN): "
+ Log() << kWARNING << "Input expression resolves to infinite value (NaN): "
<< formula->GetTitle() << Endl;
}
}
@@ -809,13 +799,13 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
// the targets
for (UInt_t itrgt=0; itrgt<ntgts; itrgt++) {
formula = fTargetFormulas[itrgt];
- Int_t ndata = formula->GetNdata();
- tgts[itrgt] = (ndata == 1 ?
- formula->EvalInstance(0) :
+ Int_t ndata = formula->GetNdata();
+ tgts[itrgt] = (ndata == 1 ?
+ formula->EvalInstance(0) :
formula->EvalInstance(idata));
if (TMath::IsNaN(tgts[itrgt])) {
containsNaN = kTRUE;
- Log() << kWARNING << "Target expression resolves to infinite value (NaN): "
+ Log() << kWARNING << "Target expression resolves to infinite value (NaN): "
<< formula->GetTitle() << Endl;
}
}
@@ -823,13 +813,13 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
// the spectators
for (UInt_t itVis=0; itVis<nvis; itVis++) {
formula = fSpectatorFormulas[itVis];
- Int_t ndata = formula->GetNdata();
- vis[itVis] = (ndata == 1 ?
- formula->EvalInstance(0) :
+ Int_t ndata = formula->GetNdata();
+ vis[itVis] = (ndata == 1 ?
+ formula->EvalInstance(0) :
formula->EvalInstance(idata));
if (TMath::IsNaN(vis[itVis])) {
containsNaN = kTRUE;
- Log() << kWARNING << "Spectator expression resolves to infinite value (NaN): "
+ Log() << kWARNING << "Spectator expression resolves to infinite value (NaN): "
<< formula->GetTitle() << Endl;
}
}
@@ -845,24 +835,24 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
formula->EvalInstance(idata));
if (TMath::IsNaN(weight)) {
containsNaN = kTRUE;
- Log() << kWARNING << "Weight expression resolves to infinite value (NaN): "
+ Log() << kWARNING << "Weight expression resolves to infinite value (NaN): "
<< formula->GetTitle() << Endl;
}
}
-
- // Count the events before rejection due to cut or NaN value
- // (weighted and unweighted)
- nEvBeforeCut.at(cl) ++;
+
+ // Count the events before rejection due to cut or NaN
+ // value (weighted and unweighted)
+ classEventCounts.nEvBeforeCut++;
if (!TMath::IsNaN(weight))
- nWeEvBeforeCut.at(cl) += weight;
+ classEventCounts.nWeEvBeforeCut += weight;
- // apply the cut
- // skip rest if cut is not fulfilled
+ // apply the cut, skip rest if cut is not fulfilled
if (cutVal<0.5) continue;
- // global flag if negative weights exist -> can be used by classifiers who may
- // require special data treatment (also print warning)
- if (weight < 0) nNegWeights.at(cl)++;
+ // global flag if negative weights exist -> can be used
+ // by classifiers who may require special data
+ // treatment (also print warning)
+ if (weight < 0) classEventCounts.nNegWeights++;
// now read the event-values (variables and regression targets)
@@ -874,90 +864,93 @@ void TMVA::DataSetFactory::BuildEventVector( TMVA::DataSetInfo& dsi,
// Count the events after rejection due to cut or NaN value
// (weighted and unweighted)
- nEvAfterCut.at(cl) ++;
- nWeEvAfterCut.at(cl) += weight;
+ classEventCounts.nEvAfterCut++;
+ classEventCounts.nWeEvAfterCut += weight;
// event accepted, fill temporary ntuple
- tmpEventVector.find(currentInfo.GetTreeType())->second.at(cl).push_back(new Event(vars, tgts , vis, cl , weight));
-
+ event_v.push_back(new Event(vars, tgts , vis, cl , weight));
}
}
-
currentInfo.GetTree()->ResetBranchAddresses();
}
}
- // for output, check maximum class name length
+ // for output format, get the maximum class name length
Int_t maxL = dsi.GetClassNameMaxLength();
-
+
Log() << kINFO << "Number of events in input trees (after possible flattening of arrays):" << Endl;
for (UInt_t cl = 0; cl < dsi.GetNClasses(); cl++) {
- Log() << kINFO << " "
- << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ Log() << kINFO << " "
+ << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
<< " -- number of events : "
- << std::setw(5) << nEvBeforeCut.at(cl)
- << " / sum of weights: " << std::setw(5) << nWeEvBeforeCut.at(cl) << Endl;
+ << std::setw(5) << eventCounts[cl].nEvBeforeCut
+ << " / sum of weights: " << std::setw(5) << eventCounts[cl].nWeEvBeforeCut << Endl;
}
for (UInt_t cl = 0; cl < dsi.GetNClasses(); cl++) {
- Log() << kINFO << " " << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
- <<" tree -- total number of entries: "
+ Log() << kINFO << " " << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ <<" tree -- total number of entries: "
<< std::setw(5) << dataInput.GetEntries(dsi.GetClassInfo(cl)->GetName()) << Endl;
}
- Log() << kINFO << "Preselection:" << Endl;
+ Log() << kINFO << "Preselection: (will effect number of requested training and testing events)" << Endl;
if (dsi.HasCuts()) {
for (UInt_t cl = 0; cl< dsi.GetNClasses(); cl++) {
- Log() << kINFO << " " << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ Log() << kINFO << " " << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
<< " requirement: \"" << dsi.GetClassInfo(cl)->GetCut() << "\"" << Endl;
- Log() << kINFO << " "
- << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ Log() << kINFO << " "
+ << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
<< " -- number of events passed: "
- << std::setw(5) << nEvAfterCut.at(cl)
- << " / sum of weights: " << std::setw(5) << nWeEvAfterCut.at(cl) << Endl;
- Log() << kINFO << " "
- << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
+ << std::setw(5) << eventCounts[cl].nEvAfterCut
+ << " / sum of weights: " << std::setw(5) << eventCounts[cl].nWeEvAfterCut << Endl;
+ Log() << kINFO << " "
+ << setiosflags(ios::left) << std::setw(maxL) << dsi.GetClassInfo(cl)->GetName()
<< " -- efficiency : "
- << std::setw(6) << nWeEvAfterCut.at(cl)/nWeEvBeforeCut.at(cl) << Endl;
+ << std::setw(6) << eventCounts[cl].nWeEvAfterCut/eventCounts[cl].nWeEvBeforeCut << Endl;
}
}
else Log() << kINFO << " No preselection cuts applied on event classes" << Endl;
delete[] varIsArray;
- for (size_t i=0; i<varAvLength.size(); i++)
- delete[] varAvLength[i];
}
//_______________________________________________________________________
-TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
- TMVA::EventVectorOfClassesOfTreeType& tmpEventVector,
- TMVA::NumberPerClassOfTreeType& nTrainTestEvents,
- const TString& splitMode,
- const TString& mixMode,
- const TString& normMode,
- UInt_t splitSeed)
+TMVA::DataSet*
+TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
+ EventVectorOfClassesOfTreeType& tmpEventVector,
+ EvtStatsPerClass& eventCounts,
+ const TString& splitMode,
+ const TString& mixMode,
+ const TString& normMode,
+ UInt_t splitSeed)
{
// Select and distribute unassigned events to kTraining and kTesting
- Bool_t emptyUndefined = kTRUE;
+ //Bool_t emptyUndefined = kTRUE;
-// // check if the vectors of all classes are empty
- for( Int_t cls = 0, clsEnd = dsi.GetNClasses(); cls < clsEnd; ++cls ){
- emptyUndefined &= tmpEventVector[Types::kMaxTreeType].at(cls).empty();
- }
+ // check if the vectors of all classes are empty
+ //for( Int_t cls = 0, clsEnd = dsi.GetNClasses(); cls < clsEnd; ++cls ){
+ // emptyUndefined &= tmpEventVector[Types::kMaxTreeType].at(cls).empty();
+ //}
TMVA::RandomGenerator rndm( splitSeed );
-
+
// ==== splitting of undefined events to kTraining and kTesting
// if splitMode contains "RANDOM", then shuffle the undefined events
- if (splitMode.Contains( "RANDOM" ) && !emptyUndefined ) {
- Log() << kDEBUG << "randomly shuffling events which are not yet associated to testing or training"<<Endl;
