ATLAS Offline Software
Public Types | Public Member Functions | Private Member Functions | Private Attributes | List of all members
TJetNet Class Reference

#include <TJetNet.h>

Inheritance diagram for TJetNet:
Collaboration diagram for TJetNet:

Public Types

enum  TActivationFunction {
  afSigmoid = 1, afTanh = 2, afExp = 3, afLinear = 4,
  afSigmoidEntropy = 5
}
 

Public Member Functions

 TJetNet (void)
 
 TJetNet (Int_t aTestCount, Int_t aTrainCount, const Int_t aLayersCnt, const Int_t *aLayers)
 
virtual ~TJetNet (void)
 
void Print (void)
 
Int_t GetTrainSetCnt (void) const
 
Int_t GetTestSetCnt (void) const
 
Int_t GetInputDim (void) const
 
Int_t GetHiddenLayerDim (void) const
 
Int_t GetHiddenLayerSize (Int_t number) const
 
Int_t GetOutputDim (void) const
 
void SetInputTrainSet (Int_t aPatternInd, Int_t aInputInd, Double_t aValue)
 
void SetOutputTrainSet (Int_t aPatternInd, Int_t aOutputInd, Double_t aValue)
 
void SetInputTestSet (Int_t aPatternInd, Int_t aInputInd, Double_t aValue)
 
void SetOutputTestSet (Int_t aPatternInd, Int_t aOutputInd, Double_t aValue)
 
void SetEventWeightTrainSet (Int_t aPatternInd, Double_t aValue)
 
void SetEventWeightTestSet (Int_t aPatternInd, Double_t aValue)
 
Double_t GetInputTrainSet (Int_t aPatternInd, Int_t aInputInd)
 
Double_t GetOutputTrainSet (Int_t aPatternInd, Int_t aOutputInd)
 
Double_t GetInputTestSet (Int_t aPatternInd, Int_t aInputInd)
 
Double_t GetOutputTestSet (Int_t aPatternInd, Int_t aOutputInd)
 
Double_t GetEventWeightTrainSet (Int_t aPatternInd)
 
Double_t GetEventWeightTestSet (Int_t aPatternInd)
 
Double_t GetWeight (Int_t aLayerInd, Int_t aNodeInd, Int_t aConnectedNodeInd) const
 
Double_t GetThreshold (Int_t aLayerInd, Int_t aNodeInd) const
 
Int_t GetEpochs (void)
 
void SetEpochs (const Int_t aEpochs)
 
void Init (void)
 
Double_t Train (void)
 
Int_t Epoch (void)
 
Double_t Test (void)
 
Double_t TestBTAG (void)
 
void Shuffle (Bool_t aShuffleTrainSet=true, Bool_t aShuffleTestSet=true)
 
void SaveDataAscii (TString aFileName="jndata.dat")
 
void SaveDataRoot (TString aFileName="jndata.root")
 
void LoadDataAscii (TString aFileName="jndata.dat")
 
void LoadDataRoot (TString aFileName="jndata.root")
 
void DumpToFile (TString aFileName="fort.8")
 
void ReadFromFile (TString aFileName="fort.8")
 
Double_t GetOutput (Int_t aIndex=0)
 
void SetInputs (Int_t aIndex=0, Double_t aValue=0.0)
 
void Evaluate (Int_t aPattern)
 
void Evaluate ()
 
void writeNetworkInfo (Int_t typeOfInfo=0)
 
Int_t GetLayerCount (void)
 
Int_t GetUnitCount (Int_t aLayer)
 
void SelectiveFields (Int_t aLayerA, Int_t aNodeA1, Int_t aNodeA2, Int_t aNodeB1, Int_t aNodeB2, Int_t aSwitch=0)
 
void SetUpdatesPerEpoch (Int_t aValue)
 
void SetUpdatingProcedure (Int_t aValue)
 
void SetErrorMeasure (Int_t aValue)
 
void SetActivationFunction (Int_t aValue)
 
void SetActivationFunction (Int_t aValue, Int_t layerN)
 
void SetPatternsPerUpdate (Int_t aValue)
 
void SetLearningRate (Double_t aValue)
 
void SetMomentum (Double_t aValue)
 
void SetInitialWeightsWidth (Double_t aValue)
 
void SetLearningRateDecrease (Double_t aValue)
 
void SetPottsUnits (Int_t aValue)
 
Int_t GetUpdatesPerEpoch (void)
 
Int_t GetUpdatingProcedure (void)
 
Int_t GetErrorMeasure (void)
 
Int_t GetActivationFunction (void) const
 
Int_t GetActivationFunction (Int_t layerN) const
 
Int_t GetPatternsPerUpdate (void)
 
Double_t GetLearningRate (void)
 
Double_t GetMomentum (void)
 
Double_t GetInitialWeightsWidth (void)
 
Double_t GetLearningRateDecrease (void)
 
Int_t GetPottsUnits ()
 
void LockInit (void)
 
void UnlockInit (void)
 
Int_t GetMSTJN (Int_t aIndex)
 
Double_t GetPARJN (Int_t aIndex)
 
void SetMSTJN (Int_t aIndex, Int_t aValue)
 
void SetPARJN (Int_t aIndex, Double_t aValue)
 
void Normalize (void)
 
void Randomize (void)
 
TTrainedNetwork * createTrainedNetwork () const
 
void readBackTrainedNetwork (const TTrainedNetwork *)
 
void NormalizeOutput (bool isTrue=true)
 

Private Member Functions

void mSetWeight (Double_t weight, Int_t aLayerInd, Int_t aNodeInd, Int_t aConnectedNodeInd)
 
void mSetThreshold (Double_t threshold, Int_t aLayerInd, Int_t aNodeInd)
 
Int_t CopyFile (TString aSrcFile, TString aDestFile)
 
void Reinitialize (void)
 

Private Attributes

TActivationFunction m_enActFunction {}
 
Int_t m_LayerCount
 
Int_t * m_pLayers
 
TNeuralDataSet * m_pInputTrainSet
 Array which contains the number of units in each layer. More...
 
TNeuralDataSet * m_pOutputTrainSet
 
TNeuralDataSet * m_pInputTestSet
 
TNeuralDataSet * m_pOutputTestSet
 
Int_t m_TrainSetCnt
 
Int_t m_TestSetCnt
 
Int_t m_InputDim
 
Int_t m_HiddenLayerDim
 
Int_t m_OutputDim
 
Int_t m_Epochs
 
Int_t m_CurrentEpoch
 
Bool_t m_Debug
 
Bool_t m_IsInitialized
 
Bool_t m_InitLocked
 
Bool_t m_NormalizeOutput
 

Detailed Description

Definition at line 40 of file TJetNet.h.

Member Enumeration Documentation

◆ TActivationFunction

enum TJetNet::TActivationFunction
Enumerator
afSigmoid 
afTanh 
afExp 
afLinear 
afSigmoidEntropy 

Definition at line 150 of file TJetNet.h.

150  {
151  afSigmoid = 1,
152  afTanh = 2,
153  afExp = 3,
154  afLinear = 4,
155  afSigmoidEntropy = 5
156  };

Constructor & Destructor Documentation

◆ TJetNet() [1/2]

TJetNet::TJetNet ( void  )

◆ TJetNet() [2/2]

TJetNet::TJetNet ( Int_t  aTestCount,
Int_t  aTrainCount,
const Int_t  aLayersCnt,
const Int_t *  aLayers 
)

Definition at line 38 of file TJetNet.cxx.

41  : m_Debug (kTRUE)
42 #else
43  : m_Debug (kFALSE)
44 #endif
45 {
46  // Creates neural network with aLayersCnt number of layers,
47  // aTestCount number of patterns for the test set,
48  // aTrainCount patterns for the train set.
49  // aLayers contains the information for number of the units in the different layers
50 
51  int i;
52 
53  if( m_Debug ){ std::cout << "=====> Entering TJetNet::TJetNet(...)" << std::endl; }
54 
55  m_TestSetCnt = aTestCount;
56  m_TrainSetCnt = aTrainCount;
57  m_LayerCount = aLayersCnt; // Get the number of layers
58 
59  if( m_LayerCount > 0 )
60  {
61  //Perform deep copy of the array holding Layers count
62  m_pLayers = new Int_t[ m_LayerCount ];
63  for( i = 0; i < m_LayerCount; ++i )
64  {
65  m_pLayers[ i ] = aLayers[ i ];
66  }
67  }
68 
69  m_InputDim = m_pLayers[ 0 ];
70  m_OutputDim = m_pLayers[ m_LayerCount - 1 ];
71  m_HiddenLayerDim = m_LayerCount-2;
72 
73 
74  m_IsInitialized = kFALSE;
75  m_InitLocked = kFALSE;
76 
77  m_pInputTrainSet = new TNeuralDataSet( m_TrainSetCnt, GetInputDim() );
78  m_pInputTestSet = new TNeuralDataSet( m_TestSetCnt, GetInputDim() );
79  m_pOutputTrainSet = new TNeuralDataSet( m_TrainSetCnt, GetOutputDim() );
80  m_pOutputTestSet = new TNeuralDataSet( m_TestSetCnt, GetOutputDim() );
81 
82  m_NormalizeOutput=false;
83 
84  m_enActFunction=afSigmoid;
85 
86  SetEpochs( -1 );
87 
88  if( m_Debug ){ std::cout << "=====> Leaving TJetNet::TJetNet(...)" << std::endl; }
89 }

◆ ~TJetNet()

TJetNet::~TJetNet ( void  )
virtual

Definition at line 91 of file TJetNet.cxx.

