EIC code displayed by LXR |
|
|
|
|||
File indexing completed on 2025-01-18 10:12:28
0001 // Author: Kerstin Tackmann, Andreas Hoecker, Heiko Lacker 0002 0003 /********************************************************************************** 0004 * * 0005 * Project: TSVDUnfold - data unfolding based on Singular Value Decomposition * 0006 * Package: ROOT * 0007 * Class : TSVDUnfold * 0008 * * 0009 * Description: * 0010 * Single class implementation of SVD data unfolding based on: * 0011 * A. Hoecker, V. Kartvelishvili, * 0012 * "SVD approach to data unfolding" * 0013 * NIM A372, 469 (1996) [hep-ph/9509307] * 0014 * * 0015 * Authors: * 0016 * Kerstin Tackmann <Kerstin.Tackmann@cern.ch> - CERN, Switzerland * 0017 * Andreas Hoecker <Andreas.Hoecker@cern.ch> - CERN, Switzerland * 0018 * Heiko Lacker <lacker@physik.hu-berlin.de> - Humboldt U, Germany * 0019 * * 0020 * Copyright (c) 2010: * 0021 * CERN, Switzerland * 0022 * Humboldt University, Germany * 0023 * * 0024 **********************************************************************************/ 0025 0026 ////////////////////////////////////////////////////////////////////////// 0027 // // 0028 // TSVDUnfold // 0029 // // 0030 // Data unfolding using Singular Value Decomposition (hep-ph/9509307) // 0031 // Authors: Kerstin Tackmann, Andreas Hoecker, Heiko Lacker // 0032 // // 0033 ////////////////////////////////////////////////////////////////////////// 0034 0035 #ifndef TSVDUNFOLD_H 0036 #define TSVDUNFOLD_H 0037 0038 #include "TObject.h" 0039 #include "TMatrixD.h" 0040 #include "TVectorD.h" 0041 #include "TMatrixDSym.h" 0042 0043 class TH1D; 0044 class TH2D; 0045 0046 class TSVDUnfold : public TObject { 0047 0048 public: 0049 0050 // Constructor 0051 // Initialisation of unfolding 0052 // "bdat" - measured data distribution (number of events) 0053 // "Bcov" - covariance matrix for measured data distribution 0054 // "bini" - reconstructed MC distribution (number of events) 0055 // "xini" - truth MC distribution (number of events) 0056 // "Adet" - detector response matrix (number of events) 0057 TSVDUnfold( const TH1D* bdat, const TH1D* bini, const TH1D* xini, const TH2D* Adet ); 0058 TSVDUnfold( const TH1D* bdat, TH2D* Bcov, const TH1D* bini, const TH1D* xini, const TH2D* Adet ); 0059 TSVDUnfold( const TSVDUnfold& other ); 0060 0061 // Destructor 0062 ~TSVDUnfold() override; 0063 0064 // Set option to normalize unfolded spectrum to unit area 0065 // "normalize" - switch 0066 void SetNormalize ( Bool_t normalize ) { fNormalize = normalize; } 0067 0068 // Do the unfolding 0069 // "kreg" - number of singular values used (regularisation) 0070 TH1D* Unfold ( Int_t kreg ); 0071 0072 // Determine for given input error matrix covariance matrix of unfolded 0073 // spectrum from toy simulation 0074 // "cov" - covariance matrix on the measured spectrum, to be propagated 0075 // "ntoys" - number of pseudo experiments used for the propagation 0076 // "seed" - seed for pseudo experiments 0077 TH2D* GetUnfoldCovMatrix( const TH2D* cov, Int_t ntoys, Int_t seed = 1 ); 0078 0079 // Determine covariance matrix of unfolded spectrum from finite statistics in 0080 // response matrix 0081 // "ntoys" - number of pseudo experiments used for the propagation 0082 // "seed" - seed for pseudo experiments 0083 TH2D* GetAdetCovMatrix( Int_t ntoys, Int_t seed=1 ); 0084 0085 // Regularisation parameter 0086 Int_t GetKReg() const { return fKReg; } 0087 0088 // Obtain the distribution of |d| (for determining the regularization) 0089 TH1D* GetD() const; 0090 0091 // Obtain the distribution of singular values 0092 TH1D* GetSV() const; 0093 0094 // Obtain the computed regularized covariance matrix 0095 TH2D* GetXtau() const; 0096 0097 // Obtain the computed inverse of the covariance matrix 0098 TH2D* GetXinv() const; 0099 0100 //Obtain the covariance matrix on the data 0101 TH2D* GetBCov() const; 0102 0103 // Helper functions 0104 Double_t ComputeChiSquared( const TH1D& truspec, const TH1D& unfspec ); 0105 0106 private: 0107 0108 // Helper functions for vector and matrix operations 0109 void FillCurvatureMatrix( TMatrixD& tCurv, TMatrixD& tC ) const; 0110 static Double_t GetCurvature ( const TVectorD& vec, const TMatrixD& curv ); 0111 0112 void InitHistos ( ); 0113 0114 // Helper functions 0115 static void H2V ( const TH1D* histo, TVectorD& vec ); 0116 static void H2Verr ( const TH1D* histo, TVectorD& vec ); 0117 static void V2H ( const TVectorD& vec, TH1D& histo ); 0118 static void H2M ( const TH2D* histo, TMatrixD& mat ); 0119 static void M2H ( const TMatrixD& mat, TH2D& histo ); 0120 static TMatrixD MatDivVec( const TMatrixD& mat, const TVectorD& vec, Int_t zero=0 ); 0121 static TVectorD CompProd ( const TVectorD& vec1, const TVectorD& vec2 ); 0122 0123 static TVectorD VecDiv ( const TVectorD& vec1, const TVectorD& vec2, Int_t zero = 0 ); 0124 static void RegularisedSymMatInvert( TMatrixDSym& mat, Double_t eps = 1e-3 ); 0125 0126 /// @name Class members 0127 ///@{ 0128 Int_t fNdim; ///<! Truth and reconstructed dimensions 0129 Int_t fDdim; ///<! Derivative for curvature matrix 0130 Bool_t fNormalize; ///<! Normalize unfolded spectrum to 1 0131 Int_t fKReg; ///<! Regularisation parameter 0132 TH1D* fDHist; ///<! Distribution of d (for checking regularization) 0133 TH1D* fSVHist; ///<! Distribution of singular values 0134 TH2D* fXtau; ///<! Computed regularized covariance matrix 0135 TH2D* fXinv; ///<! Computed inverse of covariance matrix 0136 ///@} 0137 0138 /// @name Input histos 0139 ///@{ 0140 const TH1D* fBdat; ///< Measured distribution (data) 0141 TH2D* fBcov; ///< Covariance matrix of measured distribution (data) 0142 const TH1D* fBini; ///< Reconstructed distribution (MC) 0143 const TH1D* fXini; ///< Truth distribution (MC) 0144 const TH2D* fAdet; ///< Detector response matrix 0145 ///@} 0146 0147 /// @name Evaluation of covariance matrices 0148 ///@{ 0149 TH1D* fToyhisto; ///<! Toy MC histogram 0150 TH2D* fToymat; ///<! Toy MC detector response matrix 0151 Bool_t fToyMode; ///<! Internal switch for covariance matrix propagation 0152 Bool_t fMatToyMode; ///<! Internal switch for evaluation of statistical uncertainties from response matrix 0153 ///@} 0154 0155 0156 ClassDefOverride( TSVDUnfold, 0 ) // Data unfolding using Singular Value Decomposition (hep-ph/9509307) 0157 }; 0158 0159 #endif
| [ Source navigation ] | [ Diff markup ] | [ Identifier search ] | [ general search ] |
|
This page was automatically generated by the 2.3.7 LXR engine. The LXR team |
|