Logo ROOT   6.14/05
Reference Guide
rs401d_FeldmanCousins.C
Go to the documentation of this file.
1 /// \file
2 /// \ingroup tutorial_roostats
3 /// \notebook
4 /// 'Neutrino Oscillation Example from Feldman & Cousins'
5 ///
6 /// This tutorial shows a more complex example using the FeldmanCousins utility
7 /// to create a confidence interval for a toy neutrino oscillation experiment.
8 /// The example attempts to faithfully reproduce the toy example described in Feldman & Cousins'
9 /// original paper, Phys.Rev.D57:3873-3889,1998.
10 ///
11 /// \macro_image
12 /// \macro_output
13 /// \macro_code
14 ///
15 /// \author Kyle Cranmer
16 
17 #include "RooGlobalFunc.h"
18 #include "RooStats/ConfInterval.h"
25 #include "RooStats/MCMCInterval.h"
26 
27 #include "RooDataSet.h"
28 #include "RooDataHist.h"
29 #include "RooRealVar.h"
30 #include "RooConstVar.h"
31 #include "RooAddition.h"
32 #include "RooProduct.h"
33 #include "RooProdPdf.h"
34 #include "RooAddPdf.h"
35 
36 #include "TROOT.h"
37 #include "RooPolynomial.h"
38 #include "RooRandom.h"
39 
40 #include "RooNLLVar.h"
41 #include "RooProfileLL.h"
42 
43 #include "RooPlot.h"
44 
45 #include "TCanvas.h"
46 #include "TH1F.h"
47 #include "TH2F.h"
48 #include "TTree.h"
49 #include "TMarker.h"
50 #include "TStopwatch.h"
51 
52 #include <iostream>
53 
54 // PDF class created for this macro
55 #if !defined(__CINT__) || defined(__MAKECINT__)
56 #include "../tutorials/roostats/NuMuToNuE_Oscillation.h"
57 #include "../tutorials/roostats/NuMuToNuE_Oscillation.cxx" // so that it can be executed directly
58 #else
59 #include "../tutorials/roostats/NuMuToNuE_Oscillation.cxx+" // so that it can be executed directly
60 #endif
61 
62 // use this order for safety on library loading
63 using namespace RooFit;
64 using namespace RooStats ;
65 
66 
67 void rs401d_FeldmanCousins(bool doFeldmanCousins=false, bool doMCMC = true)
68 {
69 
70  // to time the macro
71  TStopwatch t;
72  t.Start();
73 
74 
75 
76  // Taken from Feldman & Cousins paper, Phys.Rev.D57:3873-3889,1998.
77  // e-Print: physics/9711021 (see page 13.)
78  //
79  // Quantum mechanics dictates that the probability of such a transformation is given by the formula
80  // $P (\nu\mu \rightarrow \nu e ) = sin^2 (2\theta) sin^2 (1.27 \Delta m^2 L /E )$
81  // where P is the probability for a $\nu\mu$ to transform into a $\nu e$ , L is the distance in km between
82  // the creation of the neutrino from meson decay and its interaction in the detector, E is the
83  // neutrino energy in GeV, and $\Delta m^2 = |m^2 - m^2 |$ in $(eV/c^2 )^2$ .
84  //
85  // To demonstrate how this works in practice, and how it compares to alternative approaches
86  // that have been used, we consider a toy model of a typical neutrino oscillation experiment.
87  // The toy model is defined by the following parameters: Mesons are assumed to decay to
88  // neutrinos uniformly in a region 600 m to 1000 m from the detector. The expected background
89  // from conventional $\nu e$ interactions and misidentified $\nu\mu$ interactions is assumed to be 100
90  // events in each of 5 energy bins which span the region from 10 to 60 GeV. We assume that
91  // the $\nu\mu$ flux is such that if $P (\nu\mu \rightarrow \nu e ) = 0.01$ averaged over any bin, then that bin would
92  // have an expected additional contribution of 100 events due to $\nu\mu \rightarrow \nu e$ oscillations.