92 {
93  // Default destructor
94  if( m_Debug ){ std::cout << "=====> Entering TJetNet::~TJetNet(...)" << std::endl; }
95  delete [] m_pLayers;
96  delete m_pInputTestSet;
97  delete m_pInputTrainSet;
98  delete m_pOutputTestSet;
99  delete m_pOutputTrainSet;
100  if( m_Debug ){ std::cout << "=====> Leaving TJetNet::~TJetNet(...)" << std::endl; }
101 }

Member Function Documentation

◆ CopyFile()

Int_t TJetNet::CopyFile ( TString  aSrcFile,
TString  aDestFile 
)
private

◆ createTrainedNetwork()

TTrainedNetwork * TJetNet::createTrainedNetwork ( ) const

Definition at line 104 of file TJetNet.cxx.

105 {
106 
107  Int_t nInput=GetInputDim();
108  Int_t nHidden=GetHiddenLayerDim();
109  std::vector<Int_t> nHiddenLayerSize;
110  // Int_t* nHiddenLayerSize=new Int_t[nHidden];
111 
112  for (Int_t o=0;o<nHidden;++o)
113  {
114  nHiddenLayerSize.push_back(GetHiddenLayerSize(o+1));
115  }
116  Int_t nOutput=GetOutputDim();
117 
118  std::vector<TVectorD*> thresholdVectors;
119  std::vector<TMatrixD*> weightMatrices;
120 
121  for (Int_t o=0;o<nHidden+1;++o)
122  {
123  int sizeActualLayer=(o<nHidden)?nHiddenLayerSize[o]:nOutput;
124  int sizePreviousLayer=(o==0)?nInput:nHiddenLayerSize[o-1];
125  thresholdVectors.push_back(new TVectorD(sizeActualLayer));
126  weightMatrices.push_back(new TMatrixD(sizePreviousLayer,sizeActualLayer));
127  }
128 
129  for (Int_t o=0;o<nHidden+1;++o)
130  {
131 
132  if (m_Debug)
133  if (o<nHidden)
134  {
135  cout << " Iterating on hidden layer n.: " << o << endl;
136  }
137  else
138  {
139  cout << " Considering output layer " << endl;
140  }
141 
142  int sizeActualLayer=(o<nHidden)?nHiddenLayerSize[o]:nOutput;
143 
144  for (Int_t s=0;s<sizeActualLayer;++s)
145  {
146  if (o<nHidden)
147  {
148  if (m_Debug)
149  cout << " To hidden node: " << s << endl;
150  }
151  else
152  {
153  if (m_Debug)
154  cout << " To output node: " << s << endl;
155  }
156  if (o==0)
157  {
158  for (Int_t p=0;p<nInput;++p)
159  {
160  if (m_Debug)
161  cout << " W from inp nod: " << p << "weight: " <<
162  GetWeight(o+1,s+1,p+1) << endl;
163  weightMatrices[o]->operator() (p,s) = GetWeight(o+1,s+1,p+1);
164  }
165  }
166  else
167  {
168  for (Int_t p=0;p<nHiddenLayerSize[o-1];++p)
169  {
170  if (m_Debug)
171  cout << " W from lay : " << o-1 << " nd: " <<
172  p << " weight: " <<
173  GetWeight(o+1,s+1,p+1) << endl;
174  weightMatrices[o]->operator() (p,s)=GetWeight(o+1,s+1,p+1);
175  }
176  }
177  if (m_Debug)
178  cout << " Threshold for node " << s << " : " <<
179  GetThreshold(o+1,s+1) << endl;
180  thresholdVectors[o]->operator() (s) = GetThreshold(o+1,s+1);
181  }
182  }
183 
184  bool linearOutput=false;
185  if (this->GetActivationFunction(nHidden)==4)
186  {
187  // cout << " Creating TTrainedNetwork with linear output function" << endl;
188  linearOutput=true;
189  }
190 
191  TTrainedNetwork* trainedNetwork=
192  new TTrainedNetwork(nInput,
193  nHidden,
194  nOutput,
195  nHiddenLayerSize,
196  thresholdVectors,
197  weightMatrices,
198  m_enActFunction,
199  linearOutput,
200  m_NormalizeOutput);
201 
202  return trainedNetwork;
203 
204 }

◆ DumpToFile()

void TJetNet::DumpToFile ( TString  aFileName = "fort.8")

Definition at line 954 of file TJetNet.cxx.

955 {
956  // Dumps the network data into JETNET specific format
957  JNDUMP( -8 );
958  std::cout << close( 8 ) << std::endl;
959  rename( "./fort.8", aFileName );
960 }

◆ Epoch()

Int_t TJetNet::Epoch ( void  )

Definition at line 696 of file TJetNet.cxx.

697 {
698  // Initiate one train/test step the network.
699 
700  Double_t aTrain, aTest;
701  if ( m_CurrentEpoch < m_Epochs )
702  {
703  m_CurrentEpoch++;
704  aTrain = Train();
705 
706  // if (m_CurrentEpoch%2)
707 
708  // std::cout << " Calls to MSTJN: " << GetMSTJN(6) <<
709  // std::endl;
710 
711  if ( m_Debug )
712  {
713 
714 
715  std::cout << "[ " << m_CurrentEpoch << " ] Train: " << aTrain << std::endl;
716  }
717  if ( ( m_CurrentEpoch % 2 ) == 0 )
718  {
719  aTest = Test();
720  // if ( m_Debug )
721  std::cout << "[" << m_CurrentEpoch << "]: " << GetPARJN(8) << " ";
722  std::cout << "Test: " << aTest << std::endl;
723  }
724  }
725  return m_CurrentEpoch;
726 }

◆ Evaluate() [1/2]

void TJetNet::Evaluate ( )

Definition at line 926 of file TJetNet.cxx.

927 {
928  //evaluates directly the input provided through SetInputs()
929  JNTEST();
930 }

◆ Evaluate() [2/2]

void TJetNet::Evaluate ( Int_t  aPattern)

Definition at line 932 of file TJetNet.cxx.

933 {
934  // Evaluates the network output form the input data specified by the Test Pattern
935  for( Int_t i = 0; i < GetInputDim(); i++ )
936  {
937  JNDAT1.OIN[ i ] = float ( GetInputTestSet( aPattern, i ) );
938  }
939  JNTEST();
940 }

◆ GetActivationFunction() [1/2]

Int_t TJetNet::GetActivationFunction ( Int_t  layerN) const

Definition at line 1167 of file TJetNet.cxx.

1168 {
1169  return JNDAT2.IGFN[ layerN ];
1170 }

◆ GetActivationFunction() [2/2]

Int_t TJetNet::GetActivationFunction ( void  ) const

Definition at line 1161 of file TJetNet.cxx.

1162 {
1163  return JNDAT1.MSTJN[ 2 ];
1164 }

◆ GetEpochs()

Int_t TJetNet::GetEpochs ( void  )
inline

Definition at line 78 of file TJetNet.h.

78 { return m_Epochs; };

◆ GetErrorMeasure()

Int_t TJetNet::GetErrorMeasure ( void  )

Definition at line 1156 of file TJetNet.cxx.

1157 {
1158  return JNDAT1.MSTJN[ 3 ];
1159 }

◆ GetEventWeightTestSet()

Double_t TJetNet::GetEventWeightTestSet ( Int_t  aPatternInd)
inline

Definition at line 213 of file TJetNet.h.

214 {
215  return m_pInputTestSet->GetEventWeight( aPatternInd);
216 }

◆ GetEventWeightTrainSet()

Double_t TJetNet::GetEventWeightTrainSet ( Int_t  aPatternInd)
inline

Definition at line 208 of file TJetNet.h.

209 {
210  return m_pInputTrainSet->GetEventWeight( aPatternInd);
211 }

◆ GetHiddenLayerDim()

Int_t TJetNet::GetHiddenLayerDim ( void  ) const
inline

Definition at line 55 of file TJetNet.h.

55 { return mHiddenLayerDim; };

◆ GetHiddenLayerSize()

Int_t TJetNet::GetHiddenLayerSize ( Int_t  number) const
inline

Definition at line 56 of file TJetNet.h.

56 { return m_pLayers[ number ]; };

◆ GetInitialWeightsWidth()

Double_t TJetNet::GetInitialWeightsWidth ( void  )

Definition at line 1187 of file TJetNet.cxx.

1188 {
1189  return JNDAT1.PARJN[ 3 ];
1190 }

◆ GetInputDim()

Int_t TJetNet::GetInputDim ( void  ) const
inline

Definition at line 54 of file TJetNet.h.

54 { return m_pLayers[ 0 ]; };

◆ GetInputTestSet()

Double_t TJetNet::GetInputTestSet ( Int_t  aPatternInd,
Int_t  aInputInd 
)

Definition at line 763 of file TJetNet.cxx.

764 {
765  // Returns the value of the cell corresponding to unit aInputInd in pattern aPatternInd into INPUT TEST set
766  return m_pInputTestSet->GetData( aPatternInd, aInputInd );
767 }

◆ GetInputTrainSet()

Double_t TJetNet::GetInputTrainSet ( Int_t  aPatternInd,
Int_t  aInputInd 
)
inline

Definition at line 202 of file TJetNet.h.

203 {
204  // Returns the value of the cell corresponding to unit aInputInd in pattern aPatternInd into INPUT TRAIN set
205  return m_pInputTrainSet->GetData( aPatternInd, aInputInd );
206 }

◆ GetLayerCount()

Int_t TJetNet::GetLayerCount ( void  )
inline

Definition at line 105 of file TJetNet.h.

105 { return m_LayerCount; };

◆ GetLearningRate()

Double_t TJetNet::GetLearningRate ( void  )

Definition at line 1177 of file TJetNet.cxx.

