TMVAClassification_MLPBNN.class.C

Go to the documentation of this file.

// Class: ReadMLPBNN
// Automatically generated by MethodBase::MakeClass
//
 
/* configuration options =====================================================
 
#GEN -*-*-*-*-*-*-*-*-*-*-*- general info -*-*-*-*-*-*-*-*-*-*-*-
 
Method         : MLP::MLPBNN
TMVA Release   : 4.2.1         [262657]
ROOT Release   : 6.16/01       [397313]
Creator        : sftnight
Date           : Sun Dec 19 22:13:35 2021
Host           : Linux root-ubuntu-2004-3 5.4.0-73-generic #82-Ubuntu SMP Wed Apr 14 17:39:42 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
Dir            : /home/sftnight/build/workspace/root-makedoc-v616/rootspi/rdoc/src/v6-16-00-patches/documentation/doxygen
Training events: 2000
Analysis type  : [Classification]
 
 
#OPT -*-*-*-*-*-*-*-*-*-*-*-*- options -*-*-*-*-*-*-*-*-*-*-*-*-
 
# Set by User:
NCycles: "60" [Number of training cycles]
HiddenLayers: "N+5" [Specification of hidden layer architecture]
NeuronType: "tanh" [Neuron activation function type]
V: "False" [Verbose output (short form of "VerbosityLevel" below - overrides the latter one)]
VarTransform: "N" [List of variable transformations performed before training, e.g., "D_Background,P_Signal,G,N_AllClasses" for: "Decorrelation, PCA-transformation, Gaussianisation, Normalisation, each for the given class of events ('AllClasses' denotes all events of all classes, if no class indication is given, 'All' is assumed)"]
H: "True" [Print method-specific help message]
TrainingMethod: "BFGS" [Train with Back-Propagation (BP), BFGS Algorithm (BFGS), or Genetic Algorithm (GA - slower and worse)]
TestRate: "5" [Test for overtraining performed at each #th epochs]
UseRegulator: "True" [Use regulator to avoid over-training]
# Default:
RandomSeed: "1" [Random seed for initial synapse weights (0 means unique seed for each run; default value '1')]
EstimatorType: "CE" [MSE (Mean Square Estimator) for Gaussian Likelihood or CE(Cross-Entropy) for Bernoulli Likelihood]
NeuronInputType: "sum" [Neuron input function type]
VerbosityLevel: "Default" [Verbosity level]
CreateMVAPdfs: "False" [Create PDFs for classifier outputs (signal and background)]
IgnoreNegWeightsInTraining: "False" [Events with negative weights are ignored in the training (but are included for testing and performance evaluation)]
LearningRate: "2.000000e-02" [ANN learning rate parameter]
DecayRate: "1.000000e-02" [Decay rate for learning parameter]
EpochMonitoring: "False" [Provide epoch-wise monitoring plots according to TestRate (caution: causes big ROOT output file!)]
Sampling: "1.000000e+00" [Only 'Sampling' (randomly selected) events are trained each epoch]
SamplingEpoch: "1.000000e+00" [Sampling is used for the first 'SamplingEpoch' epochs, afterwards, all events are taken for training]
SamplingImportance: "1.000000e+00" [ The sampling weights of events in epochs which successful (worse estimator than before) are multiplied with SamplingImportance, else they are divided.]
SamplingTraining: "True" [The training sample is sampled]
SamplingTesting: "False" [The testing sample is sampled]
ResetStep: "50" [How often BFGS should reset history]
Tau: "3.000000e+00" [LineSearch "size step"]
BPMode: "sequential" [Back-propagation learning mode: sequential or batch]
BatchSize: "-1" [Batch size: number of events/batch, only set if in Batch Mode, -1 for BatchSize=number_of_events]
ConvergenceImprove: "1.000000e-30" [Minimum improvement which counts as improvement (<0 means automatic convergence check is turned off)]
ConvergenceTests: "-1" [Number of steps (without improvement) required for convergence (<0 means automatic convergence check is turned off)]
UpdateLimit: "10000" [Maximum times of regulator update]
CalculateErrors: "False" [Calculates inverse Hessian matrix at the end of the training to be able to calculate the uncertainties of an MVA value]
WeightRange: "1.000000e+00" [Take the events for the estimator calculations from small deviations from the desired value to large deviations only over the weight range]
##
 
