Logo ROOT   6.21/01
Reference Guide
DataLoader.h
Go to the documentation of this file.
1 // @(#)root/tmva $Id$
2 // Author: Andreas Hoecker, Peter Speckmayer, Joerg Stelzer, Helge Voss, Kai Voss, Eckhard von Toerne, Jan Therhaag, Omar Zapata, Lorenzo Moneta, Sergei Gleyzer
3 //NOTE: Based on TMVA::Factory
4 
5 /**********************************************************************************
6  * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
7  * Package: TMVA *
8  * Class : DataLoader *
9  * Web : http://tmva.sourceforge.net *
10  * *
11  * Description: *
12  * This is a class to load datasets into every booked method *
13  * *
14  * Authors (alphabetical): *
15  * Lorenzo Moneta <Lorenzo.Moneta@cern.ch> - CERN, Switzerland *
16  * Omar Zapata <andresete.chaos@gmail.com> - ITM/UdeA, Colombia *
17  * Sergei Gleyzer<sergei.gleyzer@cern.ch> - CERN, Switzerland *
18  * *
19  * Copyright (c) 2005-2011: *
20  * CERN, Switzerland *
21  * ITM/UdeA, Colombia *
22  * *
23  * Redistribution and use in source and binary forms, with or without *
24  * modification, are permitted according to the terms listed in LICENSE *
25  * (http://tmva.sourceforge.net/LICENSE) *
26  **********************************************************************************/
27 
28 #ifndef ROOT_TMVA_DataLoader
29 #define ROOT_TMVA_DataLoader
30 
31 #include <string>
32 #include <vector>
33 #include <map>
34 #include "TCut.h"
35 
36 #include "TMVA/Configurable.h"
37 #include "TMVA/Types.h"
38 #include "TMVA/DataSet.h"
39 
40 class TFile;
41 class TTree;
42 class TH2;
43 
44 namespace TMVA {
45 
46  class CvSplit;
47  class DataInputHandler;
48  class DataSetInfo;
49  class DataSetManager;
50  class VariableTransformBase;
51 
52  class DataLoader : public Configurable {
53  public:
54 
55  DataLoader(TString thedlName="default");
56 
57  // default destructor
58  virtual ~DataLoader();
59 
60 
61  // add events to training and testing trees
62  void AddSignalTrainingEvent ( const std::vector<Double_t>& event, Double_t weight = 1.0 );
63  void AddBackgroundTrainingEvent( const std::vector<Double_t>& event, Double_t weight = 1.0 );
64  void AddSignalTestEvent ( const std::vector<Double_t>& event, Double_t weight = 1.0 );
65  void AddBackgroundTestEvent ( const std::vector<Double_t>& event, Double_t weight = 1.0 );
66  void AddTrainingEvent( const TString& className, const std::vector<Double_t>& event, Double_t weight );
67  void AddTestEvent ( const TString& className, const std::vector<Double_t>& event, Double_t weight );
68  void AddEvent ( const TString& className, Types::ETreeType tt, const std::vector<Double_t>& event, Double_t weight );
71 
73  DataSetInfo& AddDataSet( const TString& );
75  DataLoader* VarTransform(TString trafoDefinition);
76 
77  // special case: signal/background
78 
79  // Data input related
80  void SetInputTrees( const TString& signalFileName, const TString& backgroundFileName,
81  Double_t signalWeight=1.0, Double_t backgroundWeight=1.0 );
82  void SetInputTrees( TTree* inputTree, const TCut& SigCut, const TCut& BgCut );
83  // Set input trees at once
84  void SetInputTrees( TTree* signal, TTree* background,
85  Double_t signalWeight=1.0, Double_t backgroundWeight=1.0) ;
86 
87  void AddSignalTree( TTree* signal, Double_t weight=1.0, Types::ETreeType treetype = Types::kMaxTreeType );
88  void AddSignalTree( TString datFileS, Double_t weight=1.0, Types::ETreeType treetype = Types::kMaxTreeType );
89  void AddSignalTree( TTree* signal, Double_t weight, const TString& treetype );
90 
91  // ... depreciated, kept for backwards compatibility
92  void SetSignalTree( TTree* signal, Double_t weight=1.0);
93 
94  void AddBackgroundTree( TTree* background, Double_t weight=1.0, Types::ETreeType treetype = Types::kMaxTreeType );
95  void AddBackgroundTree( TString datFileB, Double_t weight=1.0, Types::ETreeType treetype = Types::kMaxTreeType );
96  void AddBackgroundTree( TTree* background, Double_t weight, const TString & treetype );
97 
98  // ... depreciated, kept for backwards compatibility