1178 {
1179  return JNDAT1.PARJN[ 0 ];
1180 }

◆ GetLearningRateDecrease()

Double_t TJetNet::GetLearningRateDecrease ( void  )

Definition at line 1192 of file TJetNet.cxx.

1193 {
1194  return JNDAT1.PARJN[ 10 ];
1195 }

◆ GetMomentum()

Double_t TJetNet::GetMomentum ( void  )

Definition at line 1182 of file TJetNet.cxx.

1183 {
1184  return JNDAT1.PARJN[ 1 ];
1185 }

◆ GetMSTJN()

Int_t TJetNet::GetMSTJN ( Int_t  aIndex)

Definition at line 1202 of file TJetNet.cxx.

1203 {
1204  return JNDAT1.MSTJN[ aIndex ];
1205 }

◆ GetOutput()

Double_t TJetNet::GetOutput ( Int_t  aIndex = 0)

Definition at line 948 of file TJetNet.cxx.

949 {
950  // Returns the output of the network
951  return Double_t ( JNDAT1.OUT[ aIndex ] );
952 }

◆ GetOutputDim()

Int_t TJetNet::GetOutputDim ( void  ) const
inline

Definition at line 57 of file TJetNet.h.

57 { return m_pLayers[ m_LayerCount - 1 ]; };

◆ GetOutputTestSet()

Double_t TJetNet::GetOutputTestSet ( Int_t  aPatternInd,
Int_t  aOutputInd 
)

Definition at line 769 of file TJetNet.cxx.

770 {
771  // Returns the value of the cell corresponding to unit aInputInd in pattern aPatternInd into OUTPUT TEST set
772  return m_pOutputTestSet->GetData( aPatternInd, aOutputInd );
773 }

◆ GetOutputTrainSet()

Double_t TJetNet::GetOutputTrainSet ( Int_t  aPatternInd,
Int_t  aOutputInd 
)

Definition at line 746 of file TJetNet.cxx.

747 {
748  // Returns the value of the cell corresponding to unit aInputInd in pattern aPatternInd into OUTPUT TRAIN set
749  return m_pOutputTrainSet->GetData( aPatternInd, aOutputInd );
750 }

◆ GetPARJN()

Double_t TJetNet::GetPARJN ( Int_t  aIndex)

Definition at line 1207 of file TJetNet.cxx.

1208 {
1209  return JNDAT1.PARJN[ aIndex ];
1210 }

◆ GetPatternsPerUpdate()

Int_t TJetNet::GetPatternsPerUpdate ( void  )

Definition at line 1172 of file TJetNet.cxx.

1173 {
1174  return JNDAT1.MSTJN[ 1 ];
1175 }

◆ GetPottsUnits()

Int_t TJetNet::GetPottsUnits ( )

Definition at line 1197 of file TJetNet.cxx.

1198 {
1199  return JNINT2.IPOTT;
1200 }

◆ GetTestSetCnt()

Int_t TJetNet::GetTestSetCnt ( void  ) const
inline

Definition at line 53 of file TJetNet.h.

53 { return m_TestSetCnt; };

◆ GetThreshold()

Double_t TJetNet::GetThreshold ( Int_t  aLayerInd,
Int_t  aNodeInd 
) const

Definition at line 979 of file TJetNet.cxx.

980 {
981  //Returns the node threshold in the specific layer
982  return Double_t ( JNINT1.T[ JNINDX( aLayerInd, aNodeInd, 0 )-1 ] );
983  //GP: ONE HAS TO PAY ATTENTION TO THIS STUPID -1!!!
984 }

◆ GetTrainSetCnt()

Int_t TJetNet::GetTrainSetCnt ( void  ) const
inline

Definition at line 52 of file TJetNet.h.

52 { return m_TrainSetCnt; };

◆ GetUnitCount()

Int_t TJetNet::GetUnitCount ( Int_t  aLayer)

Definition at line 1064 of file TJetNet.cxx.

1065 {
1066  // Returns the number of the units in specfic layer
1067  if( ( aLayer > -1 ) && ( aLayer < m_LayerCount ) )
1068  return JNDAT1.MSTJN[ 9 + aLayer ];
1069 }

◆ GetUpdatesPerEpoch()

Int_t TJetNet::GetUpdatesPerEpoch ( void  )

Definition at line 1146 of file TJetNet.cxx.

1147 {
1148  return JNDAT1.MSTJN[ 8 ];
1149 }

◆ GetUpdatingProcedure()

Int_t TJetNet::GetUpdatingProcedure ( void  )

Definition at line 1151 of file TJetNet.cxx.

1152 {
1153  return JNDAT1.MSTJN[ 3 ];
1154 }

◆ GetWeight()

Double_t TJetNet::GetWeight ( Int_t  aLayerInd,
Int_t  aNodeInd,
Int_t  aConnectedNodeInd 
) const

Definition at line 972 of file TJetNet.cxx.

973 {
974  // Returns the node weight in specific Layer
975  return Double_t ( JNINT1.W[ JNINDX( aLayerInd, aNodeInd, aConnectedNodeInd )-1 ] );
976  //GP: ONE HAS TO PAY ATTENTION TO THIS STUPID -1!!!
977 }

◆ Init()

void TJetNet::Init ( void  )

Definition at line 670 of file TJetNet.cxx.

671 {
672  // Initializes the neuaral network
673  Int_t i;
674  JNDAT1.MSTJN[ 0 ] = m_LayerCount; // Set the number of layers
675 
676  // Set the number of nodes for each layer
677  for( i = 0; i < m_LayerCount; i++ )
678  {
679  if ( m_Debug ) std::cout << "Layer " << i + 1 << " has " << m_pLayers[ i ] << " units." << std::endl;
680  JNDAT1.MSTJN[ 9 + i ] = m_pLayers[ i ];
681  }
682 
683  cout << " calling JNINIT " << endl;
684  JNINIT();
685 
686  if (m_NormalizeOutput)
687  {
688  std::cout << " Setting to normalize output nodes: POTT nodes " << std::endl;
689  SetPottsUnits(m_pLayers[m_LayerCount-1]);
690  }
691 
692  cout << " finishing calling JNINIT " << endl;
693  m_IsInitialized = kTRUE;
694 }

◆ LoadDataAscii()

void TJetNet::LoadDataAscii ( TString  aFileName = "jndata.dat")

Definition at line 815 of file TJetNet.cxx.

816 {
817  // Loads the input/output test/train data from plain text file
818  ifstream in;
819  int i, j, k, l, m;
820  int aiParam[ 5 ];//iTrainCount, iTestCount, iInputDim, iHiddenDim, iOutputDim;
821  Bool_t bFlag;
822  Double_t tmp;
823  Int_t iPatternLength;
824 
825  in.open( aFileName );
826  bFlag = Bool_t( in.is_open() );
827  if ( in )
828  {
829  in >> m_LayerCount;
830  if( m_Debug ){ std::cout << "Layers Count Set to " << m_LayerCount << std::endl;}
831  i = 0;
832 
833  delete [] m_pLayers;
834  m_pLayers = new Int_t[ m_LayerCount ];
835 
836  if( m_Debug ){ std::cout << "Updating the Layers Nodes Counters..." << std::endl; }
837  while( ( i < m_LayerCount ) && ( !in.eof() ) )
838  {
839  in >> m_pLayers[ i ];
840  if( m_Debug ){ std::cout << "Layer [ " << i + 1 << " ] has " << m_pLayers[ i ] << " units" << std::endl; }
841  i++;
842  }
843 
844  m_InputDim = m_pLayers[ 0 ];
845  m_OutputDim = m_pLayers[ m_LayerCount - 1 ];
846  m_HiddenLayerDim = m_LayerCount-2;
847 
848  //Get the patterns count per line
849  iPatternLength = m_InputDim + m_OutputDim;
850  if( m_Debug ){ std::cout << "Patterns per line = " << iPatternLength << std::endl; }
851  in >> m_TrainSetCnt;
852  if( m_Debug ){ std::cout << "Train Set has " << m_TrainSetCnt << " patterns." << std::endl; }
853  in >> m_TestSetCnt;
854  if( m_Debug ){ std::cout << "Test Set has " << m_TestSetCnt << " patterns." << std::endl; }
855 
856  delete m_pInputTestSet;
857  delete m_pInputTrainSet;
858  delete m_pOutputTestSet;
859  delete m_pOutputTrainSet;
860 
861  m_pInputTrainSet = new TNeuralDataSet( m_TrainSetCnt, GetInputDim() );
862  m_pInputTestSet = new TNeuralDataSet( m_TestSetCnt, GetInputDim() );
863  m_pOutputTrainSet = new TNeuralDataSet( m_TrainSetCnt, GetOutputDim() );
864  m_pOutputTestSet = new TNeuralDataSet( m_TestSetCnt, GetOutputDim() );
865 
866  i = 0;
867  j = 0;
868 
869  while( ( i < ( m_TrainSetCnt + m_TestSetCnt ) ) && ( !in.eof() ) )
870  {
871  j = 0;
872  while( ( j < iPatternLength ) && ( !in.eof() ) )
873  {
874  if( i < m_TrainSetCnt )
875  {
876  if( j < m_InputDim )
877  {
878  //Train Input Set
879  in >> tmp;
880  SetInputTrainSet( i, j, tmp );
881  }
882  else
883  {
884  //Train Output Set
885  m = j - m_InputDim;
886  in >> tmp;
887  SetOutputTrainSet( i, m, tmp );
888  }
889  }
890  else
891  {
892  l = i - m_TrainSetCnt;
893  if( j < m_InputDim )
894  {
895  //Test Input Set
896  in >> tmp;
897  SetInputTestSet( l, j, tmp );
898  }
899  else
900  {
901  //Test Output Set
902  m = j - m_InputDim;
903  in >> tmp;
904  SetOutputTestSet( l, m, tmp );
905  }
906 
907  }
908  j++;
909  }
910  i++;
911  }
912  }
913  in.close();
914 }

◆ LoadDataRoot()

void TJetNet::LoadDataRoot ( TString  aFileName = "jndata.root")

Definition at line 921 of file TJetNet.cxx.