 
#VAR -*-*-*-*-*-*-*-*-*-*-*-* variables *-*-*-*-*-*-*-*-*-*-*-*-
 
NVar 4
var1+var2                     myvar1                        myvar1                        myvar1                                                          'F'    [-9.33803939819,7.69307804108]
var1-var2                     myvar2                        myvar2                        Expression 2                                                    'F'    [-3.25508260727,4.02912044525]
var3                          var3                          var3                          Variable 3                    units                             'F'    [-5.2777428627,4.64297914505]
var4                          var4                          var4                          Variable 4                    units                             'F'    [-5.6007027626,4.67435789108]
NSpec 2
var1*2                        spec1                         spec1                         Spectator 1                   units                             'F'    [-9.91655540466,8.7030172348]
var1*3                        spec2                         spec2                         Spectator 2                   units                             'F'    [-14.874833107,13.0545253754]
 
 
============================================================================ */
 
#include <array>
#include <vector>
#include <cmath>
#include <string>
#include <iostream>
 
#ifndef IClassifierReader__def
#define IClassifierReader__def
 
class IClassifierReader {
 
 public:
 
   // constructor
   IClassifierReader() : fStatusIsClean( true ) {}
   virtual ~IClassifierReader() {}
 
   // return classifier response
   virtual double GetMvaValue( const std::vector<double>& inputValues ) const = 0;
 
   // returns classifier status
   bool IsStatusClean() const { return fStatusIsClean; }
 
 protected:
 
   bool fStatusIsClean;
};
 
#endif
 
class ReadMLPBNN : public IClassifierReader {
 
 public:
 
   // constructor
   ReadMLPBNN( std::vector<std::string>& theInputVars )
      : IClassifierReader(),
        fClassName( "ReadMLPBNN" ),
        fNvars( 4 )
   {
      // the training input variables
      const char* inputVars[] = { "var1+var2", "var1-var2", "var3", "var4" };
 
      // sanity checks
      if (theInputVars.size() <= 0) {
         std::cout << "Problem in class \"" << fClassName << "\": empty input vector" << std::endl;
         fStatusIsClean = false;
      }
 
      if (theInputVars.size() != fNvars) {
         std::cout << "Problem in class \"" << fClassName << "\": mismatch in number of input values: "
                   << theInputVars.size() << " != " << fNvars << std::endl;
         fStatusIsClean = false;
      }
 
      // validate input variables
      for (size_t ivar = 0; ivar < theInputVars.size(); ivar++) {
         if (theInputVars[ivar] != inputVars[ivar]) {
            std::cout << "Problem in class \"" << fClassName << "\": mismatch in input variable names" << std::endl
                      << " for variable [" << ivar << "]: " << theInputVars[ivar].c_str() << " != " << inputVars[ivar] << std::endl;
            fStatusIsClean = false;
         }
      }
 
      // initialize min and max vectors (for normalisation)
      fVmin[0] = -1;
      fVmax[0] = 1;
      fVmin[1] = -1;
      fVmax[1] = 1;
      fVmin[2] = -1;
      fVmax[2] = 1;
      fVmin[3] = -1;
      fVmax[3] = 0.99999988079071;
 
      // initialize input variable types
      fType[0] = 'F';
      fType[1] = 'F';
      fType[2] = 'F';
      fType[3] = 'F';
 
      // initialize constants
      Initialize();
 
      // initialize transformation
      InitTransform();
   }
 
   // destructor
   virtual ~ReadMLPBNN() {
      Clear(); // method-specific
   }
 
   // the classifier response
   // "inputValues" is a vector of input values in the same order as the
   // variables given to the constructor
   double GetMvaValue( const std::vector<double>& inputValues ) const override;
 
 private:
 
   // method-specific destructor
   void Clear();
 
   // input variable transformation
 
   double fOff_1[3][4];
   double fScal_1[3][4];
   void InitTransform_1();
   void Transform_1( std::vector<double> & iv, int sigOrBgd ) const;
   void InitTransform();
   void Transform( std::vector<double> & iv, int sigOrBgd ) const;
 
   // common member variables
   const char* fClassName;
 
   const size_t fNvars;
   size_t GetNvar()           const { return fNvars; }
   char   GetType( int ivar ) const { return fType[ivar]; }
 