93 
94 
95  // Make signal model model
96  RooRealVar E("E","", 15,10,60,"GeV");
97  RooRealVar L("L","", .800,.600, 1.0,"km"); // need these units in formula
98  RooRealVar deltaMSq("deltaMSq","#Delta m^{2}",40,1,300,"eV/c^{2}");
99  RooRealVar sinSq2theta("sinSq2theta","sin^{2}(2#theta)", .006,.0,.02);
100  //RooRealVar deltaMSq("deltaMSq","#Delta m^{2}",40,20,70,"eV/c^{2}");
101  // RooRealVar sinSq2theta("sinSq2theta","sin^{2}(2#theta)", .006,.001,.01);
102  // PDF for oscillation only describes deltaMSq dependence, sinSq2theta goes into sigNorm
103  // 1) The code for this PDF was created by issuing these commands
104  // root [0] RooClassFactory x
105  // root [1] x.makePdf("NuMuToNuE_Oscillation","L,E,deltaMSq","","pow(sin(1.27*deltaMSq*L/E),2)")
106  NuMuToNuE_Oscillation PnmuTone("PnmuTone","P(#nu_{#mu} #rightarrow #nu_{e}",L,E,deltaMSq);
107 
108  // only E is observable, so create the signal model by integrating out L
109  RooAbsPdf* sigModel = PnmuTone.createProjection(L);
110 
111  // create $ \int dE' dL' P(E',L' | \Delta m^2)$.
112  // Given RooFit will renormalize the PDF in the range of the observables,
113  // the average probability to oscillate in the experiment's acceptance
114  // needs to be incorporated into the extended term in the likelihood.
115  // Do this by creating a RooAbsReal representing the integral and divide by
116  // the area in the E-L plane.
117  // The integral should be over "primed" observables, so we need
118  // an independent copy of PnmuTone not to interfere with the original.
119 
120  // Independent copy for Integral
121  RooRealVar EPrime("EPrime","", 15,10,60,"GeV");
122  RooRealVar LPrime("LPrime","", .800,.600, 1.0,"km"); // need these units in formula
123  NuMuToNuE_Oscillation PnmuTonePrime("PnmuTonePrime","P(#nu_{#mu} #rightarrow #nu_{e}",
124  LPrime,EPrime,deltaMSq);
125  RooAbsReal* intProbToOscInExp = PnmuTonePrime.createIntegral(RooArgSet(EPrime,LPrime));
126 
127  // Getting the flux is a bit tricky. It is more clear to include a cross section term that is not
128  // explicitly referred to in the text, eg.
129  // number events in bin = flux * cross-section for nu_e interaction in E bin * average prob nu_mu osc. to nu_e in bin
130  // let maxEventsInBin = flux * cross-section for nu_e interaction in E bin
131  // maxEventsInBin * 1% chance per bin = 100 events / bin
132  // therefore maxEventsInBin = 10,000.
133  // for 5 bins, this means maxEventsTot = 50,000
134  RooConstVar maxEventsTot("maxEventsTot","maximum number of sinal events",50000);
135  RooConstVar inverseArea("inverseArea","1/(#Delta E #Delta L)",
136  1./(EPrime.getMax()-EPrime.getMin())/(LPrime.getMax()-LPrime.getMin()));
137 
138  // $sigNorm = maxEventsTot \cdot \int dE dL \frac{P_{oscillate\ in\ experiment}}{Area} \cdot {sin}^2(2\theta)$
139  RooProduct sigNorm("sigNorm", "", RooArgSet(maxEventsTot, *intProbToOscInExp, inverseArea, sinSq2theta));
140  // bkg = 5 bins * 100 events / bin
141  RooConstVar bkgNorm("bkgNorm","normalization for background",500);
142 
143  // flat background (0th order polynomial, so no arguments for coefficients)
144  RooPolynomial bkgEShape("bkgEShape","flat bkg shape", E);
145 
146  // total model
147  RooAddPdf model("model","",RooArgList(*sigModel,bkgEShape),
148  RooArgList(sigNorm,bkgNorm));
149 
150  // for debugging, check model tree
151  // model.printCompactTree();
152  // model.graphVizTree("model.dot");
153 
154 
155  // turn off some messages
159 
160 
161  // --------------------------------------
162  // n events in data to data, simply sum of sig+bkg
163  Int_t nEventsData = bkgNorm.getVal()+sigNorm.getVal();
164  cout << "generate toy data with nEvents = " << nEventsData << endl;
165  // adjust random seed to get a toy dataset similar to one in paper.