922 {
923  // Loads the neural network from ROOT file
924 }

◆ LockInit()

void TJetNet::LockInit ( void  )
inline

Definition at line 137 of file TJetNet.h.

137 { m_InitLocked = kTRUE; };

◆ mSetThreshold()

void TJetNet::mSetThreshold ( Double_t  threshold,
Int_t  aLayerInd,
Int_t  aNodeInd 
)
private

Definition at line 290 of file TJetNet.cxx.

291 {
292  JNINT1.T[ JNINDX( aLayerInd, aNodeInd, 0 )-1 ]=threshold;
293 }

◆ mSetWeight()

void TJetNet::mSetWeight ( Double_t  weight,
Int_t  aLayerInd,
Int_t  aNodeInd,
Int_t  aConnectedNodeInd 
)
private

Definition at line 285 of file TJetNet.cxx.

286 {
287  JNINT1.W[ JNINDX( aLayerInd, aNodeInd, aConnectedNodeInd )-1 ]=weight;
288 }

◆ Normalize()

void TJetNet::Normalize ( void  )

Definition at line 1048 of file TJetNet.cxx.

1049 {
1050  // Normilizes Inputs (both test and train)
1051  m_pInputTrainSet->Normalize();
1052  m_pInputTestSet->Normalize();
1053 }

◆ NormalizeOutput()

void TJetNet::NormalizeOutput ( bool  isTrue = true)

Definition at line 1141 of file TJetNet.cxx.

1142 {
1143  m_NormalizeOutput=isTrue;
1144 }

◆ Print()

void TJetNet::Print ( void  )

Definition at line 295 of file TJetNet.cxx.

296 {
297  // Prints on the screen, information for the neural network
298  Int_t i;
299 
300  std::cout << "TJetNet" << std::endl;
301  std::cout << "Number of layers: " << m_LayerCount << std::endl;
302 
303  for( i = 0; i < m_LayerCount; i++ )
304  {
305  std::cout << "\t\tNumber of units in layer " << i << " : " << m_pLayers[ i ] << std::endl;
306  }
307 
308  std::cout << "Epochs: " << GetEpochs() << std::endl;
309  std::cout << "Updates Per Epoch: " << GetUpdatesPerEpoch() << std::endl;
310  std::cout << "Updating Procedure: " << GetUpdatingProcedure() << std::endl;
311  std::cout << "Error Measure: " << GetErrorMeasure() << std::endl;
312  std::cout << "Patterns Per Update: " << GetPatternsPerUpdate() << std::endl;
313  std::cout << "Learning Rate: " << GetLearningRate() << std::endl;
314  std::cout << "Momentum: " << GetMomentum() << std::endl;
315  std::cout << "Initial Weights Width: " << GetInitialWeightsWidth() << std::endl;
316  std::cout << "Learning Rate Decrease: " << GetLearningRateDecrease() << std::endl;
317  std::cout << "Activation Function: " << GetActivationFunction() << std::endl;
318 }

◆ Randomize()

void TJetNet::Randomize ( void  )

Definition at line 1055 of file TJetNet.cxx.

1056 {
1057  // Randomizes Inputs and Outputs of both train and test sets
1058  m_pInputTrainSet->Randomize();
1059  m_pInputTestSet->Randomize();
1060  m_pOutputTrainSet->Randomize();
1061  m_pOutputTestSet->Randomize();
1062 }

◆ readBackTrainedNetwork()

void TJetNet::readBackTrainedNetwork ( const TTrainedNetwork *  trainedNetwork)

Definition at line 207 of file TJetNet.cxx.

208 {
209 
210  Int_t nInput=GetInputDim();
211  Int_t nHidden=GetHiddenLayerDim();
212  std::vector<Int_t> nHiddenLayerSize;
213 
214  if (trainedNetwork->getnHidden()!=nHidden)
215  {
216  cout << " Network doesn't match.. not loading.." << endl;
217  return;
218  }
219 
220  for (Int_t o=0;o<nHidden;++o)
221  {
222  nHiddenLayerSize.push_back(GetHiddenLayerSize(o+1));
223  if (nHiddenLayerSize[o]!=trainedNetwork->getnHiddenLayerSize()[o])
224  {
225  cout << " Network doesn't match... not loading..." << endl;
226  return;
227  }
228  }
229  Int_t nOutput=GetOutputDim();
230 
231  if (trainedNetwork->getnInput()!=nInput)
232  {
233  cout << " Network doesn't match... not loading.." << endl;
234  return;
235  }
236 
237 
238  if (trainedNetwork->getnOutput()!=nOutput)
239  {
240  cout << " Network doesn't match.. not loading.." << endl;
241  return;
242  }
243 
244  //OK, everything matches... can go on...
245 
246  std::vector<TVectorD*> thresholdVectors=trainedNetwork->getThresholdVectors();
247  std::vector<TMatrixD*> weightMatrices=trainedNetwork->weightMatrices();
248  //ownership remains of the TTrainedNetwork
249 
250  for (Int_t o=0;o<nHidden+1;++o)
251  {
252  int sizeActualLayer=(o<nHidden)?nHiddenLayerSize[o]:nOutput;
253  int sizePreviousLayer=(o==0)?nInput:nHiddenLayerSize[o-1];
254 
255  for (Int_t s=0;s<sizeActualLayer;++s)
256  {
257  Double_t nodeValue=0.;
258  if (o==0)
259  {
260  for (Int_t p=0;p<nInput;++p)
261  {
262  mSetWeight(weightMatrices[o]->operator() (p,s),o+1,s+1,p+1);
263  }
264  }
265  else
266  {
267  for (Int_t p=0;p<nHiddenLayerSize[o-1];++p)
268  {
269  mSetWeight(weightMatrices[o]->operator() (p,s),o+1,s+1,p+1);
270  }
271  }
272  mSetThreshold(thresholdVectors[o]->operator() (s),o+1,s+1);
273  }
274  }
275  if (trainedNetwork->getIfLinearOutput()==true)
276  {
277  cout << " Setting linear output function " << endl;
278  this->SetActivationFunction(nHidden,4);
279  }
280 
281  cout << " Successfully read back Trained Network " << endl;
282 }

◆ ReadFromFile()

void TJetNet::ReadFromFile ( TString  aFileName = "fort.8")

Definition at line 962 of file TJetNet.cxx.

963 {
964  // Loads the network from JETNET specific file
965  rename( aFileName, "./fort.12" );
966  JNREAD( -12 );
967  Reinitialize();
968  rename( "./fort.12", aFileName );
969  //std::cout << close( 12 ) << std::endl;
970 }

◆ Reinitialize()

void TJetNet::Reinitialize ( void  )
private

Definition at line 1020 of file TJetNet.cxx.

1021 {
1022  //Initializes the settings of the network
1023  Int_t i;
1024 
1025  m_LayerCount = JNDAT1.MSTJN[ 0 ]; // Set the number of layers
1026 
1027  delete [] m_pLayers;
1028  m_pLayers = new Int_t[ m_LayerCount ];
1029 
1030  // Set the number of nodes for each layer
1031  for( i = 0; i < m_LayerCount; i++ )
1032  {
1033  m_pLayers[ i ] = JNDAT1.MSTJN[ 9 + i ];
1034  }
1035 
1036  m_pInputTrainSet = new TNeuralDataSet( m_TrainSetCnt, GetInputDim() );
1037  m_pInputTestSet = new TNeuralDataSet( m_TestSetCnt, GetInputDim() );
1038  m_pOutputTrainSet = new TNeuralDataSet( m_TrainSetCnt, GetOutputDim() );
1039  m_pOutputTestSet = new TNeuralDataSet( m_TestSetCnt, GetOutputDim() );
1040 
1041  m_InputDim = m_pLayers[ 0 ];
1042  m_OutputDim = m_pLayers[ m_LayerCount - 1 ];
1043  m_HiddenLayerDim = m_LayerCount-2;
1044 
1045 
1046 }

◆ SaveDataAscii()

void TJetNet::SaveDataAscii ( TString  aFileName = "jndata.dat")

Definition at line 775 of file TJetNet.cxx.

776 {
777  // Saves the Input/Output test and train data in plain text file
778  ofstream out;
779  int i, j;
780 
781  // Open ASCII file
782  out.open( aFileName );
783 
784  //Write the number of layers, including the input and output
785  out << m_LayerCount << std::endl;
786 
787  // Write into the file the number of units in input, hidden and output layers
788  for ( i = 0; i < m_LayerCount; i++ ) out << m_pLayers[ i ] << " ";
789  out << std::endl;
790 
791  // Write the size of Train and Test sets
792  out << m_TrainSetCnt << " " << m_TestSetCnt << std::endl;
793 
794  // Dump the Train set : Input1 Input2 ... InputN Output1 Output2 ... OutputN
795  for ( i = 0; i < m_TrainSetCnt; i++ )
796  {
797  out << GetInputTrainSet( i, 0 );
798  for( j = 1; j < m_pLayers[ 0 ]; j++ ) out << " " << GetInputTrainSet( i, j );
799  for( j = 0; j < m_pLayers[ m_LayerCount - 1 ]; j++ ) out << " " << GetOutputTrainSet( i, j );
800  out << std::endl;
801  }
802 
803  // Dump the Test set : Input1 Input2 ... InputN Output1 Output2 ... OutputN
804  for ( i = 0; i < m_TestSetCnt; i++ )
805  {
806  out << GetInputTestSet( i, 0 );
807  for( j = 1; j < m_pLayers[ 0 ]; j++ ) out << " " << GetInputTestSet( i, j );
808  for( j = 0; j < m_pLayers[ m_LayerCount - 1 ]; j++ ) out << " " << GetOutputTestSet( i, j );
809  out << std::endl;
810  }
811  // Close the file
812  out.close();
813 }

◆ SaveDataRoot()

void TJetNet::SaveDataRoot ( TString  aFileName = "jndata.root")

Definition at line 916 of file TJetNet.cxx.