99  void SetBackgroundTree( TTree* background, Double_t weight=1.0 );
100 
101  void SetSignalWeightExpression( const TString& variable );
102  void SetBackgroundWeightExpression( const TString& variable );
103 
104  // special case: regression
105  void AddRegressionTree( TTree* tree, Double_t weight = 1.0,
106  Types::ETreeType treetype = Types::kMaxTreeType ) {
107  AddTree( tree, "Regression", weight, "", treetype );
108  }
109 
110  // general
111 
112  // Data input related
113  void SetTree( TTree* tree, const TString& className, Double_t weight ); // depreciated
114  void AddTree( TTree* tree, const TString& className, Double_t weight=1.0,
115  const TCut& cut = "",
117  void AddTree( TTree* tree, const TString& className, Double_t weight, const TCut& cut, const TString& treeType );
118 
119  // set input variable
120  void SetInputVariables ( std::vector<TString>* theVariables ); // deprecated
121 
122  void AddVariable ( const TString& expression, const TString& title, const TString& unit,
123  char type='F', Double_t min = 0, Double_t max = 0 );
124  void AddVariable ( const TString& expression, char type='F',
125  Double_t min = 0, Double_t max = 0 );
126 
127  // NEW: add an array of variables (e.g. for image data) with the provided size
128  void AddVariablesArray(const TString &expression, int size, char type = 'F',
129  Double_t min = 0, Double_t max = 0);
130 
131 
132  void AddTarget ( const TString& expression, const TString& title = "", const TString& unit = "",
133  Double_t min = 0, Double_t max = 0 );
134  void AddRegressionTarget( const TString& expression, const TString& title = "", const TString& unit = "",
135  Double_t min = 0, Double_t max = 0 )
136  {
137  AddTarget( expression, title, unit, min, max );
138  }
139  void AddSpectator ( const TString& expression, const TString& title = "", const TString& unit = "",
140  Double_t min = 0, Double_t max = 0 );
141 
142  // set weight for class
143  void SetWeightExpression( const TString& variable, const TString& className = "" );
144 
145  // set cut for class
146  void SetCut( const TString& cut, const TString& className = "" );
147  void SetCut( const TCut& cut, const TString& className = "" );
148  void AddCut( const TString& cut, const TString& className = "" );
149  void AddCut( const TCut& cut, const TString& className = "" );
150 
151 
152  // prepare input tree for training
153  void PrepareTrainingAndTestTree( const TCut& cut, const TString& splitOpt );
154  void PrepareTrainingAndTestTree( TCut sigcut, TCut bkgcut, const TString& splitOpt );
155 
156  // ... deprecated, kept for backwards compatibility
157  void PrepareTrainingAndTestTree( const TCut& cut, Int_t Ntrain, Int_t Ntest = -1 );
158 
159  void PrepareTrainingAndTestTree( const TCut& cut, Int_t NsigTrain, Int_t NbkgTrain, Int_t NsigTest, Int_t NbkgTest,
160  const TString& otherOpt="SplitMode=Random:!V" );
161 
162  // Cross validation
163  void MakeKFoldDataSet(CvSplit & s);
166 
168 
169  TH2* GetCorrelationMatrix(const TString& className);
170 
171  //Copy method use in VI and CV DEPRECATED: you can just call Clone DataLoader *dl2=(DataLoader *)dl1->Clone("dl2")
173  friend void DataLoaderCopy(TMVA::DataLoader* des, TMVA::DataLoader* src);
175 
176  private:
177 
178 
181 
182 
183  private:
184 
185  // data members
186 
187 
189 
190 
192 
193  std::vector<TMVA::VariableTransformBase*> fDefaultTrfs; // list of transformations on default DataSet
194 
195  // cd to local directory
196  TString fOptions; // option string given by construction (presently only "V")
197  TString fTransformations; // List of transformations to test
198  Bool_t fVerbose; // verbose mode
199 
200  // flag determining the way training and test data are assigned to DataLoader
204  DataAssignType fDataAssignType; // flags for data assigning
205  std::vector<TTree*> fTrainAssignTree; // for each class: tmp tree if user wants to assign the events directly
206  std::vector<TTree*> fTestAssignTree; // for each class: tmp tree if user wants to assign the events directly
207 