+ if (splitMode.Contains( "RANDOM" ) /*&& !emptyUndefined*/ ) {
// random shuffle the undefined events of each class
for( UInt_t cls = 0; cls < dsi.GetNClasses(); ++cls ){
- std::random_shuffle(tmpEventVector[Types::kMaxTreeType].at(cls).begin(),
- tmpEventVector[Types::kMaxTreeType].at(cls).end(),
- rndm );
+ EventVector& unspecifiedEvents = tmpEventVector[Types::kMaxTreeType].at(cls);
+ if( ! unspecifiedEvents.empty() ) {
+ Log() << kDEBUG << "randomly shuffling "
+ << unspecifiedEvents.size()
+ << " events of class " << cls
+ << " which are not yet associated to testing or training" << Endl;
+ std::random_shuffle( unspecifiedEvents.begin(),
+ unspecifiedEvents.end(),
+ rndm );
+ }
}
}
@@ -968,59 +961,72 @@ TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
Log() << kDEBUG << "check number of training/testing events, requested and available number of events and for class " << cls << Endl;
// check if enough or too many events are already in the training/testing eventvectors of the class cls
- EventVector& eventVectorTraining = tmpEventVector.find( Types::kTraining )->second.at(cls);
- EventVector& eventVectorTesting = tmpEventVector.find( Types::kTesting )->second.at(cls);
- EventVector& eventVectorUndefined= tmpEventVector.find( Types::kMaxTreeType )->second.at(cls);
-
- Int_t alreadyAvailableTraining = eventVectorTraining.size();
- Int_t alreadyAvailableTesting = eventVectorTesting.size();
- Int_t availableUndefined = eventVectorUndefined.size();
-
- Int_t requestedTraining = nTrainTestEvents.find( Types::kTraining )->second.at(cls);
- Int_t requestedTesting = nTrainTestEvents.find( Types::kTesting )->second.at(cls);
-
- Log() << kDEBUG << "availableTraining " << alreadyAvailableTraining << Endl;
- Log() << kDEBUG << "availableTesting " << alreadyAvailableTesting << Endl;
- Log() << kDEBUG << "availableUndefined " << availableUndefined << Endl;
- Log() << kDEBUG << "requestedTraining " << requestedTraining << Endl;
- Log() << kDEBUG << "requestedTesting " << requestedTesting << Endl;
- //
- // nomenclature r=available training
- // s=available testing
- // u=available undefined
- // R= requested training
- // S= requested testing
- // nR = used for selection of training events
- // nS = used for selection of test events
- // we have: nR + nS = r+s+u
+ EventVector& eventVectorTraining = tmpEventVector[ Types::kTraining ].at(cls);
+ EventVector& eventVectorTesting = tmpEventVector[ Types::kTesting ].at(cls);
+ EventVector& eventVectorUndefined = tmpEventVector[ Types::kMaxTreeType ].at(cls);
+
+ Int_t availableTraining = eventVectorTraining.size();
+ Int_t availableTesting = eventVectorTesting.size();
+ Int_t availableUndefined = eventVectorUndefined.size();
+
+ Float_t presel_scale = eventCounts[cls].cutScaling();
+
+ Int_t requestedTraining = Int_t(eventCounts[cls].nTrainingEventsRequested * presel_scale);
+ Int_t requestedTesting = Int_t(eventCounts[cls].nTestingEventsRequested * presel_scale);
+
+ Log() << kDEBUG << "events in training trees : " << availableTraining << Endl;
+ Log() << kDEBUG << "events in testing trees : " << availableTesting << Endl;
+ Log() << kDEBUG << "events in unspecified trees : " << availableUndefined << Endl;
+ Log() << kDEBUG << "requested for training : " << requestedTraining;
+ if(presel_scale<1)
+ Log() << " ( " << eventCounts[cls].nTrainingEventsRequested
+ << " * " << presel_scale << " preselection efficiency)" << Endl;
+ else
+ Log() << Endl;
+ Log() << kDEBUG << "requested for testing : " << requestedTesting;
+ if(presel_scale<1)
+ Log() << " ( " << eventCounts[cls].nTestingEventsRequested
+ << " * " << presel_scale << " preselection efficiency)" << Endl;
+ else
+ Log() << Endl;
+
+ // nomenclature r = available training
+ // s = available testing
+ // u = available undefined
+ // R = requested training
+ // S = requested testing
+ // nR = to be used to select training events
+ // nS = to be used to select test events
+ // we have the constraint: nR + nS < r+s+u,
+ // since we can not use more events than we have
// free events: Nfree = u-Thet(R-r)-Thet(S-s)
- // nomenclature: Thet(x) = x, if x>0 else 0;
+ // nomenclature: Thet(x) = x, if x>0 else 0
// nR = max(R,r) + 0.5 * Nfree
// nS = max(S,s) + 0.5 * Nfree
// nR +nS = R+S + u-R+r-S+s = u+r+s= ok! for R>r
// nR +nS = r+S + u-S+s = u+r+s= ok! for r>R
- //EVT three different cases might occur here
+ // three different cases might occur here
//
// Case a
- // requestedTraining and requestedTesting >0
+ // requestedTraining and requestedTesting >0
// free events: Nfree = u-Thet(R-r)-Thet(S-s)
// nR = Max(R,r) + 0.5 * Nfree
// nS = Max(S,s) + 0.5 * Nfree
- //
+ //
// Case b
// exactly one of requestedTraining or requestedTesting >0
// assume training R >0
- // nR = max(R,r)
+ // nR = max(R,r)
// nS = s+u+r-nR
// and s=nS
//
- //Case c:
- // requestedTraining=0, requestedTesting=0
+ // Case c
+ // requestedTraining=0, requestedTesting=0
// Nfree = u-|r-s|
// if NFree >=0
// R = Max(r,s) + 0.5 * Nfree = S
- // else if r>s
+ // else if r>s
// R = r; S=s+u
// else
// R = r+u; S=s
@@ -1030,90 +1036,119 @@ TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
// distribute undefined events according to nR, nS
// finally determine actual sub samples from nR and nS to be used in training / testing
//
- // implementation of case C)
- int useForTesting,useForTraining;
- if( (requestedTraining == 0) && (requestedTesting == 0)){
- // 0 means automatic distribution of events
- Log() << kDEBUG << "requested 0" << Endl;
- // try to get the same number of events in training and testing for this class (balance)
- Int_t NFree = availableUndefined - TMath::Abs(alreadyAvailableTraining - alreadyAvailableTesting);
- if (NFree >=0){
- requestedTraining = TMath::Max(alreadyAvailableTraining,alreadyAvailableTesting) + NFree/2;
- requestedTesting = availableUndefined+alreadyAvailableTraining+alreadyAvailableTesting - requestedTraining; // the rest
- } else if (alreadyAvailableTraining > alreadyAvailableTesting){ //r>s
- requestedTraining = alreadyAvailableTraining;
- requestedTesting = alreadyAvailableTesting +availableUndefined;
- }
- else {
- requestedTraining = alreadyAvailableTraining+availableUndefined;
- requestedTesting = alreadyAvailableTesting;
+
+ Int_t useForTesting(0),useForTraining(0);
+ Int_t allAvailable(availableUndefined + availableTraining + availableTesting);
+
+ if( (requestedTraining == 0) && (requestedTesting == 0)){
+
+ // Case C: balance the number of training and testing events
+
+ if ( availableUndefined >= TMath::Abs(availableTraining - availableTesting) ) {
+ // enough unspecified are available to equal training and testing
+ useForTraining = useForTesting = allAvailable/2;
+ } else {
+ // all unspecified are assigned to the smaller of training / testing
+ useForTraining = availableTraining;
+ useForTesting = availableTesting;
+ if (availableTraining < availableTesting)
+ useForTraining += availableUndefined;
+ else
+ useForTesting += availableUndefined;
}
- useForTraining = requestedTraining;
- useForTesting = requestedTesting;
+ requestedTraining = useForTraining;
+ requestedTesting = useForTesting;
}
- else if (requestedTesting == 0){ // case B)
- useForTraining = TMath::Max(requestedTraining,alreadyAvailableTraining);
- useForTesting= availableUndefined+alreadyAvailableTraining+alreadyAvailableTesting - useForTraining; // the rest
+
+ else if (requestedTesting == 0){