917 {
918  // Saves the neural network in ROOT file
919 }

◆ SelectiveFields()

void TJetNet::SelectiveFields ( Int_t  aLayerA,
Int_t  aNodeA1,
Int_t  aNodeA2,
Int_t  aNodeB1,
Int_t  aNodeB2,
Int_t  aSwitch = 0 
)

Definition at line 986 of file TJetNet.cxx.

987 {
988  // JetNet Selective Fields
989  Int_t tmp, i1, i2, j1, j2;
990 
991  if( ( aLayerA > 0 ) && ( aLayerA < m_LayerCount ) )
992  {
993  i1 = TMath::Abs( aNodeA1 );
994  i2 = TMath::Abs( aNodeA2 );
995  j1 = TMath::Abs( aNodeB1 );
996  j2 = TMath::Abs( aNodeB2 );
997 
998  if( i1 > i2 )
999  {
1000  tmp = i1;
1001  i1 = i2;
1002  i2 = i1;
1003  }//if
1004 
1005  if( i1 > i2 )
1006  {
1007  tmp = i1;
1008  i1 = i2;
1009  i2 = i1;
1010  }//if
1011 
1012  if( ( i1 < m_pLayers[ aLayerA ] ) && ( i2 < m_pLayers[ aLayerA ] ) &&
1013  ( j1 < m_pLayers[ aLayerA - 1 ] ) && ( j2 < m_pLayers[ aLayerA - 1 ] ) )
1014  {
1015  JNSEFI( aLayerA, i1, i2, j1, j2, aSwitch );
1016  }//if
1017  } //if
1018 }

◆ SetActivationFunction() [1/2]

void TJetNet::SetActivationFunction ( Int_t  aValue)

Definition at line 1091 of file TJetNet.cxx.

1092 {
1093  // Set the kind of activation function used
1094  JNDAT1.MSTJN[ 2 ] = aValue;
1095  if( !m_InitLocked ) this->Init();
1096 }

◆ SetActivationFunction() [2/2]

void TJetNet::SetActivationFunction ( Int_t  aValue,
Int_t  layerN 
)

Definition at line 1098 of file TJetNet.cxx.

1099 {
1100  // Set the kind of activation function used
1101  JNDAT2.IGFN[ layerN ] = aValue;
1102  if( !m_InitLocked ) this->Init();
1103 }

◆ SetEpochs()

void TJetNet::SetEpochs ( const Int_t  aEpochs)
inline

Definition at line 79 of file TJetNet.h.

79 { m_Epochs = aEpochs; m_CurrentEpoch = 0; };

◆ SetErrorMeasure()

void TJetNet::SetErrorMeasure ( Int_t  aValue)

Definition at line 1085 of file TJetNet.cxx.

1086 {
1087  JNDAT1.MSTJN[ 3 ] = aValue;
1088  if( !m_InitLocked ) this->Init();
1089 }

◆ SetEventWeightTestSet()

void TJetNet::SetEventWeightTestSet ( Int_t  aPatternInd,
Double_t  aValue 
)

Definition at line 758 of file TJetNet.cxx.

759 {
760  m_pInputTestSet->SetEventWeight(aPatternInd,aValue);
761 }

◆ SetEventWeightTrainSet()

void TJetNet::SetEventWeightTrainSet ( Int_t  aPatternInd,
Double_t  aValue 
)

Definition at line 752 of file TJetNet.cxx.

753 {
754  m_pInputTrainSet->SetEventWeight(aPatternInd,aValue);
755 }

◆ SetInitialWeightsWidth()

void TJetNet::SetInitialWeightsWidth ( Double_t  aValue)

Definition at line 1124 of file TJetNet.cxx.

1125 {
1126  JNDAT1.PARJN[ 3 ] = aValue;
1127  if( !m_InitLocked ) this->Init();
1128 }

◆ SetInputs()

void TJetNet::SetInputs ( Int_t  aIndex = 0,
Double_t  aValue = 0.0 
)

Definition at line 942 of file TJetNet.cxx.

943 {
944  // Directly sets the inputs of the network
945  JNDAT1.OIN[ aIndex ] = float ( aValue );
946 }

◆ SetInputTestSet()

void TJetNet::SetInputTestSet ( Int_t  aPatternInd,
Int_t  aInputInd,
Double_t  aValue 
)

Definition at line 740 of file TJetNet.cxx.

741 {
742  // Changes the value of the cell corresponding to unit aInputInd in pattern aPatternInd into INPUT TEST set
743  m_pInputTestSet->SetData( aPatternInd, aInputInd, aValue );
744 }

◆ SetInputTrainSet()

void TJetNet::SetInputTrainSet ( Int_t  aPatternInd,
Int_t  aInputInd,
Double_t  aValue 
)

Definition at line 728 of file TJetNet.cxx.

729 {
730  // Changes the value of the cell corresponding to unit aInputInd in pattern aPatternInd into INPUT TRAIN set
731  m_pInputTrainSet->SetData( aPatternInd, aInputInd, aValue );
732 }

◆ SetLearningRate()

void TJetNet::SetLearningRate ( Double_t  aValue)

Definition at line 1111 of file TJetNet.cxx.

1112 {
1113  // Change the Learning Rate
1114  JNDAT1.PARJN[ 0 ] = aValue;
1115  if( !m_InitLocked ) this->Init();
1116 }

◆ SetLearningRateDecrease()

void TJetNet::SetLearningRateDecrease ( Double_t  aValue)

Definition at line 1130 of file TJetNet.cxx.

1131 {
1132  JNDAT1.PARJN[ 10 ] = aValue;
1133  if( !m_InitLocked ) this->Init();
1134 }

◆ SetMomentum()

void TJetNet::SetMomentum ( Double_t  aValue)

Definition at line 1118 of file TJetNet.cxx.

1119 {
1120  JNDAT1.PARJN[ 1 ] = aValue;
1121  if( !m_InitLocked ) this->Init();
1122 }

◆ SetMSTJN()

void TJetNet::SetMSTJN ( Int_t  aIndex,
Int_t  aValue 
)

Definition at line 1212 of file TJetNet.cxx.

1213 {
1214  JNDAT1.MSTJN[ aIndex ] = aValue;
1215 }

◆ SetOutputTestSet()

void TJetNet::SetOutputTestSet ( Int_t  aPatternInd,
Int_t  aOutputInd,
Double_t  aValue 
)
inline

Definition at line 196 of file TJetNet.h.

197 {
198  // Changes the value of the cell corresponding to unit aInputInd in pattern aPatternInd into OUTPUT TEST set
199  m_pOutputTestSet->SetData( aPatternInd, aOutputInd, aValue );
200 }

◆ SetOutputTrainSet()

void TJetNet::SetOutputTrainSet ( Int_t  aPatternInd,
Int_t  aOutputInd,
Double_t  aValue 
)

Definition at line 734 of file TJetNet.cxx.

735 {
736  // Changes the value of the cell corresponding to unit aInputInd in pattern aPatternInd into OUTPUT TRAIN set
737  m_pOutputTrainSet->SetData( aPatternInd, aOutputInd, aValue );
738 }

◆ SetPARJN()

void TJetNet::SetPARJN ( Int_t  aIndex,
Double_t  aValue 
)

Definition at line 1217 of file TJetNet.cxx.

1218 {
1219  JNDAT1.PARJN[ aIndex ] = aValue;
1220 }

◆ SetPatternsPerUpdate()

void TJetNet::SetPatternsPerUpdate ( Int_t  aValue)

Definition at line 1105 of file TJetNet.cxx.

1106 {
1107  JNDAT1.MSTJN[ 1 ] = aValue;
1108  if( !m_InitLocked ) this->Init();
1109 }

◆ SetPottsUnits()

void TJetNet::SetPottsUnits ( Int_t  aValue)

Definition at line 1136 of file TJetNet.cxx.

1137 {
1138  JNINT2.IPOTT = aValue;
1139 }

◆ SetUpdatesPerEpoch()

void TJetNet::SetUpdatesPerEpoch ( Int_t  aValue)

Definition at line 1071 of file TJetNet.cxx.

1072 {
1073  // Sets the number of the updates per epoch
1074  JNDAT1.MSTJN[ 8 ] = aValue;
1075  if( !m_InitLocked ) this->Init();
1076 }

◆ SetUpdatingProcedure()

void TJetNet::SetUpdatingProcedure ( Int_t  aValue)

Definition at line 1078 of file TJetNet.cxx.

1079 {
1080  // Set specific weights update function
1081  JNDAT1.MSTJN[ 4 ] = aValue;
1082  if( !m_InitLocked ) this->Init();
1083 }

◆ Shuffle()

void TJetNet::Shuffle ( Bool_t  aShuffleTrainSet = true,
Bool_t  aShuffleTestSet = true 
)

Definition at line 1222 of file TJetNet.cxx.