208  Int_t fATreeType = 0; // type of event (=classIndex)
209  Float_t fATreeWeight = 0.0; // weight of the event
210  std::vector<Float_t> fATreeEvent; // event variables
211 
212  Types::EAnalysisType fAnalysisType; // the training type
213 
214  protected:
215 
216  ClassDef(DataLoader,4);
217  };
219 } // namespace TMVA
220 
221 #endif
222 
void AddBackgroundTree(TTree *background, Double_t weight=1.0, Types::ETreeType treetype=Types::kMaxTreeType)
number of signal events (used to compute significance)
Definition: DataLoader.cxx:387
DataSetManager * fDataSetManager
Definition: DataLoader.h:188
virtual ~DataLoader()
Definition: DataLoader.cxx:82
auto * tt
Definition: textangle.C:16
void AddTrainingEvent(const TString &className, const std::vector< Double_t > &event, Double_t weight)
add signal training event
Definition: DataLoader.cxx:245
std::vector< TMVA::VariableTransformBase * > fDefaultTrfs
Definition: DataLoader.h:193
DataLoader(TString thedlName="default")
Definition: DataLoader.cxx:66
float Float_t
Definition: RtypesCore.h:53
void AddRegressionTarget(const TString &expression, const TString &title="", const TString &unit="", Double_t min=0, Double_t max=0)
Definition: DataLoader.h:134
DataSetInfo & GetDataSetInfo()
Definition: DataLoader.cxx:123
A ROOT file is a suite of consecutive data records (TKey instances) with a well defined format...
Definition: TFile.h:48
EAnalysisType
Definition: Types.h:127
TTree * CreateEventAssignTrees(const TString &name)
create the data assignment tree (for event-wise data assignment by user)
Definition: DataLoader.cxx:180
DataSetInfo & DefaultDataSetInfo()
default creation
Definition: DataLoader.cxx:518
DataLoader * VarTransform(TString trafoDefinition)
Transforms the variables and return a new DataLoader with the transformed variables.
Definition: DataLoader.cxx:132
Basic string class.
Definition: TString.h:131
int Int_t
Definition: RtypesCore.h:41
bool Bool_t
Definition: RtypesCore.h:59
void DataLoaderCopy(TMVA::DataLoader *des, TMVA::DataLoader *src)
void SetBackgroundTree(TTree *background, Double_t weight=1.0)
Definition: DataLoader.cxx:424
DataInputHandler * fDataInputHandler
Definition: DataLoader.h:191
Types::EAnalysisType fAnalysisType
Definition: DataLoader.h:212
void AddBackgroundTestEvent(const std::vector< Double_t > &event, Double_t weight=1.0)
add signal training event
Definition: DataLoader.cxx:237
TH2 * GetCorrelationMatrix(const TString &className)
returns the correlation matrix of datasets
Definition: DataLoader.cxx:702
void AddVariable(const TString &expression, const TString &title, const TString &unit, char type='F', Double_t min=0, Double_t max=0)
user inserts discriminating variable in data set info
Definition: DataLoader.cxx:470
void MakeKFoldDataSet(CvSplit &s)
Function required to split the training and testing datasets into a number of folds.
Definition: DataLoader.cxx:647
#define ClassDef(name, id)
Definition: Rtypes.h:326
void AddTestEvent(const TString &className, const std::vector< Double_t > &event, Double_t weight)
add signal test event
Definition: DataLoader.cxx:253
void SetInputTrees(const TString &signalFileName, const TString &backgroundFileName, Double_t signalWeight=1.0, Double_t backgroundWeight=1.0)
Definition: DataLoader.cxx:449
void SetTree(TTree *tree, const TString &className, Double_t weight)
set background tree
Definition: DataLoader.cxx:432
void AddVariablesArray(const TString &expression, int size, char type='F', Double_t min=0, Double_t max=0)
user inserts discriminating array of variables in data set info in case input tree provides an array ...
Definition: DataLoader.cxx:489
Class that contains all the data information.
Definition: DataSetInfo.h:60
static constexpr double s
void SetInputVariables(std::vector< TString > *theVariables)
fill input variables in data set
Definition: DataLoader.cxx:526
DataSetInfo & AddDataSet(DataSetInfo &)
Definition: DataLoader.cxx:105
void AddCut(const TString &cut, const TString &className="")
Definition: DataLoader.cxx:573
A specialized string object used for TTree selections.