+ // case B
+ useForTraining = TMath::Max(requestedTraining,availableTraining);
+ if (allAvailable < useForTraining) {
+ Log() << kFATAL << "More events requested for training ("
+ << requestedTraining << ") than available ("
+ << allAvailable << ")!" << Endl;
+ }
+ useForTesting = allAvailable - useForTraining; // the rest
requestedTesting = useForTesting;
}
+
else if (requestedTraining == 0){ // case B)
- useForTesting = TMath::Max(requestedTesting,alreadyAvailableTesting);
- useForTraining= availableUndefined+alreadyAvailableTraining+alreadyAvailableTesting - useForTesting; // the rest
+ useForTesting = TMath::Max(requestedTesting,availableTesting);
+ if (allAvailable < useForTesting) {
+ Log() << kFATAL << "More events requested for testing ("
+ << requestedTesting << ") than available ("
+ << allAvailable << ")!" << Endl;
+ }
+ useForTraining= allAvailable - useForTesting; // the rest
requestedTraining = useForTraining;
}
- else{ // case A
- int NFree = availableUndefined-TMath::Max(requestedTraining-alreadyAvailableTraining,0)-TMath::Max(requestedTesting-alreadyAvailableTesting,0);
+
+ else {
+ // Case A
+ // requestedTraining R and requestedTesting S >0
+ // free events: Nfree = u-Thet(R-r)-Thet(S-s)
+ // nR = Max(R,r) + 0.5 * Nfree
+ // nS = Max(S,s) + 0.5 * Nfree
+ Int_t stillNeedForTraining = TMath::Max(requestedTraining-availableTraining,0);
+ Int_t stillNeedForTesting = TMath::Max(requestedTesting-availableTesting,0);
+
+ int NFree = availableUndefined - stillNeedForTraining - stillNeedForTesting;
if (NFree <0) NFree = 0;
- useForTraining = TMath::Max(requestedTraining,alreadyAvailableTraining) + NFree/2;
- useForTesting= availableUndefined+alreadyAvailableTraining+alreadyAvailableTesting - useForTraining; // the rest
+ useForTraining = TMath::Max(requestedTraining,availableTraining) + NFree/2;
+ useForTesting= allAvailable - useForTraining; // the rest
}
+
Log() << kDEBUG << "determined event sample size to select training sample from="<<useForTraining<<Endl;
Log() << kDEBUG << "determined event sample size to select test sample from="<<useForTesting<<Endl;
-
- // associate undefined events
+
+
+
+
+
+ // associate undefined events
if( splitMode == "ALTERNATE" ){
Log() << kDEBUG << "split 'ALTERNATE'" << Endl;
- Int_t nTraining = alreadyAvailableTraining;
- Int_t nTesting = alreadyAvailableTesting;
+ Int_t nTraining = availableTraining;
+ Int_t nTesting = availableTesting;
for( EventVector::iterator it = eventVectorUndefined.begin(), itEnd = eventVectorUndefined.end(); it != itEnd; ){
- ++nTraining;
- if( nTraining <= requestedTraining ){
- eventVectorTraining.insert( eventVectorTraining.end(), (*it) );
- ++it;
- }
+ ++nTraining;
+ if( nTraining <= requestedTraining ){
+ eventVectorTraining.insert( eventVectorTraining.end(), (*it) );
+ ++it;
+ }
if( it != itEnd ){
- ++nTesting;
+ ++nTesting;
eventVectorTesting.insert( eventVectorTesting.end(), (*it) );
++it;
}
}
- }else{
+ } else {
Log() << kDEBUG << "split '" << splitMode << "'" << Endl;
- // test if enough events are available
- Log() << kDEBUG << "availableundefined : " << availableUndefined << Endl;
- Log() << kDEBUG << "useForTraining : " << useForTraining << Endl;
- Log() << kDEBUG << "useForTesting : " << useForTesting << Endl;
- Log() << kDEBUG << "alreadyAvailableTraining : " << alreadyAvailableTraining << Endl;
- Log() << kDEBUG << "alreadyAvailableTesting : " << alreadyAvailableTesting << Endl;
-
- if( availableUndefined<(useForTraining-alreadyAvailableTraining) ||
- availableUndefined<(useForTesting -alreadyAvailableTesting ) ||
- availableUndefined<(useForTraining+useForTesting-alreadyAvailableTraining-alreadyAvailableTesting ) ){
- Log() << kFATAL << "More events requested than available!" << Endl;
- }
-
- // select the events
- if (useForTraining>alreadyAvailableTraining){
- eventVectorTraining.insert( eventVectorTraining.end() , eventVectorUndefined.begin(), eventVectorUndefined.begin()+ useForTraining- alreadyAvailableTraining );
- eventVectorUndefined.erase( eventVectorUndefined.begin(), eventVectorUndefined.begin() + useForTraining- alreadyAvailableTraining);
+ // test if enough events are available
+ Log() << kDEBUG << "availableundefined : " << availableUndefined << Endl;
+ Log() << kDEBUG << "useForTraining : " << useForTraining << Endl;
+ Log() << kDEBUG << "useForTesting : " << useForTesting << Endl;
+ Log() << kDEBUG << "availableTraining : " << availableTraining << Endl;
+ Log() << kDEBUG << "availableTesting : " << availableTesting << Endl;
+
+ if( availableUndefined<(useForTraining-availableTraining) ||
+ availableUndefined<(useForTesting -availableTesting ) ||
+ availableUndefined<(useForTraining+useForTesting-availableTraining-availableTesting ) ){
+ Log() << kFATAL << "More events requested than available!" << Endl;
+ }
+
+ // select the events
+ if (useForTraining>availableTraining){
+ eventVectorTraining.insert( eventVectorTraining.end() , eventVectorUndefined.begin(), eventVectorUndefined.begin()+ useForTraining- availableTraining );
+ eventVectorUndefined.erase( eventVectorUndefined.begin(), eventVectorUndefined.begin() + useForTraining- availableTraining);
}
- if (useForTesting>alreadyAvailableTesting){
- eventVectorTesting.insert( eventVectorTesting.end() , eventVectorUndefined.begin(), eventVectorUndefined.begin()+ useForTesting- alreadyAvailableTesting );
+ if (useForTesting>availableTesting){
+ eventVectorTesting.insert( eventVectorTesting.end() , eventVectorUndefined.begin(), eventVectorUndefined.begin()+ useForTesting- availableTesting );
}
}
- eventVectorUndefined.clear();
+ eventVectorUndefined.clear();
// finally shorten the event vectors to the requested size by removing random events
if (splitMode.Contains( "RANDOM" )){
UInt_t sizeTraining = eventVectorTraining.size();
@@ -1153,21 +1188,21 @@ TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
}
}
else { // erase at end
- if( eventVectorTraining.size() < UInt_t(requestedTraining) )
- Log() << kWARNING << "DataSetFactory/requested number of training samples larger than size of eventVectorTraining.\n"
- << "There is probably an issue. Please contact the TMVA developers." << Endl;
+ if( eventVectorTraining.size() < UInt_t(requestedTraining) )
+ Log() << kWARNING << "DataSetFactory/requested number of training samples larger than size of eventVectorTraining.\n"
+ << "There is probably an issue. Please contact the TMVA developers." << Endl;
std::for_each( eventVectorTraining.begin()+requestedTraining, eventVectorTraining.end(), DeleteFunctor<Event>() );
eventVectorTraining.erase(eventVectorTraining.begin()+requestedTraining,eventVectorTraining.end());
- if( eventVectorTesting.size() < UInt_t(requestedTesting) )
- Log() << kWARNING << "DataSetFactory/requested number of testing samples larger than size of eventVectorTesting.\n"
- << "There is probably an issue. Please contact the TMVA developers." << Endl;
+ if( eventVectorTesting.size() < UInt_t(requestedTesting) )
+ Log() << kWARNING << "DataSetFactory/requested number of testing samples larger than size of eventVectorTesting.\n"
+ << "There is probably an issue. Please contact the TMVA developers." << Endl;
std::for_each( eventVectorTesting.begin()+requestedTesting, eventVectorTesting.end(), DeleteFunctor<Event>() );
eventVectorTesting.erase(eventVectorTesting.begin()+requestedTesting,eventVectorTesting.end());
}
}