   // normalisation of input variables
   double fVmin[4];
   double fVmax[4];
   double NormVariable( double x, double xmin, double xmax ) const {
      // normalise to output range: [-1, 1]
      return 2*(x - xmin)/(xmax - xmin) - 1.0;
   }
 
   // type of input variable: 'F' or 'I'
   char   fType[4];
 
   // initialize internal variables
   void Initialize();
   double GetMvaValue__( const std::vector<double>& inputValues ) const;
 
   // private members (method specific)
 
   double ActivationFnc(double x) const;
   double OutputActivationFnc(double x) const;
 
   double fWeightMatrix0to1[10][5];   // weight matrix from layer 0 to 1
   double fWeightMatrix1to2[1][10];   // weight matrix from layer 1 to 2
 
};
 
inline void ReadMLPBNN::Initialize()
{
   // build network structure
   // weight matrix from layer 0 to 1
   fWeightMatrix0to1[0][0] = 0.499339155868432;
   fWeightMatrix0to1[1][0] = 1.97323370224732;
   fWeightMatrix0to1[2][0] = 0.835850497820873;
   fWeightMatrix0to1[3][0] = 0.370715139038071;
   fWeightMatrix0to1[4][0] = -3.3244780544242;
   fWeightMatrix0to1[5][0] = -1.00515616054487;
   fWeightMatrix0to1[6][0] = -0.903517469691334;
   fWeightMatrix0to1[7][0] = 1.68095300796643;
   fWeightMatrix0to1[8][0] = -1.8086188219696;
   fWeightMatrix0to1[0][1] = -0.469787533689036;
   fWeightMatrix0to1[1][1] = -1.47308437821555;
   fWeightMatrix0to1[2][1] = -0.548985376762871;
   fWeightMatrix0to1[3][1] = -1.08352747767343;
   fWeightMatrix0to1[4][1] = -0.706728846639731;
   fWeightMatrix0to1[5][1] = -0.973391622880094;
   fWeightMatrix0to1[6][1] = 0.659572707267351;
   fWeightMatrix0to1[7][1] = -0.31695950236766;
   fWeightMatrix0to1[8][1] = 2.4093761098387;
   fWeightMatrix0to1[0][2] = 0.355913509480398;
   fWeightMatrix0to1[1][2] = 1.40361366583309;
   fWeightMatrix0to1[2][2] = -0.210743959789661;
   fWeightMatrix0to1[3][2] = -0.978127205874705;
   fWeightMatrix0to1[4][2] = -1.01949963347685;
   fWeightMatrix0to1[5][2] = 0.897826904860834;
   fWeightMatrix0to1[6][2] = -1.25600410498056;
   fWeightMatrix0to1[7][2] = -0.743241454883446;
   fWeightMatrix0to1[8][2] = 1.48931411037645;
   fWeightMatrix0to1[0][3] = -2.17450079985913;
   fWeightMatrix0to1[1][3] = -1.60207885420351;
   fWeightMatrix0to1[2][3] = 0.552988361084977;
   fWeightMatrix0to1[3][3] = 2.28075263985257;
   fWeightMatrix0to1[4][3] = 4.00452196921293;
   fWeightMatrix0to1[5][3] = 1.33661120213445;
   fWeightMatrix0to1[6][3] = 0.0806551453627355;
   fWeightMatrix0to1[7][3] = -0.523552944740624;
   fWeightMatrix0to1[8][3] = 0.925130649598083;
   fWeightMatrix0to1[0][4] = -0.565154618007088;
   fWeightMatrix0to1[1][4] = 2.36601716651037;
   fWeightMatrix0to1[2][4] = -1.05353132940242;
   fWeightMatrix0to1[3][4] = -0.163502849122647;
   fWeightMatrix0to1[4][4] = -0.194079160790969;
   fWeightMatrix0to1[5][4] = 0.98449705780042;
   fWeightMatrix0to1[6][4] = 1.9250180795564;
   fWeightMatrix0to1[7][4] = 1.61703868436564;
   fWeightMatrix0to1[8][4] = -0.686657299656178;
   // weight matrix from layer 1 to 2
   fWeightMatrix1to2[0][0] = -2.23742003371891;
   fWeightMatrix1to2[0][1] = 0.277464222426375;
   fWeightMatrix1to2[0][2] = -0.203363715034144;
   fWeightMatrix1to2[0][3] = 1.57205142891781;
   fWeightMatrix1to2[0][4] = 5.94689352294752;
   fWeightMatrix1to2[0][5] = 1.15566920637399;
   fWeightMatrix1to2[0][6] = 1.26453893993009;
   fWeightMatrix1to2[0][7] = -1.34996651290595;
   fWeightMatrix1to2[0][8] = 1.53068569815131;
   fWeightMatrix1to2[0][9] = -1.7128873155142;
}
 