166  // Found by trial and error (3 trials, so not very "fine tuned")
168  // create a toy dataset
169  RooDataSet* data = model.generate(RooArgSet(E), nEventsData);
170 
171  // --------------------------------------
172  // make some plots
173  TCanvas* dataCanvas = new TCanvas("dataCanvas");
174  dataCanvas->Divide(2,2);
175 
176  // plot the PDF
177  dataCanvas->cd(1);
178  TH1* hh = PnmuTone.createHistogram("hh",E,Binning(40),YVar(L,Binning(40)),Scaling(kFALSE)) ;
179  hh->SetLineColor(kBlue) ;
180  hh->SetTitle("True Signal Model");
181  hh->Draw("surf");
182 
183  // plot the data with the best fit
184  dataCanvas->cd(2);
185  RooPlot* Eframe = E.frame();
186  data->plotOn(Eframe);
187  model.fitTo(*data, Extended());
188  model.plotOn(Eframe);
189  model.plotOn(Eframe,Components(*sigModel),LineColor(kRed));
190  model.plotOn(Eframe,Components(bkgEShape),LineColor(kGreen));
191  model.plotOn(Eframe);
192  Eframe->SetTitle("toy data with best fit model (and sig+bkg components)");
193  Eframe->Draw();
194 
195  // plot the likelihood function
196  dataCanvas->cd(3);
197  RooNLLVar nll("nll", "nll", model, *data, Extended());
198  RooProfileLL pll("pll", "", nll, RooArgSet(deltaMSq, sinSq2theta));
199  // TH1* hhh = nll.createHistogram("hhh",sinSq2theta,Binning(40),YVar(deltaMSq,Binning(40))) ;
200  TH1* hhh = pll.createHistogram("hhh",sinSq2theta,Binning(40),YVar(deltaMSq,Binning(40)),Scaling(kFALSE)) ;
201  hhh->SetLineColor(kBlue) ;
202  hhh->SetTitle("Likelihood Function");
203  hhh->Draw("surf");
204 
205  dataCanvas->Update();
206 
207 
208 
209  // --------------------------------------------------------------
210  // show use of Feldman-Cousins utility in RooStats
211  // set the distribution creator, which encodes the test statistic
212  RooArgSet parameters(deltaMSq, sinSq2theta);
213  RooWorkspace* w = new RooWorkspace();
214 
215  ModelConfig modelConfig;
216  modelConfig.SetWorkspace(*w);
217  modelConfig.SetPdf(model);
218  modelConfig.SetParametersOfInterest(parameters);
219 
220  RooStats::FeldmanCousins fc(*data, modelConfig);
221  fc.SetTestSize(.1); // set size of test
222  fc.UseAdaptiveSampling(true);
223  fc.SetNBins(10); // number of points to test per parameter
224 
225  // use the Feldman-Cousins tool
226  ConfInterval* interval = 0;
227  if(doFeldmanCousins)
228  interval = fc.GetInterval();
229 
230 
231  // ---------------------------------------------------------
232  // show use of ProfileLikeihoodCalculator utility in RooStats
233  RooStats::ProfileLikelihoodCalculator plc(*data, modelConfig);
234  plc.SetTestSize(.1);
235 
236  ConfInterval* plcInterval = plc.GetInterval();
237 
238  // --------------------------------------------
239  // show use of MCMCCalculator utility in RooStats
240  MCMCInterval* mcInt = NULL;
241 
242  if (doMCMC) {
243  // turn some messages back on
246 
247  TStopwatch mcmcWatch;
248  mcmcWatch.Start();
249 
250  RooArgList axisList(deltaMSq, sinSq2theta);
251  MCMCCalculator mc(*data, modelConfig);
252  mc.SetNumIters(5000);
253  mc.SetNumBurnInSteps(100);
254  mc.SetUseKeys(true);
255  mc.SetTestSize(.1);