- TMVA::DataSetFactory::RenormEvents( dsi, tmpEventVector, normMode );
+ TMVA::DataSetFactory::RenormEvents( dsi, tmpEventVector, eventCounts, normMode );
Int_t trainingSize = 0;
Int_t testingSize = 0;
@@ -1211,35 +1246,35 @@ TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
Log() << kDEBUG << "insert class 0 into training and test vector" << Endl;
trainingEventVector->insert( trainingEventVector->end(), tmpEventVector[Types::kTraining].at(0).begin(), tmpEventVector[Types::kTraining].at(0).end() );
testingEventVector->insert( testingEventVector->end(), tmpEventVector[Types::kTesting].at(0).begin(), tmpEventVector[Types::kTesting].at(0).end() );
-
+
// insert other classes
EvtVecIt itTarget;
for( UInt_t cls = 1; cls < dsi.GetNClasses(); ++cls ){
Log() << kDEBUG << "insert class " << cls << Endl;
// training vector
itTarget = trainingEventVector->begin() - 1; // start one before begin
- // loop over source
+ // loop over source
for( itEvent = tmpEventVector[Types::kTraining].at(cls).begin(), itEventEnd = tmpEventVector[Types::kTraining].at(cls).end(); itEvent != itEventEnd; ++itEvent ){
// if( std::distance( itTarget, trainingEventVector->end()) < Int_t(cls+1) ) {
if( (trainingEventVector->end() - itTarget) < Int_t(cls+1) ) {
itTarget = trainingEventVector->end();
trainingEventVector->insert( itTarget, itEvent, itEventEnd ); // fill in the rest without mixing
break;
- }else{
+ }else{
itTarget += cls+1;
trainingEventVector->insert( itTarget, (*itEvent) ); // fill event
}
}
// testing vector
itTarget = testingEventVector->begin() - 1;
- // loop over source
+ // loop over source
for( itEvent = tmpEventVector[Types::kTesting].at(cls).begin(), itEventEnd = tmpEventVector[Types::kTesting].at(cls).end(); itEvent != itEventEnd; ++itEvent ){
// if( std::distance( itTarget, testingEventVector->end()) < Int_t(cls+1) ) {
if( ( testingEventVector->end() - itTarget ) < Int_t(cls+1) ) {
itTarget = testingEventVector->end();
testingEventVector->insert( itTarget, itEvent, itEventEnd ); // fill in the rest without mixing
break;
- }else{
+ }else{
itTarget += cls+1;
testingEventVector->insert( itTarget, (*itEvent) ); // fill event
}
@@ -1250,13 +1285,13 @@ TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
// std::cout << std::endl;
// std::cout << "TRAINING VECTOR" << std::endl;
// std::transform( trainingEventVector->begin(), trainingEventVector->end(), ostream_iterator<Int_t>(std::cout, "|"), std::mem_fun(&TMVA::Event::GetClass) );
-
+
// std::cout << std::endl;
// std::cout << "TESTING VECTOR" << std::endl;
// std::transform( testingEventVector->begin(), testingEventVector->end(), ostream_iterator<Int_t>(std::cout, "|"), std::mem_fun(&TMVA::Event::GetClass) );
// std::cout << std::endl;
- }else{
+ }else{
for( UInt_t cls = 0; cls < dsi.GetNClasses(); ++cls ){
trainingEventVector->insert( trainingEventVector->end(), tmpEventVector[Types::kTraining].at(cls).begin(), tmpEventVector[Types::kTraining].at(cls).end() );
testingEventVector->insert ( testingEventVector->end(), tmpEventVector[Types::kTesting].at(cls).begin(), tmpEventVector[Types::kTesting].at(cls).end() );
@@ -1288,20 +1323,22 @@ TMVA::DataSet* TMVA::DataSetFactory::MixEvents( DataSetInfo& dsi,
// create dataset
DataSet* ds = new DataSet(dsi);
- Log() << kINFO << "Create internal training tree" << Endl;
- ds->SetEventCollection(trainingEventVector, Types::kTraining );
- Log() << kINFO << "Create internal testing tree" << Endl;
- ds->SetEventCollection(testingEventVector, Types::kTesting );
+ Log() << kINFO << "Create internal training tree" << Endl;
+ ds->SetEventCollection(trainingEventVector, Types::kTraining );
+ Log() << kINFO << "Create internal testing tree" << Endl;
+ ds->SetEventCollection(testingEventVector, Types::kTesting );
return ds;
-
+
}
//_______________________________________________________________________
-void TMVA::DataSetFactory::RenormEvents( TMVA::DataSetInfo& dsi,
- TMVA::EventVectorOfClassesOfTreeType& tmpEventVector,
- const TString& normMode )
+void
+TMVA::DataSetFactory::RenormEvents( TMVA::DataSetInfo& dsi,
+ EventVectorOfClassesOfTreeType& tmpEventVector,
+ const EvtStatsPerClass& eventCounts,
+ const TString& normMode )
{
// ============================================================
// renormalisation
@@ -1406,20 +1443,18 @@ void TMVA::DataSetFactory::RenormEvents( TMVA::DataSetInfo& dsi,
// ---------------------------------
// now apply the normalization factors
Int_t maxL = dsi.GetClassNameMaxLength();
- for (UInt_t cls = 0, clsEnd = dsi.GetNClasses(); cls<clsEnd; ++cls) {
- Log() << kINFO << "--> Rescale " << setiosflags(ios::left) << std::setw(maxL)
+ for (UInt_t cls = 0, clsEnd = dsi.GetNClasses(); cls<clsEnd; ++cls) {
+ Log() << kINFO << "--> Rescale " << setiosflags(ios::left) << std::setw(maxL)
<< dsi.GetClassInfo(cls)->GetName() << " event weights by factor: " << renormFactor.at(cls) << Endl;
- std::for_each( tmpEventVector[Types::kTraining].at(cls).begin(),
+ std::for_each( tmpEventVector[Types::kTraining].at(cls).begin(),
tmpEventVector[Types::kTraining].at(cls).end(),
std::bind2nd(std::mem_fun(&TMVA::Event::ScaleWeight),renormFactor.at(cls)) );
- std::for_each( tmpEventVector[Types::kTesting].at(cls).begin(),
+ std::for_each( tmpEventVector[Types::kTesting].at(cls).begin(),
tmpEventVector[Types::kTesting].at(cls).end(),
std::bind2nd(std::mem_fun(&TMVA::Event::ScaleWeight),renormFactor.at(cls)) );
}
-
-
// ---------------------------------
// for information purposes
dsi.SetNormalization( normMode );
@@ -1454,20 +1489,29 @@ void TMVA::DataSetFactory::RenormEvents( TMVA::DataSetInfo& dsi,
testingSumWeights += testingSumWeightsPerClass.at(cls);
// output statistics
- Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
- << dsi.GetClassInfo(cls)->GetName() << " -- "
- << "training entries : " << trainingSizePerClass.at(cls)
- << " (" << "sum of weights: " << trainingSumWeightsPerClass.at(cls) << ")" << Endl;
- Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
- << dsi.GetClassInfo(cls)->GetName() << " -- "
- << "testing entries : " << testingSizePerClass.at(cls)
- << " (" << "sum of weights: " << testingSumWeightsPerClass.at(cls) << ")" << Endl;
- Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
- << dsi.GetClassInfo(cls)->GetName() << " -- "
- << "training and testing entries: "
- << (trainingSizePerClass.at(cls)+testingSizePerClass.at(cls))
- << " (" << "sum of weights: "
+
+ Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
+ << dsi.GetClassInfo(cls)->GetName() << " -- "
+ << "training events : " << trainingSizePerClass.at(cls)
+ << " (sum of weights: " << trainingSumWeightsPerClass.at(cls) << ")"
+ << " - requested were " << eventCounts[cls].nTrainingEventsRequested << " events" << Endl;
+ Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
+ << dsi.GetClassInfo(cls)->GetName() << " -- "
+ << "testing events : " << testingSizePerClass.at(cls)
+ << " (sum of weights: " << testingSumWeightsPerClass.at(cls) << ")"
+ << " - requested were " << eventCounts[cls].nTestingEventsRequested << " events" << Endl;
+ Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
+ << dsi.GetClassInfo(cls)->GetName() << " -- "
+ << "training and testing events: "
+ << (trainingSizePerClass.at(cls)+testingSizePerClass.at(cls))
+ << " (sum of weights: "
<< (trainingSumWeightsPerClass.at(cls)+testingSumWeightsPerClass.at(cls)) << ")" << Endl;