inline double ReadMLPBNN::GetMvaValue__( const std::vector<double>& inputValues ) const
{
   if (inputValues.size() != (unsigned int)4) {
      std::cout << "Input vector needs to be of size " << 4 << std::endl;
      return 0;
   }
 
   std::array<double, 10> fWeights1 {{}};
   std::array<double, 1> fWeights2 {{}};
   fWeights1.back() = 1.;
 
   // layer 0 to 1
   for (int o=0; o<9; o++) {
      std::array<double, 5> buffer; // no need to initialise
      for (int i = 0; i<5 - 1; i++) {
         buffer[i] = fWeightMatrix0to1[o][i] * inputValues[i];
      } // loop over i
      buffer.back() = fWeightMatrix0to1[o][4];      for (int i=0; i<5; i++) {
         fWeights1[o] += buffer[i];
      } // loop over i
    } // loop over o
   for (int o=0; o<9; o++) {
      fWeights1[o] = ActivationFnc(fWeights1[o]);
   } // loop over o
   // layer 1 to 2
   for (int o=0; o<1; o++) {
      std::array<double, 10> buffer; // no need to initialise
      for (int i=0; i<10; i++) {
         buffer[i] = fWeightMatrix1to2[o][i] * fWeights1[i];
      } // loop over i
      for (int i=0; i<10; i++) {
         fWeights2[o] += buffer[i];
      } // loop over i
    } // loop over o
   for (int o=0; o<1; o++) {
      fWeights2[o] = OutputActivationFnc(fWeights2[o]);
   } // loop over o
 
   return fWeights2[0];
}
 
double ReadMLPBNN::ActivationFnc(double x) const {
   // hyperbolic tan
   return tanh(x);
}
double ReadMLPBNN::OutputActivationFnc(double x) const {
   // sigmoid
   return 1.0/(1.0+exp(-x));
}
 
// Clean up
inline void ReadMLPBNN::Clear()
{
}
   inline double ReadMLPBNN::GetMvaValue( const std::vector<double>& inputValues ) const
   {
      // classifier response value
      double retval = 0;
 
      // classifier response, sanity check first
      if (!IsStatusClean()) {
         std::cout << "Problem in class \"" << fClassName << "\": cannot return classifier response"
                   << " because status is dirty" << std::endl;
         retval = 0;
      }
      else {
            std::vector<double> iV(inputValues);
            Transform( iV, -1 );
            retval = GetMvaValue__( iV );
      }
 
