Logo ROOT  
Reference Guide
MCFitter.cxx
Go to the documentation of this file.
1// @(#)root/tmva $Id$
2// Author: Andreas Hoecker, Peter Speckmayer, Joerg Stelzer, Helge Voss
3
4/**********************************************************************************
5 * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
6 * Package: TMVA *
7 * Class : TMVA::MCFitter *
8 * Web : http://tmva.sourceforge.net *
9 * *
10 * Description: *
11 * Implementation *
12 * *
13 * Authors (alphabetical): *
14 * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
15 * Peter Speckmayer <speckmay@mail.cern.ch> - CERN, Switzerland *
16 * Joerg Stelzer <Joerg.Stelzer@cern.ch> - CERN, Switzerland *
17 * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
18 * *
19 * Copyright (c) 2005: *
20 * CERN, Switzerland *
21 * MPI-K Heidelberg, Germany *
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/*! \class TMVA::MCFitter
29\ingroup TMVA
30
31Fitter using Monte Carlo sampling of parameters.
32
33*/
34
35#include "TMVA/MCFitter.h"
36
37#include "TMVA/Configurable.h"
38#include "TMVA/FitterBase.h"
39#include "TMVA/GeneticRange.h"
40#include "TMVA/Interval.h"
41#include "TMVA/MsgLogger.h"
42#include "TMVA/Timer.h"
43#include "TMVA/Types.h"
44#include "TRandom3.h"
45
47
48////////////////////////////////////////////////////////////////////////////////
49/// constructor
50
52 const TString& name,
53 const std::vector<Interval*>& ranges,
54 const TString& theOption )
55: TMVA::FitterBase( target, name, ranges, theOption ),
56 fSamples( 0 ),
57 fSigma ( 1 ),
58 fSeed ( 0 )
59{
62}
63
64////////////////////////////////////////////////////////////////////////////////
65/// Declare MCFitter options
66
68{
69 DeclareOptionRef( fSamples = 100000, "SampleSize", "Number of Monte Carlo events in toy sample" );
70 DeclareOptionRef( fSigma = -1.0, "Sigma",
71 "If > 0: new points are generated according to Gauss around best value and with \"Sigma\" in units of interval length" );
72 DeclareOptionRef( fSeed = 100, "Seed", "Seed for the random generator (0 takes random seeds)" );
73}
74
75////////////////////////////////////////////////////////////////////////////////
76/// set MC fitter configuration parameters
77
79{
80 fSamples = samples;
81}
82
83////////////////////////////////////////////////////////////////////////////////
84/// Execute fitting
85
86Double_t TMVA::MCFitter::Run( std::vector<Double_t>& pars )
87{
88 Log() << kHEADER << "<MCFitter> Sampling, please be patient ..." << Endl;
89
90 // sanity check
91 if ((Int_t)pars.size() != GetNpars())
92 Log() << kFATAL << "<Run> Mismatch in number of parameters: "
93 << GetNpars() << " != " << pars.size() << Endl;
94
95 // timing of MC
96 Timer timer( fSamples, GetName() );
97 if (fIPyMaxIter) *fIPyMaxIter = fSamples;
98
99 std::vector<Double_t> parameters;
100 std::vector<Double_t> bestParameters;
101
102 TRandom3*rnd = new TRandom3( fSeed );
103 rnd->Uniform(0.,1.);
104
105 std::vector<TMVA::GeneticRange*> rndRanges;
106
107 // initial parameters (given by argument) are ignored
108 std::vector< TMVA::Interval* >::const_iterator rIt;
109 Double_t val;
110 for (rIt = fRanges.begin(); rIt<fRanges.end(); ++rIt) {
111 rndRanges.push_back( new TMVA::GeneticRange( rnd, (*rIt) ) );
112 val = rndRanges.back()->Random();
113 parameters.push_back( val );