256  mc.SetAxes(axisList); // set which is x and y axis in posterior histogram
257  //mc.SetNumBins(50);
258  mcInt = (MCMCInterval*)mc.GetInterval();
259 
260  mcmcWatch.Stop();
261  mcmcWatch.Print();
262  }
263  // -------------------------------
264  // make plot of resulting interval
265 
266  dataCanvas->cd(4);
267 
268  // first plot a small dot for every point tested
269  if (doFeldmanCousins) {
270  RooDataHist* parameterScan = (RooDataHist*) fc.GetPointsToScan();
271  TH2F* hist = (TH2F*) parameterScan->createHistogram("sinSq2theta:deltaMSq",30,30);
272  // hist->Draw();
273  TH2F* forContour = (TH2F*)hist->Clone();
274 
275  // now loop through the points and put a marker if it's in the interval
276  RooArgSet* tmpPoint;
277  // loop over points to test
278  for(Int_t i=0; i<parameterScan->numEntries(); ++i){
279  // get a parameter point from the list of points to test.
280  tmpPoint = (RooArgSet*) parameterScan->get(i)->clone("temp");
281 
282  if (interval){
283  if (interval->IsInInterval( *tmpPoint ) ) {
284  forContour->SetBinContent( hist->FindBin(tmpPoint->getRealValue("sinSq2theta"),
285  tmpPoint->getRealValue("deltaMSq")), 1);
286  }else{
287  forContour->SetBinContent( hist->FindBin(tmpPoint->getRealValue("sinSq2theta"),
288  tmpPoint->getRealValue("deltaMSq")), 0);
289  }
290  }
291 
292 
293  delete tmpPoint;
294  }
295 
296  if (interval){
297  Double_t level=0.5;
298  forContour->SetContour(1,&level);
299  forContour->SetLineWidth(2);
300  forContour->SetLineColor(kRed);
301  forContour->Draw("cont2,same");
302  }
303  }
304 
305  MCMCIntervalPlot* mcPlot = NULL;
306  if (mcInt) {
307  cout << "MCMC actual confidence level: "
308  << mcInt->GetActualConfidenceLevel() << endl;
309  mcPlot = new MCMCIntervalPlot(*mcInt);
310  mcPlot->SetLineColor(kMagenta);
311  mcPlot->Draw();
312  }
313  dataCanvas->Update();
314 
315  LikelihoodIntervalPlot plotInt((LikelihoodInterval*)plcInterval);
316  plotInt.SetTitle("90% Confidence Intervals");
317  if (mcInt)
318  plotInt.Draw("same");
319  else
320  plotInt.Draw();
321  dataCanvas->Update();
322 
323  /// print timing info
324  t.Stop();
325  t.Print();
326 
327 
328 }
329 
virtual void SetLineWidth(Width_t lwidth)
Set the line width.
Definition: TAttLine.h:43
virtual Int_t FindBin(Double_t x, Double_t y=0, Double_t z=0)
Return Global bin number corresponding to x,y,z.
Definition: TH1.cxx:3572
RooAddPdf is an efficient implementation of a sum of PDFs of the form.
Definition: RooAddPdf.h:29
ModelConfig is a simple class that holds configuration information specifying how a model should be u...
Definition: ModelConfig.h:30
virtual const RooArgSet * get() const
Definition: RooDataHist.h:77
virtual RooPlot * plotOn(RooPlot *frame, const RooCmdArg &arg1=RooCmdArg::none(), const RooCmdArg &arg2=RooCmdArg::none(), const RooCmdArg &arg3=RooCmdArg::none(), const RooCmdArg &arg4=RooCmdArg::none(), const RooCmdArg &arg5=RooCmdArg::none(), const RooCmdArg &arg6=RooCmdArg::none(), const RooCmdArg &arg7=RooCmdArg::none(), const RooCmdArg &arg8=RooCmdArg::none()) const
Plot dataset on specified frame.