+ if(eventCounts[cls].nEvAfterCut<eventCounts[cls].nEvBeforeCut) {
+ Log() << kINFO << setiosflags(ios::left) << std::setw(maxL)
+ << dsi.GetClassInfo(cls)->GetName() << " -- "
+ << "due to the preselection a scaling factor has been applied to the numbers of requested events: "
+ << eventCounts[cls].cutScaling() << Endl;
+ }
}
}
diff --git a/tmva/src/DataSetInfo.cxx b/tmva/src/DataSetInfo.cxx
index e59f7c47eb0..1d773b0e851 100644
--- a/tmva/src/DataSetInfo.cxx
+++ b/tmva/src/DataSetInfo.cxx
@@ -58,7 +58,7 @@
#endif
//_______________________________________________________________________
-TMVA::DataSetInfo::DataSetInfo(const TString& name)
+TMVA::DataSetInfo::DataSetInfo(const TString& name)
: TObject(),
fDataSetManager(NULL),
fName(name),
@@ -81,7 +81,7 @@ TMVA::DataSetInfo::DataSetInfo(const TString& name)
}
//_______________________________________________________________________
-TMVA::DataSetInfo::~DataSetInfo()
+TMVA::DataSetInfo::~DataSetInfo()
{
// destructor
ClearDataSet();
@@ -101,6 +101,12 @@ void TMVA::DataSetInfo::ClearDataSet() const
if(fDataSet!=0) { delete fDataSet; fDataSet=0; }
}
+void
+TMVA::DataSetInfo::SetMsgType( EMsgType t ) const
+{
+ fLogger->SetMinType(t);
+}
+
//_______________________________________________________________________
TMVA::ClassInfo* TMVA::DataSetInfo::AddClass( const TString& className )
{
@@ -119,12 +125,6 @@ TMVA::ClassInfo* TMVA::DataSetInfo::AddClass( const TString& className )
return fClasses.back();
}
-//_______________________________________________________________________
-void TMVA::DataSetInfo::SetMsgType( EMsgType t ) const
-{
- fLogger->SetMinType(t);
-}
-
//_______________________________________________________________________
TMVA::ClassInfo* TMVA::DataSetInfo::GetClassInfo( const TString& name ) const
{
@@ -166,12 +166,12 @@ std::vector<Float_t>* TMVA::DataSetInfo::GetTargetsForMulticlass( const TMVA::E
// fTargetsForMulticlass->resize( GetNClasses() );
fTargetsForMulticlass->assign( GetNClasses(), 0.0 );
fTargetsForMulticlass->at( ev->GetClass() ) = 1.0;
- return fTargetsForMulticlass;
+ return fTargetsForMulticlass;
}
//_______________________________________________________________________
-Bool_t TMVA::DataSetInfo::HasCuts() const
+Bool_t TMVA::DataSetInfo::HasCuts() const
{
Bool_t hasCuts = kFALSE;
for (std::vector<ClassInfo*>::iterator it = fClasses.begin(); it < fClasses.end(); it++) {
@@ -181,22 +181,26 @@ Bool_t TMVA::DataSetInfo::HasCuts() const
}
//_______________________________________________________________________
-const TMatrixD* TMVA::DataSetInfo::CorrelationMatrix( const TString& className ) const
-{
+const TMatrixD* TMVA::DataSetInfo::CorrelationMatrix( const TString& className ) const
+{
ClassInfo* ptr = GetClassInfo(className);
return ptr?ptr->GetCorrelationMatrix():0;
}
//_______________________________________________________________________
-TMVA::VariableInfo& TMVA::DataSetInfo::AddVariable( const TString& expression, const TString& title, const TString& unit,
- Double_t min, Double_t max, char varType,
- Bool_t normalized, void* external )
+TMVA::VariableInfo& TMVA::DataSetInfo::AddVariable( const TString& expression,
+ const TString& title,
+ const TString& unit,
+ Double_t min, Double_t max,
+ char varType,
+ Bool_t normalized,
+ void* external )
{
- // add a variable (can be a complex expression) to the set of variables used in
- // the MV analysis
+ // add a variable (can be a complex expression) to the set of
+ // variables used in the MV analysis
TString regexpr = expression; // remove possible blanks
regexpr.ReplaceAll(" ", "" );
- fVariables.push_back(VariableInfo( regexpr, title, unit,
+ fVariables.push_back(VariableInfo( regexpr, title, unit,
fVariables.size()+1, varType, external, min, max, normalized ));
fNeedsRebuilding = kTRUE;
return fVariables.back();
@@ -211,17 +215,21 @@ TMVA::VariableInfo& TMVA::DataSetInfo::AddVariable( const VariableInfo& varInfo)
}
//_______________________________________________________________________
-TMVA::VariableInfo& TMVA::DataSetInfo::AddTarget( const TString& expression, const TString& title, const TString& unit,
- Double_t min, Double_t max,
- Bool_t normalized, void* external )
+TMVA::VariableInfo& TMVA::DataSetInfo::AddTarget( const TString& expression,
+ const TString& title,
+ const TString& unit,
+ Double_t min, Double_t max,
+ Bool_t normalized,
+ void* external )
{
- // add a variable (can be a complex expression) to the set of variables used in
- // the MV analysis
+ // add a variable (can be a complex expression) to the set of
+ // variables used in the MV analysis
TString regexpr = expression; // remove possible blanks
regexpr.ReplaceAll(" ", "" );
char type='F';
- fTargets.push_back(VariableInfo( regexpr, title, unit,
- fTargets.size()+1, type, external, min, max, normalized ));
+ fTargets.push_back(VariableInfo( regexpr, title, unit,
+ fTargets.size()+1, type, external, min,
+ max, normalized ));
fNeedsRebuilding = kTRUE;
return fTargets.back();
}
@@ -235,7 +243,9 @@ TMVA::VariableInfo& TMVA::DataSetInfo::AddTarget( const VariableInfo& varInfo){
}
//_______________________________________________________________________
-TMVA::VariableInfo& TMVA::DataSetInfo::AddSpectator( const TString& expression, const TString& title, const TString& unit,
+TMVA::VariableInfo& TMVA::DataSetInfo::AddSpectator( const TString& expression,
+ const TString& title,
+ const TString& unit,
Double_t min, Double_t max, char type,
Bool_t normalized, void* external )
{
@@ -243,7 +253,7 @@ TMVA::VariableInfo& TMVA::DataSetInfo::AddSpectator( const TString& expression,
// the MV analysis
TString regexpr = expression; // remove possible blanks
regexpr.ReplaceAll(" ", "" );
- fSpectators.push_back(VariableInfo( regexpr, title, unit,
+ fSpectators.push_back(VariableInfo( regexpr, title, unit,
fSpectators.size()+1, type, external, min, max, normalized ));
fNeedsRebuilding = kTRUE;
return fSpectators.back();
@@ -282,7 +292,7 @@ void TMVA::DataSetInfo::SetWeightExpression( const TString& expr, const TString&
if (className != "") {
TMVA::ClassInfo* ci = AddClass(className);
ci->SetWeight( expr );
- }
+ }
else {
// no class name specified, set weight for all classes
if (fClasses.size()==0) {
@@ -321,7 +331,7 @@ void TMVA::DataSetInfo::AddCut( const TCut& cut, const TString& className )
// set the cut for the classes
if (className == "") { // if no className has been given set the cut for all the classes
for (std::vector<ClassInfo*>::iterator it = fClasses.begin(); it < fClasses.end(); it++) {
- const TCut& oldCut = (*it)->GetCut();
+ const TCut& oldCut = (*it)->GetCut();
(*it)->SetCut( oldCut+cut );
}
}
@@ -344,7 +354,7 @@ std::vector<TString> TMVA::DataSetInfo::GetListOfVariables() const
//_______________________________________________________________________
void TMVA::DataSetInfo::PrintCorrelationMatrix( const TString& className )
-{
+{
// calculates the correlation matrices for signal and background,
// prints them to standard output, and fills 2D histograms
Log() << kINFO << "Correlation matrix (" << className << "):" << Endl;
@@ -367,7 +377,7 @@ TH2* TMVA::DataSetInfo::CreateCorrelationMatrixHist( const TMatrixD* m,
for (UInt_t jvar=0; jvar<nvar; jvar++) {
(*tm)(ivar, jvar) = (*m)(ivar,jvar);
}
- }
+ }
TH2F* h2 = new TH2F( *tm );
h2->SetNameTitle( hName, hTitle );
@@ -379,7 +389,7 @@ TH2* TMVA::DataSetInfo::CreateCorrelationMatrixHist( const TMatrixD* m,
