Logo ROOT   6.14/05
Reference Guide
testUnfold5d.C
Go to the documentation of this file.
1 /// \file
2 /// \ingroup tutorial_unfold
3 /// \notebook
4 /// Test program for the classes TUnfoldDensity and TUnfoldBinning.
5 ///
6 /// A toy test of the TUnfold package
7 ///
8 /// This is an example of unfolding a two-dimensional distribution
9 /// also using an auxiliary measurement to constrain some background
10 ///
11 /// The example comprises several macros
12 /// testUnfold5a.C create root files with TTree objects for
13 /// signal, background and data
14 /// -> write files testUnfold5_signal.root
15 /// testUnfold5_background.root
16 /// testUnfold5_data.root
17 ///
18 /// testUnfold5b.C create a root file with the TUnfoldBinning objects
19 /// -> write file testUnfold5_binning.root
20 ///
21 /// testUnfold5c.C loop over trees and fill histograms based on the
22 /// TUnfoldBinning objects
23 /// -> read testUnfold5_binning.root
24 /// testUnfold5_signal.root
25 /// testUnfold5_background.root
26 /// testUnfold5_data.root
27 ///
28 /// -> write testUnfold5_histograms.root
29 ///
30 /// testUnfold5d.C run the unfolding
31 /// -> read testUnfold5_histograms.root
32 /// -> write testUnfold5_result.root
33 /// testUnfold5_result.ps
34 ///
35 /// \macro_output
36 /// \macro_code
37 ///
38 /// **Version 17.6, in parallel to changes in TUnfold**
39 ///
40 /// #### History:
41 /// - Version 17.5, in parallel to changes in TUnfold
42 /// - Version 17.4, in parallel to changes in TUnfold
43 /// - Version 17.3, in parallel to changes in TUnfold
44 /// - Version 17.2, in parallel to changes in TUnfold
45 /// - Version 17.1, in parallel to changes in TUnfold
46 /// - Version 17.0 example for multi-dimensional unfolding
47 ///
48 /// This file is part of TUnfold.
49 ///
50 /// TUnfold is free software: you can redistribute it and/or modify
51 /// it under the terms of the GNU General Public License as published by
52 /// the Free Software Foundation, either version 3 of the License, or
53 /// (at your option) any later version.
54 ///
55 /// TUnfold is distributed in the hope that it will be useful,
56 /// but WITHOUT ANY WARRANTY; without even the implied warranty of
57 /// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
58 /// GNU General Public License for more details.
59 ///
60 /// You should have received a copy of the GNU General Public License
61 /// along with TUnfold. If not, see <http://www.gnu.org/licenses/>.
62 ///
63 /// \author Stefan Schmitt DESY, 14.10.2008
64 
65 #include <iostream>
66 #include <cmath>
67 #include <map>
68 #include <TMath.h>
69 #include <TCanvas.h>
70 #include <TStyle.h>
71 #include <TGraph.h>
72 #include <TFile.h>
73 #include <TH1.h>
74 #include "TUnfoldDensity.h"
75 
76 using namespace std;
77 
78 // #define PRINT_MATRIX_L
79 
80 #define TEST_INPUT_COVARIANCE
81 
82 void testUnfold5d()
83 {
84  // switch on histogram errors
86 
87  //==============================================
88  // step 1 : open output file
89  TFile *outputFile=new TFile("testUnfold5_results.root","recreate");
90 
91  //==============================================
92  // step 2 : read binning schemes and input histograms
93  TFile *inputFile=new TFile("testUnfold5_histograms.root");
94 
95  outputFile->cd();
96 
97  TUnfoldBinning *detectorBinning,*generatorBinning;
98 
99  inputFile->GetObject("detector",detectorBinning);
100  inputFile->GetObject("generator",generatorBinning);
101 
102  if((!detectorBinning)||(!generatorBinning)) {
103  cout<<"problem to read binning schemes\n";
104  }
105 
106  // save binning schemes to output file