1223 {
1224  // Shuffles the train and/or test input/output sets
1225  TTimeStamp ts;
1226  Int_t Seed = ts.GetSec();
1227  if ( aShuffleTrainSet )
1228  {
1229 
1230  m_pInputTrainSet->Shuffle( Seed );
1231  m_pOutputTrainSet->Shuffle( Seed );
1232  }
1233  //Shuffle Test Set
1234  if ( aShuffleTestSet )
1235  {
1236  Seed = ts.GetSec();
1237  m_pInputTestSet->Shuffle( Seed );
1238  m_pOutputTestSet->Shuffle( Seed );
1239  }
1240 
1241  return;
1242 }

◆ Test()

Double_t TJetNet::Test ( void  )

Definition at line 320 of file TJetNet.cxx.

321 {
322  // Initiate test cycle of the neural network
323  Int_t NRight = 0;
324  Double_t fMeanError = 0.0;
325  Double_t *TMP;
326  Int_t NPatterns = GetTestSetCnt();
327 
328 
329  for( Int_t iPattern = 0; iPattern < NPatterns; iPattern++ )
330  {
331 
332  for( Int_t i = 0; i < GetInputDim(); i++ )
333  {
334  JNDAT1.OIN[ i ] = float ( GetInputTestSet( iPattern, i ) );
335  }
336 
337  NWJNWGT.OWGT = GetEventWeightTestSet( iPattern );
338 
339  JNTEST();
340 
341  for( Int_t j = 0; j < GetOutputDim(); j++ )
342  {
343  fMeanError+= NWJNWGT.OWGT *
344  std::pow(JNDAT1.OUT[ j ]-float( GetOutputTestSet( iPattern, j )),2)/(float)GetOutputDim();
345  }
346 
347 
348 
349  if( m_Debug ) std::cout << "Testing [ " << iPattern << " ] - " << JNDAT1.OIN[ 0 ]
350  << " => " << JNDAT1.OUT[ 0 ] << std::endl;
351 
352  }
353 
354  fMeanError/=2.*NPatterns;
355 
356  if (m_Debug)
357  std::cout << " Test error: " << fMeanError << endl;
358 
359  return fMeanError;
360 }

◆ TestBTAG()

Double_t TJetNet::TestBTAG ( void  )

Definition at line 362 of file TJetNet.cxx.

363 {
364 
365  bool test=false;
366 
367  // Initiate test cycle of the neural network
368  Int_t NRight = 0;
369  Double_t fMeanError = 0.0;
370  Double_t *TMP;
371  Int_t NPatterns = GetTestSetCnt();
372  if (test)
373  {
374  NPatterns = GetTrainSetCnt();
375  }
376 
377  vector<double> eff;
378  eff.push_back(0.5);
379  eff.push_back(0.6);
380  eff.push_back(0.7);
381 
382  //test also the b-tagging performance directly during testing !!!
383  vector<TH1F*> histoEfficienciesC;
384  vector<TH1F*> histoEfficienciesL;
385  TString histoEffStringC("histoEffC");
386  TString histoEffStringL("histoEffL");
387  for (int i=0;i<GetOutputDim();i++)
388  {
389  TString string1=histoEffStringC;
390  string1+=i;
391  TH1F* histo=new TH1F(string1,
392  string1,
393  20000,
394  -2,3);
395  TString string2=histoEffStringL;
396  string2+=i;
397  TH1F* histo2=new TH1F(string2,
398  string2,
399  20000,
400  -2,3);
401  histoEfficienciesC.push_back(histo);
402  histoEfficienciesL.push_back(histo2);
403  }
404 
405  for( Int_t iPattern = 0; iPattern < NPatterns; iPattern++ )
406  {
407 
408  for( Int_t i = 0; i < GetInputDim(); i++ )
409  {
410  if (!test)
411  {
412  JNDAT1.OIN[ i ] = float ( GetInputTestSet( iPattern, i ) );
413  NWJNWGT.OWGT = GetEventWeightTestSet( iPattern );
414  }
415  else
416  {
417  JNDAT1.OIN[ i ] = float (GetInputTrainSet( iPattern, i ) );
418  NWJNWGT.OWGT = GetEventWeightTrainSet( iPattern );
419  }
420 
421  }
422 
423  JNTEST();
424 
425  int active=0;
426  for( Int_t j = 0; j < GetOutputDim(); j++ )
427  {
428  if (!test)
429  {
430  fMeanError+= NWJNWGT.OWGT *
431  std::pow(JNDAT1.OUT[ j ]-float( GetOutputTestSet( iPattern, j )),2)/(float)GetOutputDim();
432  }
433  else
434  {
435  fMeanError+= NWJNWGT.OWGT *
436  std::pow(JNDAT1.OUT[ j ]-float( GetOutputTrainSet( iPattern, j )),2)/(float)GetOutputDim();
437  }
438 
439 
440  // std::cout << " j " << j << " is " << GetOutputTestSet( iPattern, j) << std::endl;
441 
442  if (!test)
443  {
444  if (fabs(float( GetOutputTestSet( iPattern, j)) - 1) < 1e-4)
445  {
446  active = j;
447  }
448  }
449  else
450  {
451  if (fabs(float( GetOutputTrainSet( iPattern, j)) - 1) < 1e-4)
452  {
453  active = j;
454  }
455  }
456  }
457 
458  // if (m_Debug) std::cout << " active is: " << active << std::endl;
459 
460  // if (m_Debug) std::cout << " filling histograms " << std::endl;
461 
462  if (JNDAT1.OUT[ 0 ] + JNDAT1.OUT[ 1 ] >= 0)
463  {
464  histoEfficienciesC[active]->Fill( JNDAT1.OUT[ 0 ] /
465  ( JNDAT1.OUT[ 0 ] + JNDAT1.OUT[ 1 ]),
466  NWJNWGT.OWGT);
467 
468  // if( m_Debug ) std::cout << "Filled: " << JNDAT1.OUT[ 0 ] /
469  // ( JNDAT1.OUT[ 0 ] + JNDAT1.OUT[ 2 ]) << std::endl;
470 
471  }
472  else
473  {
474  std::cout << " Filled 0 " << std::endl;
475  histoEfficienciesC[active]->Fill( 0 );
476  }
477 
478 
479  if (JNDAT1.OUT[ 0 ] + JNDAT1.OUT[ 2 ] >= 0)
480  {
481  histoEfficienciesL[active]->Fill( JNDAT1.OUT[ 0 ] /
482  ( JNDAT1.OUT[ 0 ] + JNDAT1.OUT[ 2 ]),
483  NWJNWGT.OWGT);
484  // if( m_Debug ) std::cout << "Filled: " << JNDAT1.OUT[ 0 ] /
485  // ( JNDAT1.OUT[ 0 ] + JNDAT1.OUT[ 1 ]) << std::endl;
486 
487  }
488  else
489  {
490  std::cout << " Filled 0 " << std::endl;
491  histoEfficienciesL[active]->Fill( 0 );
492  }
493 
494  if( m_Debug ) std::cout << "Testing [ " << iPattern << " ] - " << JNDAT1.OIN[ 0 ]
495  << " => " << JNDAT1.OUT[ 0 ] << std::endl;
496 
497  }// finish patterns
498 
499  if (m_Debug) std::cout << " Finished patterns... " << std::endl;
500 
501  TFile* newFile=new TFile("test.root","recreate");
502  histoEfficienciesL[0]->Write();
503  histoEfficienciesL[1]->Write();
504  histoEfficienciesL[2]->Write();
505  histoEfficienciesC[0]->Write();
506  histoEfficienciesC[1]->Write();
507  histoEfficienciesC[2]->Write();
508  newFile->Write();
509  newFile->Close();
510 
511  //for C-jet rejection
512 
513  for (int u=0;u<2;u++)
514  {
515  vector<TH1F*>* myVectorHistos;
516  if (u==0)
517  {
518  std::cout << "c-rej --> ";
519  myVectorHistos=&histoEfficienciesC;
520  }
521  if (u==1)
522  {
523  std::cout << "l-rej --> ";
524  myVectorHistos=&histoEfficienciesL;
525  }
526 
527 
528  if (m_Debug) std::cout << " 1 " << std::endl;
529 
530  Double_t allb=(*myVectorHistos)[0]->GetEntries();
531  Double_t allc=(*myVectorHistos)[1]->GetEntries();
532  Double_t allu=(*myVectorHistos)[2]->GetEntries();
533 
534  if (m_Debug) std::cout << " allb " << allb << std::endl;
535 
536  Double_t allbsofar=0;
537 
538  vector<int> binN_Eff;
539  vector<bool> ok_eff;
540 
541  for (int r=0;r<eff.size();r++)
542  {
543  ok_eff.push_back(false);
544  binN_Eff.push_back(0);
545  }
546 
547  for (int s=0;s<(*myVectorHistos)[0]->GetNbinsX()+1;s++) {
548  allbsofar+=(*myVectorHistos)[0]->GetBinContent((*myVectorHistos)[0]->GetNbinsX()+1-s);
549  bool nothingMore(true);
550 
551 
552  for (int r=0;r<eff.size();r++)
553  {
554  if (m_Debug) std::cout << " actual eff: " << allbsofar / allb << std::endl;
555 
556  if ((!ok_eff[r]) && allbsofar / allb > eff[r])
557  {
558  binN_Eff[r]=s;
559  ok_eff[r]=true;
560  if (m_Debug) std::cout << " bin: " << s << " eff: " << allbsofar / allb << std::endl;
561  // std::cout << " Cut value: " << (*myVectorHistos)[0]->GetBinCenter(s) << std::endl;
562  }
563  else if (allbsofar / allb <= eff[r])
564  {
565  nothingMore=false;
566  }
567  }
568  if (nothingMore) break;
569  }
570 
571 
572  for (int r=0;r<eff.size();r++)
573  {
574 
575  std::cout << " " << eff[r];
576 
577  std::cout << " check: " << (double)(*myVectorHistos)[0]->Integral((*myVectorHistos)[0]->GetNbinsX()-binN_Eff[r],
578  (*myVectorHistos)[1]->GetNbinsX()+1)
579  / (double)allb;
580 
581  double effc=(*myVectorHistos)[1]->Integral((*myVectorHistos)[0]->GetNbinsX()-binN_Eff[r],
582  (*myVectorHistos)[1]->GetNbinsX()+1);
583  effc /= allc;
584  double effl=(*myVectorHistos)[2]->Integral((*myVectorHistos)[0]->GetNbinsX()-binN_Eff[r],
585  (*myVectorHistos)[2]->GetNbinsX()+1);
586  effl /= allu;
587 
588  if (effc!=0)
589  {
590  std::cout << " c: " << 1/effc;
591  }
592  if (effl!=0)
593  {
594  std::cout << " l: " << 1/effl;
595  }
596 
597  }
598  std::cout << std::endl;
599  }
600 
601  for( Int_t j = 0; j < GetOutputDim(); j++ )
602  {
603  delete histoEfficienciesC[j];
604  delete histoEfficienciesL[j];
605  }
606 
607 
608  fMeanError/=2.*NPatterns;
609 
610  if (m_Debug)
611  std::cout << " Test error: " << fMeanError << endl;
612 
613  return fMeanError;
614 }