// present in percent, and round off digits
// also, use absolute value of correlation coefficient (ignore sign)
- h2->Scale( 100.0 );
+ h2->Scale( 100.0 );
for (UInt_t ibin=1; ibin<=nvar; ibin++) {
for (UInt_t jbin=1; jbin<=nvar; jbin++) {
h2->SetBinContent( ibin, jbin, Int_t(h2->GetBinContent( ibin, jbin )) );
@@ -404,24 +414,24 @@ TH2* TMVA::DataSetInfo::CreateCorrelationMatrixHist( const TMatrixD* m,
// gROOT->SetStyle("Plain");
// TStyle* gStyle = gROOT->GetStyle( "Plain" );
// gStyle->SetPalette( 1, 0 );
- // TPaletteAxis* paletteAxis
+ // TPaletteAxis* paletteAxis
// = (TPaletteAxis*)h2->GetListOfFunctions()->FindObject( "palette" );
// -------------------------------------------------------------------------------------
-
+
Log() << kDEBUG << "Created correlation matrix as 2D histogram: " << h2->GetName() << Endl;
-
+
return h2;
}
//_______________________________________________________________________
-TMVA::DataSet* TMVA::DataSetInfo::GetDataSet() const
+TMVA::DataSet* TMVA::DataSetInfo::GetDataSet() const
{
// returns data set
if (fDataSet==0 || fNeedsRebuilding) {
if(fDataSet!=0) ClearDataSet();
// fDataSet = DataSetManager::Instance().CreateDataSet(GetName()); //DSMTEST replaced by following lines
if( !fDataSetManager )
- Log() << kFATAL << "DataSetManager has not been set in DataSetInfo (GetDataSet() )." << Endl;
+ Log() << kFATAL << "DataSetManager has not been set in DataSetInfo (GetDataSet() )." << Endl;
fDataSet = fDataSetManager->CreateDataSet(GetName());
fNeedsRebuilding = kFALSE;
diff --git a/tmva/src/DataSetManager.cxx b/tmva/src/DataSetManager.cxx
index caa8119d719..63c81e0ea14 100644
--- a/tmva/src/DataSetManager.cxx
+++ b/tmva/src/DataSetManager.cxx
@@ -51,7 +51,7 @@ using std::endl;
// void TMVA::DataSetManager::DestroyInstance() { if (fgDSManager) { delete fgDSManager; fgDSManager=0; } } // DSMTEST removed
//_______________________________________________________________________
-TMVA::DataSetManager::DataSetManager( DataInputHandler& dataInput )
+TMVA::DataSetManager::DataSetManager( DataInputHandler& dataInput )
: fDataInput(dataInput),
fDataSetInfoCollection(),
fLogger( new MsgLogger("DataSetManager", kINFO) )
@@ -60,20 +60,20 @@ TMVA::DataSetManager::DataSetManager( DataInputHandler& dataInput )
}
//_______________________________________________________________________
-TMVA::DataSetManager::~DataSetManager()
+TMVA::DataSetManager::~DataSetManager()
{
// destructor
// fDataSetInfoCollection.SetOwner(); // DSMTEST --> created a segfault because the DataSetInfo-objects got deleted twice
- TMVA::DataSetFactory::destroyInstance();
+ TMVA::DataSetFactory::destroyInstance();
delete fLogger;
}
//_______________________________________________________________________
-TMVA::DataSet* TMVA::DataSetManager::CreateDataSet( const TString& dsiName )
+TMVA::DataSet* TMVA::DataSetManager::CreateDataSet( const TString& dsiName )
{
- // Creates the singleton dataset
+ // Creates the singleton dataset
DataSetInfo* dsi = GetDataSetInfo( dsiName );
if (!dsi) Log() << kFATAL << "DataSetInfo object '" << dsiName << "' not found" << Endl;
diff --git a/tmva/src/Factory.cxx b/tmva/src/Factory.cxx
index 1dee0b59184..c5d839953f2 100644
--- a/tmva/src/Factory.cxx
+++ b/tmva/src/Factory.cxx
@@ -621,8 +621,8 @@ void TMVA::Factory::PrepareTrainingAndTestTree( const TCut& cut, Int_t Ntrain, I
//_______________________________________________________________________
void TMVA::Factory::PrepareTrainingAndTestTree( const TCut& cut, const TString& opt )
-{
- // prepare the training and test trees
+{
+ // prepare the training and test trees
// -> same cuts for signal and background
SetInputTreesFromEventAssignTrees();
diff --git a/tmva/src/MethodBoost.cxx b/tmva/src/MethodBoost.cxx
index 1041ad144b0..018d8690442 100644
--- a/tmva/src/MethodBoost.cxx
+++ b/tmva/src/MethodBoost.cxx
@@ -83,6 +83,7 @@ TMVA::MethodBoost::MethodBoost( const TString& jobName,
, fMethodError(0)
, fOrigMethodError(0)
, fBoostWeight(0)
+ , fDetailedMonitoring(kFALSE)
, fADABoostBeta(0)
, fRandomSeed(0)
, fBoostedMethodTitle(methodTitle)
@@ -109,6 +110,7 @@ TMVA::MethodBoost::MethodBoost( DataSetInfo& dsi,
, fMethodError(0)
, fOrigMethodError(0)
, fBoostWeight(0)
+ , fDetailedMonitoring(kFALSE)
, fADABoostBeta(0)
, fRandomSeed(0)
, fBoostedMethodTitle("")
diff --git a/tmva/src/MethodCategory.cxx b/tmva/src/MethodCategory.cxx
index f1354d07d45..cfc7ef73dcb 100644
--- a/tmva/src/MethodCategory.cxx
+++ b/tmva/src/MethodCategory.cxx
@@ -1,5 +1,5 @@
// @(#)root/tmva $Id$
-// Author: Andreas Hoecker, Joerg Stelzer, Helge Voss, Kai Voss,Or Cohen, Eckhard von Toerne
+// Author: Andreas Hoecker, Joerg Stelzer, Helge Voss, Eckhard von Toerne
/**********************************************************************************
* Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
@@ -11,17 +11,17 @@
* Virtual base class for all MVA method *
* *
* Authors (alphabetical): *
- * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
- * Nadim Sah <Nadim.Sah@cern.ch> - Berlin, Germany *
- * Peter Speckmayer <Peter.Speckmazer@cern.ch> - CERN, Switzerland *
- * Joerg Stelzer <Joerg.Stelzer@cern.ch> - CERN, Switzerland *
- * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
- * Jan Therhaag <Jan.Therhaag@cern.ch> - U of Bonn, Germany *
- * Eckhard v. Toerne <evt@uni-bonn.de> - U of Bonn, Germany *
+ * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
+ * Nadim Sah <Nadim.Sah@cern.ch> - Berlin, Germany *
+ * Peter Speckmayer <Peter.Speckmazer@cern.ch> - CERN, Switzerland *
+ * Joerg Stelzer <Joerg.Stelzer@cern.ch> - MSU East Lansing, USA *
+ * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
+ * Jan Therhaag <Jan.Therhaag@cern.ch> - U of Bonn, Germany *
+ * Eckhard v. Toerne <evt@uni-bonn.de> - U of Bonn, Germany *
* *
* Copyright (c) 2005-2011: *
* CERN, Switzerland *
- * U. of Victoria, Canada *
+ * MSU East Lansing, USA *
* MPI-K Heidelberg, Germany *
* U. of Bonn, Germany *
* *
@@ -151,12 +151,15 @@ TMVA::IMethod* TMVA::MethodCategory::AddMethod( const TCut& theCut,
// std::cout << "set input done " << std::endl;
rearrangeTransformation->SetEnabled(kFALSE);
- IMethod* addedMethod = ClassifierFactory::Instance().Create(addedMethodName,GetJobName(),theTitle,dsi,theOptions);
+ IMethod* addedMethod = ClassifierFactory::Instance().Create(addedMethodName,
+ GetJobName(),
+ theTitle,
+ dsi,
+ theOptions);
MethodBase *method = (dynamic_cast<MethodBase*>(addedMethod));
-
if(method==0) return 0;
-
+
method->SetupMethod();
method->ParseOptions();
method->GetTransformationHandler().AddTransformation( rearrangeTransformation, -1 );
@@ -186,7 +189,7 @@ TMVA::IMethod* TMVA::MethodCategory::AddMethod( const TCut& theCut,
UInt_t newSpectatorIndex = primaryDSI.GetSpectatorInfos().size();
fCategorySpecIdx.push_back(newSpectatorIndex);
-
+
primaryDSI.AddSpectator( Form("%s_cat%i:=%s", GetName(),(int)fMethods.size(),theCut.GetTitle()),
Form("%s:%s",GetName(),method->GetName()),
"pass", 0, 0, 'C' );
@@ -358,15 +361,10 @@ void TMVA::MethodCategory::Train()
// specify the minimum # of training events and set 'classification'
const Int_t MinNoTrainingEvents = 10;
- // THIS NEEDS TO BE CHANGED:
-// TString what("Classification");
-// what.ToLower();
-// Types::EAnalysisType analysisType = ( what.CompareTo("regression")==0 ? Types::kRegression : Types::kClassification );
-