107  detectorBinning->Write();
108  generatorBinning->Write();
109 
110  // read histograms
111  TH1 *histDataReco,*histDataTruth;
112  TH2 *histMCGenRec;
113 
114  inputFile->GetObject("histDataReco",histDataReco);
115  inputFile->GetObject("histDataTruth",histDataTruth);
116  inputFile->GetObject("histMCGenRec",histMCGenRec);
117 
118 #ifdef TEST_ZERO_UNCORR_ERROR
119  // special test (bug in version 17.2 and below)
120  // set all errors in hisMCGenRec to zero
121  // -> program will crash
122  for(int i=0;i<=histMCGenRec->GetNbinsX()+1;i++) {
123  for(int j=0;j<=histMCGenRec->GetNbinsY()+1;j++) {
124  histMCGenRec->SetBinError(i,j,0.0);
125  }
126  }
127 #endif
128 
129  histDataReco->Write();
130  histDataTruth->Write();
131  histMCGenRec->Write();
132 
133  if((!histDataReco)||(!histDataTruth)||(!histMCGenRec)) {
134  cout<<"problem to read input histograms\n";
135  }
136 
137  //========================
138  // Step 3: unfolding
139 
140  // preserve the area
142 
143  // basic choice of regularisation scheme:
144  // curvature (second derivative)
146 
147  // density flags
148  TUnfoldDensity::EDensityMode densityFlags=
150 
151  // detailed steering for regularisation
152  const char *REGULARISATION_DISTRIBUTION=0;
153  const char *REGULARISATION_AXISSTEERING="*[B]";
154 
155  // set up matrix of migrations
156  TUnfoldDensity unfold(histMCGenRec,TUnfold::kHistMapOutputHoriz,
157  regMode,constraintMode,densityFlags,
158  generatorBinning,detectorBinning,
159  REGULARISATION_DISTRIBUTION,
160  REGULARISATION_AXISSTEERING);
161 
162  // define the input vector (the measured data distribution)
163 
164 #ifdef TEST_INPUT_COVARIANCE
165  // special test to use input covariance matrix
166  TH2D *inputEmatrix=
167  detectorBinning->CreateErrorMatrixHistogram("input_covar",true);
168  for(int i=1;i<=inputEmatrix->GetNbinsX();i++) {
169  Double_t e=histDataReco->GetBinError(i);
170  inputEmatrix->SetBinContent(i,i,e*e);
171  // test: non-zero covariance where variance is zero
172  // if(e<=0.) inputEmatrix->SetBinContent(i,i+1,1.0);
173  }
174  unfold.SetInput(histDataReco,0.0,0.0,inputEmatrix);
175 #else
176  unfold.SetInput(histDataReco /* ,0.0,1.0 */);
177 #endif
178  // print matrix of regularisation conditions
179 #ifdef PRINT_MATRIX_L
180  TH2 *histL= unfold.GetL("L");
181  for(Int_t j=1;j<=histL->GetNbinsY();j++) {
182  cout<<"L["<<unfold.GetLBinning()->GetBinName(j)<<"]";
183  for(Int_t i=1;i<=histL->GetNbinsX();i++) {
184  Double_t c=histL->GetBinContent(i,j);
185  if(c!=0.0) cout<<" ["<<i<<"]="<<c;
186  }
187  cout<<"\n";
188  }
189 #endif
190  // run the unfolding
191  //
192  // here, tau is determined by scanning the global correlation coefficients
193 
194  Int_t nScan=30;
195  TSpline *rhoLogTau=0;
196  TGraph *lCurve=0;
197 
198  // for determining tau, scan the correlation coefficients
199  // correlation coefficients may be probed for all distributions
200  // or only for selected distributions
201  // underflow/overflow bins may be included/excluded
202  //
203  const char *SCAN_DISTRIBUTION="signal";
204  const char *SCAN_AXISSTEERING=0;
205 
206  Int_t iBest=unfold.ScanTau(nScan,0.,0.,&rhoLogTau,
208  SCAN_DISTRIBUTION,SCAN_AXISSTEERING,
209  &lCurve);
210 
211  // create graphs with one point to visualize best choice of tau
212  Double_t t[1],rho[1],x[1],y[1];
213  rhoLogTau->GetKnot(iBest,t[0],rho[0]);
214  lCurve->GetPoint(iBest,x[0],y[0]);
215  TGraph *bestRhoLogTau=new TGraph(1,t,rho);
216  TGraph *bestLCurve=new TGraph(1,x,y);
217  Double_t *tAll=new Double_t[nScan],*rhoAll=new Double_t[nScan];