Definition: RooAbsData.cxx:568
void Start(Bool_t reset=kTRUE)
Start the stopwatch.
Definition: TStopwatch.cxx:58
This class provides simple and straightforward utilities to plot a MCMCInterval object.
Class RooProfileLL implements the profile likelihood estimator for a given likelihood and set of para...
Definition: RooProfileLL.h:26
void Print(Option_t *option="") const
Print the real and cpu time passed between the start and stop events.
Definition: TStopwatch.cxx:219
RooCmdArg LineColor(Color_t color)
LikelihoodInterval is a concrete implementation of the RooStats::ConfInterval interface.
virtual void SetWorkspace(RooWorkspace &ws)
Definition: ModelConfig.h:66
Definition: Rtypes.h:59
virtual void SetContour(Int_t nlevels, const Double_t *levels=0)
Set the number and values of contour levels.
Definition: TH1.cxx:7785
TVirtualPad * cd(Int_t subpadnumber=0)
Set current canvas & pad.
Definition: TCanvas.cxx:688
This class provides simple and straightforward utilities to plot a LikelihoodInterval object...
ProfileLikelihoodCalculator is a concrete implementation of CombinedCalculator (the interface class f...
virtual Bool_t IsInInterval(const RooArgSet &) const =0
check if given point is in the interval
int Int_t
Definition: RtypesCore.h:41
void Draw(const Option_t *options=NULL)
void SetTitle(const char *name)
Set the title of the RooPlot to &#39;title&#39;.
Definition: RooPlot.cxx:1098
Definition: Rtypes.h:59
static RooMsgService & instance()
Return reference to singleton instance.
RooCmdArg Extended(Bool_t flag=kTRUE)
void setStreamStatus(Int_t id, Bool_t active)
(De)Activate stream with given unique ID
static struct mg_connection * fc(struct mg_context *ctx)
Definition: civetweb.c:3352
RooDataSet is a container class to hold N-dimensional binned data.
Definition: RooDataHist.h:40
void Stop()
Stop the stopwatch.
Definition: TStopwatch.cxx:77
virtual void SetSeed(ULong_t seed=0)
Set the random generator seed.
Definition: TRandom.cxx:589
virtual void SetPdf(const RooAbsPdf &pdf)
Set the Pdf, add to the the workspace if not already there.
Definition: ModelConfig.h:75
static constexpr double L
static TRandom * randomGenerator()
Return a pointer to a singleton random-number generator implementation.
Definition: RooRandom.cxx:54
RooConstVar represent a constant real-valued object.
Definition: RooConstVar.h:25
virtual RooAbsPdf * createProjection(const RooArgSet &iset)
Return a p.d.f that represent a projection of this p.d.f integrated over given observables.
Definition: RooAbsPdf.cxx:3012
virtual TObject * clone(const char *newname) const
Definition: RooArgSet.h:82
RooRealVar represents a fundamental (non-derived) real valued object.
Definition: RooRealVar.h:36
virtual void SetLineColor(Color_t lcolor)
Set the line color.
Definition: TAttLine.h:40
RooAbsReal * createIntegral(const RooArgSet &iset, const RooCmdArg &arg1, const RooCmdArg &arg2=RooCmdArg::none(), const RooCmdArg &arg3=RooCmdArg::none(), const RooCmdArg &arg4=RooCmdArg::none(), const RooCmdArg &arg5=RooCmdArg::none(), const RooCmdArg &arg6=RooCmdArg::none(), const RooCmdArg &arg7=RooCmdArg::none(), const RooCmdArg &arg8=RooCmdArg::none()) const
Create an object that represents the integral of the function over one or more observables listed in ...
Definition: RooAbsReal.cxx:501
Class RooNLLVar implements a a -log(likelihood) calculation from a dataset and a PDF.
Definition: RooNLLVar.h:26
virtual void Draw(Option_t *option="")
Draw this histogram with options.