Types::EAnalysisType analysisType = GetAnalysisType();
// start the training
- Log() << kINFO << "Train all sub-classifiers for "
+ Log() << kINFO << "Train all sub-classifiers for "
<< (analysisType == Types::kRegression ? "Regression" : "Classification") << " ..." << Endl;
// don't do anything if no sub-classifier booked
@@ -374,7 +372,7 @@ void TMVA::MethodCategory::Train()
Log() << kINFO << "...nothing found to train" << Endl;
return;
}
-
+
std::vector<IMethod*>::iterator itrMethod;
// iterate over all booked sub-classifiers and train them
@@ -382,9 +380,9 @@ void TMVA::MethodCategory::Train()
MethodBase* mva = dynamic_cast<MethodBase*>(*itrMethod);
if(!mva) continue;
- mva->SetAnalysisType(GetAnalysisType());
- if (!mva->HasAnalysisType( analysisType,
- mva->DataInfo().GetNClasses(),
+ mva->SetAnalysisType( analysisType );
+ if (!mva->HasAnalysisType( analysisType,
+ mva->DataInfo().GetNClasses(),
mva->DataInfo().GetNTargets() ) ) {
Log() << kWARNING << "Method " << mva->GetMethodTypeName() << " is not capable of handling " ;
if (analysisType == Types::kRegression)
@@ -394,8 +392,6 @@ void TMVA::MethodCategory::Train()
itrMethod = fMethods.erase( itrMethod );
continue;
}
-
- mva->SetAnalysisType( analysisType );
if (mva->Data()->GetNTrainingEvents() >= MinNoTrainingEvents) {
Log() << kINFO << "Train method: " << mva->GetMethodName() << " for "
diff --git a/tmva/src/MethodCompositeBase.cxx b/tmva/src/MethodCompositeBase.cxx
index 11e255187fe..313edc22c79 100644
--- a/tmva/src/MethodCompositeBase.cxx
+++ b/tmva/src/MethodCompositeBase.cxx
@@ -1,32 +1,32 @@
// @(#)root/tmva $Id$
// Author: Andreas Hoecker, Joerg Stelzer, Helge Voss, Kai Voss,Or Cohen
-/**********************************************************************************
- * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
- * Package: TMVA *
- * Class : MethodCompositeBase *
- * Web : http://tmva.sourceforge.net *
- * *
- * Description: *
- * Virtual base class for all MVA method *
- * *
- * Authors (alphabetical): *
- * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
- * Joerg Stelzer <Joerg.Stelzer@cern.ch> - CERN, Switzerland *
- * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
- * Kai Voss <Kai.Voss@cern.ch> - U. of Victoria, Canada *
- * Or Cohen <orcohenor@gmail.com> - Weizmann Inst., Israel *
- * *
- * Copyright (c) 2005: *
- * CERN, Switzerland *
- * U. of Victoria, Canada *
- * MPI-K Heidelberg, Germany *
- * LAPP, Annecy, France *
- * *
- * Redistribution and use in source and binary forms, with or without *
- * modification, are permitted according to the terms listed in LICENSE *
- * (http://tmva.sourceforge.net/LICENSE) *
- **********************************************************************************/
+/*****************************************************************************
+ * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
+ * Package: TMVA *
+ * Class : MethodCompositeBase *
+ * Web : http://tmva.sourceforge.net *
+ * *
+ * Description: *
+ * Virtual base class for all MVA method *
+ * *
+ * Authors (alphabetical): *
+ * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
+ * Joerg Stelzer <Joerg.Stelzer@cern.ch> - MSU, USA *
+ * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
+ * Kai Voss <Kai.Voss@cern.ch> - U. of Victoria, Canada *
+ * Or Cohen <orcohenor@gmail.com> - Weizmann Inst., Israel *
+ * *
+ * Copyright (c) 2005: *
+ * CERN, Switzerland *
+ * U. of Victoria, Canada *
+ * MPI-K Heidelberg, Germany *
+ * LAPP, Annecy, France *
+ * *
+ * Redistribution and use in source and binary forms, with or without *
+ * modification, are permitted according to the terms listed in LICENSE *
+ * (http://tmva.sourceforge.net/LICENSE) *
+ *****************************************************************************/
//_______________________________________________________________________
//
@@ -58,7 +58,7 @@ using std::vector;
ClassImp(TMVA::MethodCompositeBase)
//_______________________________________________________________________
-TMVA::MethodCompositeBase::MethodCompositeBase( const TString& jobName,
+TMVA::MethodCompositeBase::MethodCompositeBase( const TString& jobName,
Types::EMVA methodType,
const TString& methodTitle,
DataSetInfo& theData,
@@ -71,7 +71,7 @@ TMVA::MethodCompositeBase::MethodCompositeBase( const TString& jobName,
//_______________________________________________________________________
TMVA::MethodCompositeBase::MethodCompositeBase( Types::EMVA methodType,
DataSetInfo& dsi,
- const TString& weightFile,
+ const TString& weightFile,
TDirectory* theTargetDir )
: TMVA::MethodBase( methodType, dsi, weightFile, theTargetDir ),
fMethodIndex(0)
@@ -85,7 +85,7 @@ TMVA::IMethod* TMVA::MethodCompositeBase::GetMethod( const TString &methodTitle
vector<IMethod*>::const_iterator itrMethodEnd = fMethods.end();
for (; itrMethod != itrMethodEnd; itrMethod++) {
- MethodBase* mva = dynamic_cast<MethodBase*>(*itrMethod);
+ MethodBase* mva = dynamic_cast<MethodBase*>(*itrMethod);
if ( (mva->GetMethodName())==methodTitle ) return mva;
}
return 0;
@@ -102,26 +102,26 @@ TMVA::IMethod* TMVA::MethodCompositeBase::GetMethod( const Int_t index ) const
//_______________________________________________________________________
-void TMVA::MethodCompositeBase::AddWeightsXMLTo( void* parent ) const
+void TMVA::MethodCompositeBase::AddWeightsXMLTo( void* parent ) const
{
void* wght = gTools().AddChild(parent, "Weights");
gTools().AddAttr( wght, "NMethods", fMethods.size() );
- for (UInt_t i=0; i< fMethods.size(); i++)
+ for (UInt_t i=0; i< fMethods.size(); i++)
{
void* methxml = gTools().AddChild( wght, "Method" );
MethodBase* method = dynamic_cast<MethodBase*>(fMethods[i]);
- gTools().AddAttr(methxml,"Index", i );
- gTools().AddAttr(methxml,"Weight", fMethodWeight[i]);
+ gTools().AddAttr(methxml,"Index", i );
+ gTools().AddAttr(methxml,"Weight", fMethodWeight[i]);
gTools().AddAttr(methxml,"MethodSigCut", method->GetSignalReferenceCut());
gTools().AddAttr(methxml,"MethodSigCutOrientation", method->GetSignalReferenceCutOrientation());
gTools().AddAttr(methxml,"MethodTypeName", method->GetMethodTypeName());
- gTools().AddAttr(methxml,"MethodName", method->GetMethodName() );
+ gTools().AddAttr(methxml,"MethodName", method->GetMethodName() );
gTools().AddAttr(methxml,"JobName", method->GetJobName());
- gTools().AddAttr(methxml,"Options", method->GetOptions());
+ gTools().AddAttr(methxml,"Options", method->GetOptions());
if (method->fTransformationPointer)
- gTools().AddAttr(methxml,"UseMainMethodTransformation", TString("true"));
+ gTools().AddAttr(methxml,"UseMainMethodTransformation", TString("true"));
else
- gTools().AddAttr(methxml,"UseMainMethodTransformation", TString("false"));
+ gTools().AddAttr(methxml,"UseMainMethodTransformation", TString("false"));
method->AddWeightsXMLTo(methxml);
}
}
@@ -132,14 +132,14 @@ TMVA::MethodCompositeBase::~MethodCompositeBase( void )
// delete methods
vector<IMethod*>::iterator itrMethod = fMethods.begin();
for (; itrMethod != fMethods.end(); itrMethod++) {
- Log() << kVERBOSE << "Delete method: " << (*itrMethod)->GetName() << Endl;
+ Log() << kVERBOSE << "Delete method: " << (*itrMethod)->GetName() << Endl;
delete (*itrMethod);
}
fMethods.clear();
}
//_______________________________________________________________________
-void TMVA::MethodCompositeBase::ReadWeightsFromXML( void* wghtnode )