Definition: TCut.h:25
void SetInputTreesFromEventAssignTrees()
assign event-wise local trees to data set
Definition: DataLoader.cxx:304
Float_t fATreeWeight
Definition: DataLoader.h:209
DataInputHandler & DataInput()
Definition: DataLoader.h:174
Service class for 2-Dim histogram classes.
Definition: TH2.h:30
Class that contains all the data information.
unsigned int UInt_t
Definition: RtypesCore.h:42
std::vector< TTree * > fTestAssignTree
Definition: DataLoader.h:206
Bool_t UserAssignEvents(UInt_t clIndex)
Definition: DataLoader.cxx:296
std::vector< Float_t > fATreeEvent
Definition: DataLoader.h:210
void AddRegressionTree(TTree *tree, Double_t weight=1.0, Types::ETreeType treetype=Types::kMaxTreeType)
Definition: DataLoader.h:105
DataLoader * MakeCopy(TString name)
Copy method use in VI and CV.
Definition: DataLoader.cxx:676
const DataSetInfo & GetDefaultDataSetInfo()
Definition: DataLoader.h:167
void AddTree(TTree *tree, const TString &className, Double_t weight=1.0, const TCut &cut="", Types::ETreeType tt=Types::kMaxTreeType)
Definition: DataLoader.cxx:336
void PrepareTrainingAndTestTree(const TCut &cut, const TString &splitOpt)
prepare the training and test trees -> same cuts for signal and background
Definition: DataLoader.cxx:617
DataAssignType fDataAssignType
Definition: DataLoader.h:204
TString fTransformations
Definition: DataLoader.h:197
double Double_t
Definition: RtypesCore.h:55
void AddEvent(const TString &className, Types::ETreeType tt, const std::vector< Double_t > &event, Double_t weight)
add event vector event : the order of values is: variables + targets + spectators ...
Definition: DataLoader.cxx:262
Class that contains all the data information.
void SetBackgroundWeightExpression(const TString &variable)
Definition: DataLoader.cxx:541
int type
Definition: TGX11.cxx:120
void AddTarget(const TString &expression, const TString &title="", const TString &unit="", Double_t min=0, Double_t max=0)
user inserts target in data set info
Definition: DataLoader.cxx:497
void SetWeightExpression(const TString &variable, const TString &className="")
Definition: DataLoader.cxx:548
void AddBackgroundTrainingEvent(const std::vector< Double_t > &event, Double_t weight=1.0)
add signal training event
Definition: DataLoader.cxx:229
void PrepareFoldDataSet(CvSplit &s, UInt_t foldNumber, Types::ETreeType tt=Types::kTraining)
Function for assigning the correct folds to the testing or training set.
Definition: DataLoader.cxx:655
void SetSignalWeightExpression(const TString &variable)
Definition: DataLoader.cxx:534
create variable transformations
std::vector< TTree * > fTrainAssignTree
Definition: DataLoader.h:205
void AddSignalTestEvent(const std::vector< Double_t > &event, Double_t weight=1.0)
add signal testing event
Definition: DataLoader.cxx:221
void AddSignalTrainingEvent(const std::vector< Double_t > &event, Double_t weight=1.0)
add signal training event
Definition: DataLoader.cxx:213
friend void DataLoaderCopy(TMVA::DataLoader *des, TMVA::DataLoader *src)
void RecombineKFoldDataSet(CvSplit &s, Types::ETreeType tt=Types::kTraining)
Recombines the dataset.
Definition: DataLoader.cxx:668
TString fOptions
Definition: DataLoader.h:196
void SetSignalTree(TTree *signal, Double_t weight=1.0)
Definition: DataLoader.cxx:417
Definition: tree.py:1
A TTree represents a columnar dataset.
Definition: TTree.h:72
void AddSignalTree(TTree *signal, Double_t weight=1.0, Types::ETreeType treetype=Types::kMaxTreeType)
number of signal events (used to compute significance)
Definition: DataLoader.cxx:356
void SetCut(const TString &cut, const TString &className="")
Definition: DataLoader.cxx:560
char name[80]
Definition: TGX11.cxx:109
void AddSpectator(const TString &expression, const TString &title="", const TString &unit="", Double_t min=0, Double_t max=0)
user inserts target in data set info
Definition: DataLoader.cxx:509