Definition: TH1.cxx:2974
2-D histogram with a float per channel (see TH1 documentation)}
Definition: TH2.h:250
void SetLineColor(Color_t color)
RooDataSet is a container class to hold unbinned data.
Definition: RooDataSet.h:29
RooProduct a RooAbsReal implementation that represent the product of a given set of other RooAbsReal ...
Definition: RooProduct.h:32
constexpr Double_t E()
Base of natural log: .
Definition: TMath.h:97
A RooPlot is a plot frame and a container for graphics objects within that frame. ...
Definition: RooPlot.h:41
The FeldmanCousins class (like the Feldman-Cousins technique) is essentially a specific configuration...
const Bool_t kFALSE
Definition: RtypesCore.h:88
virtual Double_t GetActualConfidenceLevel()
virtual Double_t GetKeysPdfCutoff() { return fKeysCutoff; }
The Canvas class.
Definition: TCanvas.h:31
ConfInterval is an interface class for a generic interval in the RooStats framework.
Definition: ConfInterval.h:35
Namespace for the RooStats classes.
Definition: Asimov.h:20
virtual Int_t numEntries() const
Return the number of bins.
double Double_t
Definition: RtypesCore.h:55
RooAbsReal is the common abstract base class for objects that represent a real value and implements f...
Definition: RooAbsReal.h:53
RooCmdArg Components(const RooArgSet &compSet)
RooCmdArg YVar(const RooAbsRealLValue &var, const RooCmdArg &arg=RooCmdArg::none())
The TH1 histogram class.
Definition: TH1.h:56
RooCmdArg Scaling(Bool_t flag)
RooAbsPdf is the abstract interface for all probability density functions The class provides hybrid a...
Definition: RooAbsPdf.h:41
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
Double_t getRealValue(const char *name, Double_t defVal=0, Bool_t verbose=kFALSE) const
Get value of a RooAbsReal stored in set with given name.
Definition: RooArgSet.cxx:526
TObject * Clone(const char *newname=0) const
Make a complete copy of the underlying object.
Definition: TH1.cxx:2657
RooPolynomial implements a polynomial p.d.f of the form By default coefficient a_0 is chosen to be 1...
Definition: RooPolynomial.h:28
virtual void SetParametersOfInterest(const RooArgSet &set)
Definition: ModelConfig.h:93
Definition: Rtypes.h:59
virtual void SetTitle(const char *title)
See GetStatOverflows for more information.
Definition: TH1.cxx:6192
virtual void SetBinContent(Int_t bin, Double_t content)
Set bin content.
Definition: TH2.cxx:2500
MCMCInterval is a concrete implementation of the RooStats::ConfInterval interface.
Definition: MCMCInterval.h:30
virtual void Update()
Update canvas pad buffers.
Definition: TCanvas.cxx:2248
TH1 * createHistogram(const char *name, const RooAbsRealLValue &xvar, const RooCmdArg &arg1=RooCmdArg::none(), const RooCmdArg &arg2=RooCmdArg::none(), const RooCmdArg &arg3=RooCmdArg::none(), const RooCmdArg &arg4=RooCmdArg::none(), const RooCmdArg &arg5=RooCmdArg::none(), const RooCmdArg &arg6=RooCmdArg::none(), const RooCmdArg &arg7=RooCmdArg::none(), const RooCmdArg &arg8=RooCmdArg::none()) const
Create and fill a ROOT histogram TH1,TH2 or TH3 with the values of this dataset.
Definition: RooAbsData.cxx:672
Bayesian Calculator estimating an interval or a credible region using the Markov-Chain Monte Carlo me...
const Bool_t kTRUE
Definition: RtypesCore.h:87
The RooWorkspace is a persistable container for RooFit projects.
Definition: RooWorkspace.h:43
virtual void Draw(Option_t *options=0)
Draw this plot and all of the elements it contains.
Definition: RooPlot.cxx:558
RooCmdArg Binning(const RooAbsBinning &binning)
Stopwatch class.
Definition: TStopwatch.h:28