114 bestParameters.push_back( val );
115 }
116
117 std::vector<Double_t>::iterator parIt;
118 std::vector<Double_t>::iterator parBestIt;
119
120 Double_t estimator = 0;
121 Double_t bestFit = 0;
122
123 // loop over all MC samples
124 for (Int_t sample = 0; sample < fSamples; sample++) {
125 if (fIPyCurrentIter) *fIPyCurrentIter = sample;
126 if (fExitFromTraining && *fExitFromTraining) break;
127
128 // dice the parameters
129 parIt = parameters.begin();
130 if (fSigma > 0.0) {
131 parBestIt = bestParameters.begin();
132 for (std::vector<TMVA::GeneticRange*>::iterator rndIt = rndRanges.begin(); rndIt<rndRanges.end(); ++rndIt) {
133 (*parIt) = (*rndIt)->Random( kTRUE, (*parBestIt), fSigma );
134 ++parIt;
135 ++parBestIt;
136 }
137 }
138 else {
139 for (std::vector<TMVA::GeneticRange*>::iterator rndIt = rndRanges.begin(); rndIt<rndRanges.end(); ++rndIt) {
140 (*parIt) = (*rndIt)->Random();
141 ++parIt;
142 }
143 }
144
145 // test the estimator value for the parameters
146 estimator = EstimatorFunction( parameters );
147
148 // if the estimator ist better (=smaller), take the new parameters as the best ones
149 if (estimator < bestFit || sample==0) {
150 bestFit = estimator;
151 bestParameters.swap( parameters );
152 }
153
154 // whats the time please?
155 if ((fSamples<100) || sample%Int_t(fSamples/100.0) == 0) timer.DrawProgressBar( sample );
156 }
157 pars.swap( bestParameters ); // return best parameters found
158
159 // get elapsed time
160 Log() << kINFO << "Elapsed time: " << timer.GetElapsedTime()
161 << " " << Endl;
162
163 return bestFit;
164}
int Int_t
Definition: RtypesCore.h:45
double Double_t
Definition: RtypesCore.h:59
const Bool_t kTRUE
Definition: RtypesCore.h:100
#define ClassImp(name)
Definition: Rtypes.h:364
char name[80]
Definition: TGX11.cxx:110
virtual void ParseOptions()
options parser
Base class for TMVA fitters.
Definition: FitterBase.h:51
Double_t Run()
estimator function interface for fitting
Definition: FitterBase.cxx:74
Range definition for genetic algorithm.
Definition: GeneticRange.h:42
Interface for a fitter 'target'.
Definition: IFitterTarget.h:44
Fitter using Monte Carlo sampling of parameters.
Definition: MCFitter.h:44
void SetParameters(Int_t cycles)
set MC fitter configuration parameters
Definition: MCFitter.cxx:78
MCFitter(IFitterTarget &target, const TString &name, const std::vector< TMVA::Interval * > &ranges, const TString &theOption)
constructor
Definition: MCFitter.cxx:51
void DeclareOptions()
Declare MCFitter options.
Definition: MCFitter.cxx:67
Timing information for training and evaluation of MVA methods.
Definition: Timer.h:58
TString GetElapsedTime(Bool_t Scientific=kTRUE)
returns pretty string with elapsed time
Definition: Timer.cxx:146
void DrawProgressBar(Int_t, const TString &comment="")
draws progress bar in color or B&W caution:
Definition: Timer.cxx:202
@ kHEADER
Definition: Types.h:63
@ kINFO
Definition: Types.h:58
@ kFATAL
Definition: Types.h:61
Random number generator class based on M.
Definition: TRandom3.h:27
virtual Double_t Uniform(Double_t x1=1)
Returns a uniform deviate on the interval (0, x1).
Definition: TRandom.cxx:672
Basic string class.
Definition: TString.h:136
create variable transformations
MsgLogger & Endl(MsgLogger &ml)
Definition: MsgLogger.h:148
Double_t Log(Double_t x)
Definition: TMath.h:760