◆ Train()

Double_t TJetNet::Train ( void  )

Definition at line 618 of file TJetNet.cxx.

619 {
620  // Initiate the train phase for the neural network
621  Int_t NRight = 0;
622  Double_t fMeanError = 0.0;
623  Int_t NPatterns = GetTrainSetCnt();
624 
625  // cout << " NPatterns is: " << NPatterns << endl;
626 
627  Int_t inputDim=GetInputDim();
628  Int_t outputDim=GetOutputDim();
629  Int_t updatesPerEpoch=GetUpdatesPerEpoch();
630  Int_t patternsPerUpdate=GetPatternsPerUpdate();
631 
632  if (updatesPerEpoch*patternsPerUpdate<1./2.*NPatterns)
633  {
634  cout << "Using only: " << updatesPerEpoch*patternsPerUpdate <<
635  " patterns on available: " << NPatterns << endl;
636  } else if (updatesPerEpoch*patternsPerUpdate>NPatterns)
637  {
638  cout << " Trying to use " << updatesPerEpoch*patternsPerUpdate <<
639  " patterns, but available: " << NPatterns << endl;
640  return -100;
641  }
642 
643  for( Int_t iPattern = 0; iPattern < updatesPerEpoch*patternsPerUpdate;
644  iPattern++ )
645  {
646  for( Int_t i = 0; i < inputDim; i++ )
647  {
648  JNDAT1.OIN[ i ] = float ( GetInputTrainSet( iPattern, i ) );
649  }
650 
651  NWJNWGT.OWGT = GetEventWeightTrainSet( iPattern );
652 
653  for( Int_t j = 0; j < outputDim; j++ )
654  {
655  JNDAT1.OUT[ j ] = float ( GetOutputTrainSet( iPattern, j ) );
656  }
657 
658  JNTRAL();
659  }
660 
661  return GetPARJN(8);
662 }

◆ UnlockInit()

void TJetNet::UnlockInit ( void  )
inline

Definition at line 138 of file TJetNet.h.

138 { m_InitLocked = kFALSE; };

◆ writeNetworkInfo()

void TJetNet::writeNetworkInfo ( Int_t  typeOfInfo = 0)

Definition at line 664 of file TJetNet.cxx.

665 {
666  cout << " Invoking info of type: " << typeOfInfo << endl;
667  JNSTAT(typeOfInfo);
668 }

Member Data Documentation

◆ m_CurrentEpoch

Int_t TJetNet::m_CurrentEpoch
private

Definition at line 186 of file TJetNet.h.

◆ m_Debug

Bool_t TJetNet::m_Debug
private

Definition at line 187 of file TJetNet.h.

◆ m_enActFunction

TActivationFunction TJetNet::m_enActFunction {}
private

Definition at line 170 of file TJetNet.h.

◆ m_Epochs

Int_t TJetNet::m_Epochs
private

Definition at line 185 of file TJetNet.h.

◆ m_HiddenLayerDim

Int_t TJetNet::m_HiddenLayerDim
private

Definition at line 183 of file TJetNet.h.

◆ m_InitLocked

Bool_t TJetNet::m_InitLocked
private

Definition at line 189 of file TJetNet.h.

◆ m_InputDim

Int_t TJetNet::m_InputDim
private

Definition at line 182 of file TJetNet.h.

◆ m_IsInitialized

Bool_t TJetNet::m_IsInitialized
private

Definition at line 188 of file TJetNet.h.

◆ m_LayerCount

Int_t TJetNet::m_LayerCount
private

Definition at line 172 of file TJetNet.h.

◆ m_NormalizeOutput

Bool_t TJetNet::m_NormalizeOutput
private

Definition at line 190 of file TJetNet.h.

◆ m_OutputDim

Int_t TJetNet::m_OutputDim
private

Definition at line 184 of file TJetNet.h.

◆ m_pInputTestSet

TNeuralDataSet* TJetNet::m_pInputTestSet
private

Definition at line 177 of file TJetNet.h.

◆ m_pInputTrainSet

TNeuralDataSet* TJetNet::m_pInputTrainSet
private

Array which contains the number of units in each layer.

Definition at line 175 of file TJetNet.h.

◆ m_pLayers

Int_t* TJetNet::m_pLayers
private

Definition at line 173 of file TJetNet.h.

◆ m_pOutputTestSet

TNeuralDataSet* TJetNet::m_pOutputTestSet
private

Definition at line 178 of file TJetNet.h.

◆ m_pOutputTrainSet

TNeuralDataSet* TJetNet::m_pOutputTrainSet
private

Definition at line 176 of file TJetNet.h.

◆ m_TestSetCnt

Int_t TJetNet::m_TestSetCnt
private

Definition at line 180 of file TJetNet.h.

◆ m_TrainSetCnt

Int_t TJetNet::m_TrainSetCnt
private

Definition at line 180 of file TJetNet.h.