218  for(Int_t i=0;i<nScan;i++) {
219  rhoLogTau->GetKnot(i,tAll[i],rhoAll[i]);
220  }
221  TGraph *knots=new TGraph(nScan,tAll,rhoAll);
222 
223  cout<<"chi**2="<<unfold.GetChi2A()<<"+"<<unfold.GetChi2L()
224  <<" / "<<unfold.GetNdf()<<"\n";
225 
226 
227  //===========================
228  // Step 4: retrieve and plot unfolding results
229 
230  // get unfolding output
231  TH1 *histDataUnfold=unfold.GetOutput("unfolded signal",0,0,0,kFALSE);
232  // get Monte Carlo reconstructed data
233  TH1 *histMCReco=histMCGenRec->ProjectionY("histMCReco",0,-1,"e");
234  TH1 *histMCTruth=histMCGenRec->ProjectionX("histMCTruth",0,-1,"e");
235  Double_t scaleFactor=histDataTruth->GetSumOfWeights()/
236  histMCTruth->GetSumOfWeights();
237  histMCReco->Scale(scaleFactor);
238  histMCTruth->Scale(scaleFactor);
239  // get matrix of probabilities
240  TH2 *histProbability=unfold.GetProbabilityMatrix("histProbability");
241  // get global correlation coefficients
242  /* TH1 *histGlobalCorr=*/ unfold.GetRhoItotal("histGlobalCorr",0,0,0,kFALSE);
243  TH1 *histGlobalCorrScan=unfold.GetRhoItotal
244  ("histGlobalCorrScan",0,SCAN_DISTRIBUTION,SCAN_AXISSTEERING,kFALSE);
245  /* TH2 *histCorrCoeff=*/ unfold.GetRhoIJtotal("histCorrCoeff",0,0,0,kFALSE);
246 
247  TCanvas canvas;
248  canvas.Print("testUnfold5.ps[");
249 
250  //========== page 1 ============
251  // unfolding control plots
252  // input, matrix, output
253  // tau-scan, global correlations, correlation coefficients
254  canvas.Clear();
255  canvas.Divide(3,2);
256 
257  // (1) all bins, compare to original MC distribution
258  canvas.cd(1);
259  histDataReco->SetMinimum(0.0);
260  histDataReco->Draw("E");
261  histMCReco->SetLineColor(kBlue);
262  histMCReco->Draw("SAME HIST");
263  // (2) matrix of probabilities
264  canvas.cd(2);
265  histProbability->Draw("BOX");
266  // (3) unfolded data, data truth, MC truth
267  canvas.cd(3);
268  gPad->SetLogy();
269  histDataUnfold->Draw("E");
270  histDataTruth->SetLineColor(kBlue);
271  histDataTruth->Draw("SAME HIST");
272  histMCTruth->SetLineColor(kRed);
273  histMCTruth->Draw("SAME HIST");
274  // (4) scan of correlation vs tau
275  canvas.cd(4);
276  rhoLogTau->Draw();
277  knots->Draw("*");
278  bestRhoLogTau->SetMarkerColor(kRed);
279  bestRhoLogTau->Draw("*");
280  // (5) global correlation coefficients for the distributions
281  // used during the scan
282  canvas.cd(5);
283  //histCorrCoeff->Draw("BOX");
284  histGlobalCorrScan->Draw("HIST");
285  // (6) L-curve
286  canvas.cd(6);
287  lCurve->Draw("AL");
288  bestLCurve->SetMarkerColor(kRed);
289  bestLCurve->Draw("*");
290 
291 
292  canvas.Print("testUnfold5.ps");
293 
294  canvas.Print("testUnfold5.ps]");
295 
296 }
virtual Int_t Write(const char *name=0, Int_t option=0, Int_t bufsize=0)
Write this object to the current directory.
Definition: TObject.cxx:785
virtual void Scale(Double_t c1=1, Option_t *option="")
Multiply this histogram by a constant c1.
Definition: TH1.cxx:6101
An algorithm to unfold distributions from detector to truth level.
TH2D * CreateErrorMatrixHistogram(const char *histogramName, Bool_t originalAxisBinning, Int_t **binMap=0, const char *histogramTitle=0, const char *axisSteering=0) const
Create a TH2D histogram capable to hold a covariance matrix.
Definition: Rtypes.h:59
TH1D * ProjectionY(const char *name="_py", Int_t firstxbin=0, Int_t lastxbin=-1, Option_t *option="") const
Project a 2-D histogram into a 1-D histogram along Y.
Definition: TH2.cxx:2357
Base class for spline implementation containing the Draw/Paint methods.