      return retval;
   }
 
//_______________________________________________________________________
inline void ReadMLPBNN::InitTransform_1()
{
   double fMin_1[3][4];
   double fMax_1[3][4];
   // Normalization transformation, initialisation
   fMin_1[0][0] = -4.33593845367;
   fMax_1[0][0] = 6.3994679451;
   fScal_1[0][0] = 2.0/(fMax_1[0][0]-fMin_1[0][0]);
   fOff_1[0][0] = fMin_1[0][0]*fScal_1[0][0]+1.;
   fMin_1[1][0] = -9.33803939819;
   fMax_1[1][0] = 7.69307804108;
   fScal_1[1][0] = 2.0/(fMax_1[1][0]-fMin_1[1][0]);
   fOff_1[1][0] = fMin_1[1][0]*fScal_1[1][0]+1.;
   fMin_1[2][0] = -9.33803939819;
   fMax_1[2][0] = 7.69307804108;
   fScal_1[2][0] = 2.0/(fMax_1[2][0]-fMin_1[2][0]);
   fOff_1[2][0] = fMin_1[2][0]*fScal_1[2][0]+1.;
   fMin_1[0][1] = -3.20988440514;
   fMax_1[0][1] = 4.02912044525;
   fScal_1[0][1] = 2.0/(fMax_1[0][1]-fMin_1[0][1]);
   fOff_1[0][1] = fMin_1[0][1]*fScal_1[0][1]+1.;
   fMin_1[1][1] = -3.25508260727;
   fMax_1[1][1] = 3.36500358582;
   fScal_1[1][1] = 2.0/(fMax_1[1][1]-fMin_1[1][1]);
   fOff_1[1][1] = fMin_1[1][1]*fScal_1[1][1]+1.;
   fMin_1[2][1] = -3.25508260727;
   fMax_1[2][1] = 4.02912044525;
   fScal_1[2][1] = 2.0/(fMax_1[2][1]-fMin_1[2][1]);
   fOff_1[2][1] = fMin_1[2][1]*fScal_1[2][1]+1.;
   fMin_1[0][2] = -2.60635733604;
   fMax_1[0][2] = 3.86989831924;
   fScal_1[0][2] = 2.0/(fMax_1[0][2]-fMin_1[0][2]);
   fOff_1[0][2] = fMin_1[0][2]*fScal_1[0][2]+1.;
   fMin_1[1][2] = -5.2777428627;
   fMax_1[1][2] = 4.64297914505;
   fScal_1[1][2] = 2.0/(fMax_1[1][2]-fMin_1[1][2]);
   fOff_1[1][2] = fMin_1[1][2]*fScal_1[1][2]+1.;
   fMin_1[2][2] = -5.2777428627;
   fMax_1[2][2] = 4.64297914505;
   fScal_1[2][2] = 2.0/(fMax_1[2][2]-fMin_1[2][2]);
   fOff_1[2][2] = fMin_1[2][2]*fScal_1[2][2]+1.;
   fMin_1[0][3] = -2.1695792675;
   fMax_1[0][3] = 4.5351858139;
   fScal_1[0][3] = 2.0/(fMax_1[0][3]-fMin_1[0][3]);
   fOff_1[0][3] = fMin_1[0][3]*fScal_1[0][3]+1.;
   fMin_1[1][3] = -5.6007027626;
   fMax_1[1][3] = 4.67435789108;
   fScal_1[1][3] = 2.0/(fMax_1[1][3]-fMin_1[1][3]);
   fOff_1[1][3] = fMin_1[1][3]*fScal_1[1][3]+1.;
   fMin_1[2][3] = -5.6007027626;
   fMax_1[2][3] = 4.67435789108;
   fScal_1[2][3] = 2.0/(fMax_1[2][3]-fMin_1[2][3]);
   fOff_1[2][3] = fMin_1[2][3]*fScal_1[2][3]+1.;
}
 
//_______________________________________________________________________
inline void ReadMLPBNN::Transform_1( std::vector<double>& iv, int cls) const
{
   // Normalization transformation
   if (cls < 0 || cls > 2) {
   if (2 > 1 ) cls = 2;
      else cls = 2;
   }
   const int nVar = 4;
 
   // get indices of used variables
 
   // define the indices of the variables which are transformed by this transformation
   static std::vector<int> indicesGet;
   static std::vector<int> indicesPut;
 
   if ( indicesGet.empty() ) {
      indicesGet.reserve(fNvars);
      indicesGet.push_back( 0);
      indicesGet.push_back( 1);
      indicesGet.push_back( 2);
      indicesGet.push_back( 3);
   }
   if ( indicesPut.empty() ) {
      indicesPut.reserve(fNvars);
      indicesPut.push_back( 0);
      indicesPut.push_back( 1);
      indicesPut.push_back( 2);
      indicesPut.push_back( 3);
   }
 
   static std::vector<double> dv;
   dv.resize(nVar);
   for (int ivar=0; ivar<nVar; ivar++) dv[ivar] = iv[indicesGet.at(ivar)];
   for (int ivar=0;ivar<4;ivar++) {
      double offset = fOff_1[cls][ivar];
      double scale  = fScal_1[cls][ivar];
      iv[indicesPut.at(ivar)] = scale*dv[ivar]-offset;
   }
}
 
//_______________________________________________________________________
inline void ReadMLPBNN::InitTransform()
{
   InitTransform_1();
}
 
//_______________________________________________________________________
inline void ReadMLPBNN::Transform( std::vector<double>& iv, int sigOrBgd ) const
{
   Transform_1( iv, sigOrBgd );
}