The documentation for this class was generated from the following files:
TNeuralDataSet
Definition: TNeuralDataSet.h:26
TTrainedNetwork::getnInput
Int_t getnInput() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:46
JNDAT1
#define JNDAT1
Definition: jetnet.h:42
AllowedVariables::e
e
Definition: AsgElectronSelectorTool.cxx:37
beamspotman.r
def r
Definition: beamspotman.py:676
TJetNet::m_enActFunction
TActivationFunction m_enActFunction
Definition: TJetNet.h:170
TJetNet::m_CurrentEpoch
Int_t m_CurrentEpoch
Definition: TJetNet.h:186
TJetNet::afSigmoid
@ afSigmoid
Definition: TJetNet.h:151
TJetNet::GetTestSetCnt
Int_t GetTestSetCnt(void) const
Definition: TJetNet.h:53
python.SystemOfUnits.s
int s
Definition: SystemOfUnits.py:131
python.SystemOfUnits.m
int m
Definition: SystemOfUnits.py:91
TJetNet::GetErrorMeasure
Int_t GetErrorMeasure(void)
Definition: TJetNet.cxx:1156
TJetNet::GetLearningRate
Double_t GetLearningRate(void)
Definition: TJetNet.cxx:1177
JNSTAT
#define JNSTAT(IS)
Definition: jetnet.h:167
JNDUMP
#define JNDUMP(NF)
Definition: jetnet.h:154
NWJNWGT
#define NWJNWGT
Definition: jetnet.h:51
TJetNet::GetEventWeightTestSet
Double_t GetEventWeightTestSet(Int_t aPatternInd)
Definition: TJetNet.h:213
TJetNet::m_TrainSetCnt
Int_t m_TrainSetCnt
Definition: TJetNet.h:180
TNeuralDataSet::Randomize
void Randomize(void)
Definition: TNeuralDataSet.cxx:74
TNeuralDataSet::SetData
void SetData(const Int_t aPattern, const Int_t aIndex, Double_t aValue)
Definition: TNeuralDataSet.cxx:37
TJetNet::m_InputDim
Int_t m_InputDim
Definition: TJetNet.h:182
TJetNet::m_OutputDim
Int_t m_OutputDim
Definition: TJetNet.h:184
TJetNet::GetMomentum
Double_t GetMomentum(void)
Definition: TJetNet.cxx:1182
TJetNet::Init
void Init(void)
Definition: TJetNet.cxx:670
python.AthDsoLogger.out
out
Definition: AthDsoLogger.py:71
TJetNet::SetOutputTestSet
void SetOutputTestSet(Int_t aPatternInd, Int_t aOutputInd, Double_t aValue)
Definition: TJetNet.h:196
TJetNet::GetLearningRateDecrease
Double_t GetLearningRateDecrease(void)
Definition: TJetNet.cxx:1192
TNeuralDataSet::GetData
Double_t GetData(const Int_t aPattern, const Int_t aIndex)
Definition: TNeuralDataSet.h:68
UploadAMITag.l
list l
Definition: UploadAMITag.larcaf.py:158
TJetNet::m_TestSetCnt
Int_t m_TestSetCnt
Definition: TJetNet.h:180
TJetNet::m_pInputTrainSet
TNeuralDataSet * m_pInputTrainSet
Array which contains the number of units in each layer.
Definition: TJetNet.h:175
TrigInDetValidation_Base.test
test
Definition: TrigInDetValidation_Base.py:147
TNeuralDataSet::Shuffle
void Shuffle(Int_t aSeed=0)
Definition: TNeuralDataSet.cxx:85
TJetNet::GetOutputTrainSet
Double_t GetOutputTrainSet(Int_t aPatternInd, Int_t aOutputInd)
Definition: TJetNet.cxx:746
TTrainedNetwork::getnHiddenLayerSize
const std::vector< Int_t > & getnHiddenLayerSize() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:52
TJetNet::GetUpdatesPerEpoch
Int_t GetUpdatesPerEpoch(void)
Definition: TJetNet.cxx:1146
TJetNet::SetEpochs
void SetEpochs(const Int_t aEpochs)
Definition: TJetNet.h:79
TJetNet::m_NormalizeOutput
Bool_t m_NormalizeOutput
Definition: TJetNet.h:190
TTrainedNetwork::getnHidden
Int_t getnHidden() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:48
TTrainedNetwork::weightMatrices
const std::vector< TMatrixD * > & weightMatrices() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:58
TJetNet::m_InitLocked
Bool_t m_InitLocked
Definition: TJetNet.h:189
Trk::u
@ u
Enums for curvilinear frames.
Definition: ParamDefs.h:77
TJetNet::GetInputTrainSet
Double_t GetInputTrainSet(Int_t aPatternInd, Int_t aInputInd)
Definition: TJetNet.h:202
TJetNet::SetInputTestSet
void SetInputTestSet(Int_t aPatternInd, Int_t aInputInd, Double_t aValue)
Definition: TJetNet.cxx:740
TJetNet::mSetThreshold
void mSetThreshold(Double_t threshold, Int_t aLayerInd, Int_t aNodeInd)
Definition: TJetNet.cxx:290
JNINIT
#define JNINIT()
Definition: jetnet.h:134
dqt_zlumi_pandas.weight
int weight
Definition: dqt_zlumi_pandas.py:189
TJetNet::m_Debug
Bool_t m_Debug
Definition: TJetNet.h:187
TJetNet::Reinitialize
void Reinitialize(void)
Definition: TJetNet.cxx:1020
python.utils.AtlRunQueryDQUtils.p
p
Definition: AtlRunQueryDQUtils.py:210
Trk::active
@ active
Definition: Layer.h:48
doubleTestComp.j1
j1
Definition: doubleTestComp.py:21
lumiFormat.i
int i
Definition: lumiFormat.py:85
TJetNet::m_pOutputTrainSet
TNeuralDataSet * m_pOutputTrainSet
Definition: TJetNet.h:176
JNSEFI
#define JNSEFI(ILA, I1, I2, J1, J2, NO)
Definition: jetnet.h:173
vector< double >
TJetNet::GetInputTestSet
Double_t GetInputTestSet(Int_t aPatternInd, Int_t aInputInd)
Definition: TJetNet.cxx:763
TJetNet::afTanh
@ afTanh
Definition: TJetNet.h:152
TJetNet::GetWeight
Double_t GetWeight(Int_t aLayerInd, Int_t aNodeInd, Int_t aConnectedNodeInd) const
Definition: TJetNet.cxx:972
TJetNet::m_IsInitialized
Bool_t m_IsInitialized
Definition: TJetNet.h:188
TJetNet::m_pLayers
Int_t * m_pLayers
Definition: TJetNet.h:173
JNTEST
#define JNTEST()
Definition: jetnet.h:148
ActsTrk::Seed
Acts::Seed< xAOD::SpacePoint, 3ul > Seed
Definition: Seed.h:12
TJetNet::m_pInputTestSet
TNeuralDataSet * m_pInputTestSet
Definition: TJetNet.h:177
TJetNet::SetOutputTrainSet
void SetOutputTrainSet(Int_t aPatternInd, Int_t aOutputInd, Double_t aValue)
Definition: TJetNet.cxx:734
TJetNet::SetInputTrainSet
void SetInputTrainSet(Int_t aPatternInd, Int_t aInputInd, Double_t aValue)
Definition: TJetNet.cxx:728
xAOD::double
double
Definition: CompositeParticle_v1.cxx:159
DeMoUpdate.tmp
string tmp
Definition: DeMoUpdate.py:1167
TTrainedNetwork::getThresholdVectors
const std::vector< TVectorD * > & getThresholdVectors() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:56
TJetNet::GetPatternsPerUpdate
Int_t GetPatternsPerUpdate(void)
Definition: TJetNet.cxx:1172
TJetNet::mSetWeight
void mSetWeight(Double_t weight, Int_t aLayerInd, Int_t aNodeInd, Int_t aConnectedNodeInd)
Definition: TJetNet.cxx:285
TTrainedNetwork
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:21
TJetNet::GetTrainSetCnt
Int_t GetTrainSetCnt(void) const
Definition: TJetNet.h:52
TJetNet::GetEpochs
Int_t GetEpochs(void)
Definition: TJetNet.h:78
TJetNet::m_Epochs
Int_t m_Epochs
Definition: TJetNet.h:185
TJetNet::GetActivationFunction
Int_t GetActivationFunction(void) const
Definition: TJetNet.cxx:1161
TJetNet::GetOutputDim
Int_t GetOutputDim(void) const
Definition: TJetNet.h:57
python.selection.number
number
Definition: selection.py:20
TJetNet::afLinear
@ afLinear
Definition: TJetNet.h:154
TJetNet::m_HiddenLayerDim
Int_t m_HiddenLayerDim
Definition: TJetNet.h:183
TJetNet::GetThreshold
Double_t GetThreshold(Int_t aLayerInd, Int_t aNodeInd) const
Definition: TJetNet.cxx:979
TJetNet::GetInitialWeightsWidth
Double_t GetInitialWeightsWidth(void)
Definition: TJetNet.cxx:1187
JNTRAL
#define JNTRAL()
Definition: jetnet.h:141
TNeuralDataSet::SetEventWeight
void SetEventWeight(const Int_t aPattern, Double_t aValue)
Definition: TNeuralDataSet.cxx:32
JNINT1
#define JNINT1
Definition: jetnet.h:103
TJetNet::afSigmoidEntropy
@ afSigmoidEntropy
Definition: TJetNet.h:155
JNREAD
#define JNREAD(NF)
Definition: jetnet.h:161
JNINDX
#define JNINDX(IL, I, J)
Definition: jetnet.h:112
TJetNet::SetActivationFunction
void SetActivationFunction(Int_t aValue)
Definition: TJetNet.cxx:1091
TNeuralDataSet::GetEventWeight
Double_t GetEventWeight(const Int_t aPattern)
Definition: TNeuralDataSet.h:74
TJetNet::afExp
@ afExp
Definition: TJetNet.h:153
TJetNet::GetInputDim
Int_t GetInputDim(void) const
Definition: TJetNet.h:54
TJetNet::GetHiddenLayerSize
Int_t GetHiddenLayerSize(Int_t number) const
Definition: TJetNet.h:56
TTrainedNetwork::getIfLinearOutput
bool getIfLinearOutput() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:62
TTrainedNetwork::getnOutput
Int_t getnOutput() const
Definition: InnerDetector/InDetCalibAlgs/PixelCalibAlgs/NNClusteringCalibration_RunI/TTrainedNetwork.h:50
TJetNet::GetPARJN
Double_t GetPARJN(Int_t aIndex)
Definition: TJetNet.cxx:1207
dqt_zlumi_alleff_HIST.eff
int eff
Definition: dqt_zlumi_alleff_HIST.py:113
python.TrigEgammaMonitorHelper.TH1F
def TH1F(name, title, nxbins, bins_par2, bins_par3=None, path='', **kwargs)
Definition: TrigEgammaMonitorHelper.py:24
TNeuralDataSet::Normalize
void Normalize(void)
Definition: TNeuralDataSet.cxx:43
TJetNet::GetOutputTestSet
Double_t GetOutputTestSet(Int_t aPatternInd, Int_t aOutputInd)
Definition: TJetNet.cxx:769
TJetNet::m_pOutputTestSet
TNeuralDataSet * m_pOutputTestSet
Definition: TJetNet.h:178
TJetNet::Test
Double_t Test(void)
Definition: TJetNet.cxx:320
TJetNet::SetPottsUnits
void SetPottsUnits(Int_t aValue)
Definition: TJetNet.cxx:1136
python.CaloScaleNoiseConfig.ts
ts
Definition: CaloScaleNoiseConfig.py:86
plotBeamSpotCompare.histo
histo
Definition: plotBeamSpotCompare.py:415
pow
constexpr int pow(int base, int exp) noexcept
Definition: ap_fixedTest.cxx:15
JNDAT2
#define JNDAT2
Definition: jetnet.h:65
dumpTgcDigiThreshold.threshold
list threshold
Definition: dumpTgcDigiThreshold.py:34
readCCLHist.float
float
Definition: readCCLHist.py:83
TJetNet::GetEventWeightTrainSet
Double_t GetEventWeightTrainSet(Int_t aPatternInd)
Definition: TJetNet.h:208
doubleTestComp.j2
j2
Definition: doubleTestComp.py:22
TJetNet::Train
Double_t Train(void)
Definition: TJetNet.cxx:618
TJetNet::m_LayerCount
Int_t m_LayerCount
Definition: TJetNet.h:172
fitman.k
k
Definition: fitman.py:528
TJetNet::GetUpdatingProcedure
Int_t GetUpdatingProcedure(void)
Definition: TJetNet.cxx:1151
JNINT2
#define JNINT2
Definition: jetnet.h:83
TJetNet::GetHiddenLayerDim
Int_t GetHiddenLayerDim(void) const
Definition: TJetNet.h:55
Generated on Sun Dec 22 2024 21:49:09 for ATLAS Offline Software by   doxygen 1.8.18