Definition: TSpline.h:20
virtual Double_t GetSumOfWeights() const
Return the sum of weights excluding under/overflows.
Definition: TH1.cxx:7269
TVirtualPad * cd(Int_t subpadnumber=0)
Set current canvas & pad.
Definition: TCanvas.cxx:688
TH1D * ProjectionX(const char *name="_px", Int_t firstybin=0, Int_t lastybin=-1, Option_t *option="") const
Project a 2-D histogram into a 1-D histogram along X.
Definition: TH2.cxx:2317
virtual void SetMinimum(Double_t minimum=-1111)
Definition: TH1.h:391
int Int_t
Definition: RtypesCore.h:41
STL namespace.
EDensityMode
choice of regularisation scale factors to cinstruct the matrix L
virtual void Draw(Option_t *chopt="")
Draw this graph with its current attributes.
Definition: TGraph.cxx:745
static void SetDefaultSumw2(Bool_t sumw2=kTRUE)
When this static function is called with sumw2=kTRUE, all new histograms will automatically activate ...
Definition: TH1.cxx:6177
Double_t x[n]
Definition: legend1.C:17
truth level on x-axis of the response matrix
Definition: TUnfold.h:143
scale factors from multidimensional bin width
EConstraint
type of extra constraint
Definition: TUnfold.h:110
virtual void SetMarkerColor(Color_t mcolor=1)
Set the marker color.
Definition: TAttMarker.h:38
void Clear(Option_t *option="")
Remove all primitives from the canvas.
Definition: TCanvas.cxx:707
maximum global correlation coefficient (from TUnfold::GetRhoI())
virtual void SetBinError(Int_t bin, Double_t error)
Set the bin Error Note that this resets the bin eror option to be of Normal Type and for the non-empt...
Definition: TH1.cxx:8498
ERegMode
choice of regularisation scheme
Definition: TUnfold.h:120
virtual void SetLineColor(Color_t lcolor)
Set the line color.
Definition: TAttLine.h:40
virtual Int_t GetPoint(Int_t i, Double_t &x, Double_t &y) const
Get x and y values for point number i.
Definition: TGraph.cxx:1580
Service class for 2-Dim histogram classes.
Definition: TH2.h:30
virtual void Draw(Option_t *option="")
Draw this histogram with options.
Definition: TH1.cxx:2974
virtual void Print(const char *filename="") const
Save Pad contents in a file in one of various formats.
Definition: TPad.cxx:4617
enforce preservation of the area
Definition: TUnfold.h:116
const Bool_t kFALSE
Definition: RtypesCore.h:88
virtual void GetKnot(Int_t i, Double_t &x, Double_t &y) const =0
Binning schemes for use with the unfolding algorithm TUnfoldDensity.
The Canvas class.
Definition: TCanvas.h:31
double Double_t
Definition: RtypesCore.h:55
virtual void Draw(Option_t *option="")
Draw this function with its current attributes.
Definition: TSpline.cxx:97
Double_t y[n]
Definition: legend1.C:17
The TH1 histogram class.
Definition: TH1.h:56
you should not use this method at all Int_t Int_t Double_t Double_t Double_t e
Definition: TRolke.cxx:630
virtual void Divide(Int_t nx=1, Int_t ny=1, Float_t xmargin=0.01, Float_t ymargin=0.01, Int_t color=0)
Automatic pad generation by division.
Definition: TPad.cxx:1162
A Graph is a graphics object made of two arrays X and Y with npoints each.
Definition: TGraph.h:41
virtual Double_t GetBinContent(Int_t bin) const
Return content of bin number bin.
Definition: TH2.h:84
#define gPad
Definition: TVirtualPad.h:285
#define c(i)
Definition: RSha256.hxx:101
Definition: Rtypes.h:59
virtual void SetBinContent(Int_t bin, Double_t content)
Set bin content.
Definition: TH2.cxx:2500
virtual Int_t GetNbinsX() const
Definition: TH1.h:291
regularize the 2nd derivative of the output distribution
Definition: TUnfold.h:132
virtual Int_t GetNbinsY() const
Definition: TH1.h:292
virtual Double_t GetBinError(Int_t bin) const
Return value of error associated to bin number bin.
Definition: TH1.cxx:8356
2-D histogram with a double per channel (see TH1 documentation)}
Definition: TH2.h:291