+void TMVA::MethodCompositeBase::ReadWeightsFromXML( void* wghtnode )
{
// XML streamer
UInt_t nMethods;
@@ -162,13 +162,13 @@ void TMVA::MethodCompositeBase::ReadWeightsFromXML( void* wghtnode )
Bool_t rerouteTransformation = kFALSE;
if (gTools().HasAttr( ch, "UseMainMethodTransformation")) {
- TString rerouteString("");
- gTools().ReadAttr( ch, "UseMainMethodTransformation", rerouteString );
- rerouteString.ToLower();
- if (rerouteString=="true")
- rerouteTransformation=kTRUE;
+ TString rerouteString("");
+ gTools().ReadAttr( ch, "UseMainMethodTransformation", rerouteString );
+ rerouteString.ToLower();
+ if (rerouteString=="true")
+ rerouteTransformation=kTRUE;
}
-
+
//remove trailing "~" to signal that options have to be reused
optionString.ReplaceAll("~","");
//ignore meta-options for method Boost
diff --git a/tmva/src/PDEFoamVect.cxx b/tmva/src/PDEFoamVect.cxx
index d26b012f68f..94dfb64f78d 100644
--- a/tmva/src/PDEFoamVect.cxx
+++ b/tmva/src/PDEFoamVect.cxx
@@ -35,7 +35,7 @@
using namespace std;
-//#define SW2 std::setw(12)
+#define SW2 std::setprecision(7) << std::setw(12)
ClassImp(TMVA::PDEFoamVect)
@@ -200,13 +200,11 @@ TMVA::PDEFoamVect& TMVA::PDEFoamVect::operator =(Double_t x)
//_____________________________________________________________________
void TMVA::PDEFoamVect::Print(Option_t *option) const
{
- streamsize wid = cout.width(); // saving current field width
// Printout of all vector components
if(!option) Error( "Print ", "No option set \n");
cout << "(";
for(Int_t i=0; i<fDim-1; i++)
- cout << std::setw(12) << *(fCoords+i) << ",";
- cout << std::setw(12) << *(fCoords+fDim-1);
+ cout << SW2 << *(fCoords+i) << ",";
+ cout << SW2 << *(fCoords+fDim-1);
cout << ")";
- cout.width(wid);
}
diff --git a/tmva/src/RegressionVariance.cxx b/tmva/src/RegressionVariance.cxx
index a4bbe0cc458..d0ab388a66d 100644
--- a/tmva/src/RegressionVariance.cxx
+++ b/tmva/src/RegressionVariance.cxx
@@ -61,7 +61,7 @@ Double_t TMVA::RegressionVariance::GetSeparationGain(const Double_t &nLeft,
Double_t parentIndex = nTot * this->GetSeparationIndex(nTot,targetTot,target2Tot);
Double_t leftIndex = ( (nTot - nLeft) * this->GetSeparationIndex(nTot-nLeft,targetTot-targetLeft,target2Tot-target2Left) );
Double_t rightIndex = nLeft * this->GetSeparationIndex(nLeft,targetLeft,target2Left);
-
+
// return 1/ (leftIndex + rightIndex);
return (parentIndex - leftIndex - rightIndex)/(parentIndex);
}
diff --git a/tmva/src/VariableNormalizeTransform.cxx b/tmva/src/VariableNormalizeTransform.cxx
index eebe6bced5b..ea6fc57ca74 100644
--- a/tmva/src/VariableNormalizeTransform.cxx
+++ b/tmva/src/VariableNormalizeTransform.cxx
@@ -401,6 +401,7 @@ void TMVA::VariableNormalizeTransform::ReadFromXML( void* trfnode )
gTools().ReadAttr(trfnode, "NVariables", nvars);
// coverity[tainted_data_argument]
gTools().ReadAttr(trfnode, "NTargets", ntgts);
+ // coverity[tainted_data_argument]
for( UInt_t ivar = 0; ivar < nvars; ++ivar ){
fGet.push_back(std::make_pair<Char_t,UInt_t>('v',ivar));
diff --git a/tmva/test/TMVAClassificationCategory.cxx b/tmva/test/TMVAClassificationCategory.cxx
index 60433d839ca..f2768c7a60b 100644
--- a/tmva/test/TMVAClassificationCategory.cxx
+++ b/tmva/test/TMVAClassificationCategory.cxx
@@ -21,11 +21,11 @@
* macros (simply say: root -l <../macros/macro.C>), which can be conveniently *
* invoked through a GUI launched by the command *
* *
- * root -l TMVAGui.C *
+ * root -l TMVAGui.C *
**********************************************************************************/
#include <cstdlib>
-#include <iostream>
+#include <iostream>
#include <map>
#include <string>
@@ -44,12 +44,13 @@
// two types of category methods are implemented
Bool_t UseOffsetMethod = kTRUE;
-int main( int argc, char** argv )
+int main( int argc, char** argv )
{
//---------------------------------------------------------------
- // Example for usage of different event categories with classifiers
+ // Example for usage of different event categories with classifiers
- std::cout << std::endl << "==> Start TMVAClassificationCategory" << std::endl;
+ std::cout << std::endl
+ << "==> Start TMVAClassificationCategory" << std::endl;
bool batchMode = false;
@@ -62,7 +63,9 @@ int main( int argc, char** argv )
std::string factoryOptions( "!V:!Silent:Transformations=I;D;P;G,D" );
if (batchMode) factoryOptions += ":!Color:!DrawProgressBar";
- TMVA::Factory *factory = new TMVA::Factory( "TMVAClassificationCategory", outputFile, factoryOptions );
+ TMVA::Factory *factory = new TMVA::Factory( "TMVAClassificationCategory",
+ outputFile,
+ factoryOptions );
// Define the input variables used for the MVA training
factory->AddVariable( "var1", 'F' );
@@ -70,16 +73,17 @@ int main( int argc, char** argv )
factory->AddVariable( "var3", 'F' );
factory->AddVariable( "var4", 'F' );
- // You can add so-called "Spectator variables", which are not used in the MVA training,
- // but will appear in the final "TestTree" produced by TMVA. This TestTree will contain the
- // input variables, the response values of all trained MVAs, and the spectator variables
+ // You can add so-called "Spectator variables", which are not used
+ // in the MVA training, but will appear in the final "TestTree"
+ // produced by TMVA. This TestTree will contain the input
+ // variables, the response values of all trained MVAs, and the
+ // spectator variables
factory->AddSpectator( "eta" );
// Load the signal and background event samples from ROOT trees
TFile *input(0);
- TString fname( "" );
- if (UseOffsetMethod) fname = "data/toy_sigbkg_categ_offset.root";
- else fname = "data/toy_sigbkg_categ_varoff.root";
+ TString fname = UseOffsetMethod ? "data/toy_sigbkg_categ_offset.root" : "data/toy_sigbkg_categ_varoff.root";
+
if (!gSystem->AccessPathName( fname )) {
// first we try to find tmva_example.root in the local directory
std::cout << "--- TMVAClassificationCategory: Accessing " << fname << std::endl;
@@ -108,7 +112,7 @@ int main( int argc, char** argv )
// Tell the factory how to use the training and testing events
factory->PrepareTrainingAndTestTree( mycuts, mycutb,
- "nTrain_Signal=0:nTrain_Background=0:SplitMode=Random:NormMode=NumEvents:!V" );
+ "nTest_Signal=5500:nTrain_Background=3400:SplitMode=Random:NormMode=NumEvents" );
// ---- Book MVA methods
@@ -117,14 +121,14 @@ int main( int argc, char** argv )
// Likelihood
factory->BookMethod( TMVA::Types::kLikelihood, "Likelihood",
- "!H:!V:TransformOutput:PDFInterpol=Spline2:NSmoothSig[0]=20:NSmoothBkg[0]=20:NSmoothBkg[1]=10:NSmooth=1:NAvEvtPerBin=50" );
+ "!H:!V:TransformOutput:PDFInterpol=Spline2:NSmoothSig[0]=20:NSmoothBkg[0]=20:NSmoothBkg[1]=10:NSmooth=1:NAvEvtPerBin=50" );
// --- Categorised classifier
TMVA::MethodCategory* mcat = 0;
// The variable sets
TString theCat1Vars = "var1:var2:var3:var4";
- TString theCat2Vars = (UseOffsetMethod ? "var1:var2:var3:var4" : "var1:var2:var3");
+ TString theCat2Vars = UseOffsetMethod ? "var1:var2:var3:var4" : "var1:var2:var3";
// Fisher with categories
TMVA::MethodBase* fiCat = factory->BookMethod( TMVA::Types::kCategory, "FisherCat","" );
diff --git a/tmva/test/setup.sh b/tmva/test/setup.sh
index 17fd0edfe7d..8edab3e449b 100755
--- a/tmva/test/setup.sh
+++ b/tmva/test/setup.sh
@@ -23,6 +23,9 @@ if [ ! $ROOTSYS ]; then
return 1
fi
+export TMVASYS=$PWD
+
+
# On MacOS X $DYLD_LIBRARY_PATH has to be modified, so:
if [[ `root-config --platform` == "macosx" ]]; then
--
GitLab