ROOT » HIST » HIST » TEfficiency

class TEfficiency: public TNamed, public TAttLine, public TAttFill, public TAttMarker


TEfficiency - a class to handle efficiency histograms

  1. Overview
  2. Creating a TEfficiency object
  3. Fill in events
  4. Statistic options
    1. Coverage probabilities for different methods
  5. Merging and combining TEfficiency objects
    1. When should I use merging?
    2. When should I use combining?
  6. Further operations
    1. Information about the internal histograms
    2. Fitting
    3. Draw a TEfficiency object
  7. TEfficiency class

I. Overview

This class handles the calculation of efficiencies and their uncertainties. It provides several statistical methods for calculating frequentist and bayesian confidence intervals as well as a function for combining several efficiencies.
Efficiencies have a lot of applications and meanings but in principle they can be described by the fraction of good/passed events k out of sample containing N events. One is usually interested in the dependency of the efficiency on other (binned) variables. The number of passed and total events is therefore stored internally in two histograms (fTotalHistogram and fPassedHistogram). Then the efficiency as well as its upper and lower error an be calculated for each bin individually.
As the efficiency can be regarded as a parameter of a binomial distribution, the number of pass ed and total events must always be integer numbers. Therefore a filling with weights is not possible however you can assign a global weight to each TEfficiency object (SetWeight). It is necessary to create one TEfficiency object for each weight if you investigate a process involving different weights. This procedure needs more effort but enables you to re-use the filled object in cases where you want to change one or more weights. This would not be possible if all events with different weights were filled in the same histogram.

II. Creating a TEfficiency object

If you start a new analysis, it is highly recommended to use the TEfficiency class from the beginning. You can then use one of the constructors for fixed or variable bin size and your desired dimension. These constructors append the created TEfficiency object to the current directory. So it will be written automatically to a file during the next TFile::Write command.
 Example:
 create a twodimensional TEfficiency object with
 - name = "eff"
 - title = "my efficiency"
 - axistitles: x, y and LaTeX formated epsilon as label for Z axis
 - 10 bins with constant bin width (= 1) along X axis starting at 0 (lower edge
 from first bin) upto 10 (upper edge of last bin)
 - 20 bins with constant bin width (= 0.5) along Y axis starting at -5 (lower
 edge from first bin) upto 5 (upper edge of last bin)
TEfficiency* pEff = new TEfficiency("eff","my efficiency;x;y;#epsilon",10,0,10,20,-5,5);
If you already have two histograms filled with the number of passed and total events, you will use the constructor TEfficiency(const TH1& passed,const TH1& total) to construct the TEfficiency object. The histograms "passed" and "total" have to fullfill the conditions mentioned in CheckConsistency, otherwise the construction will fail. As the histograms already exist, the new TEfficiency is by default not attached to the current directory to avoid duplication of data. If you want to store the new object anyway, you can either write it directly by calling Write or attach it to a directory using SetDirectory. This also applies for TEfficiency objects created by the copy constructor TEfficiency(const TEfficiency& rEff).
Example 1:
TEfficiency* pEff = 0;
TFile* pFile = new TFile("myfile.root","recreate");
//h_pass and h_total are valid and consistent histograms
if(TEfficiency::CheckConsistency(h_pass,h_total))
{
  pEff = new TEfficiency(h_pass,h_total);
  // this will write the TEfficiency object to "myfile.root"
  // AND pEff will be attached to the current directory
  pEff->Write();
}
Example 2:
TEfficiency* pEff = 0;
TFile* pFile = new TFile("myfile.root","recreate");
//h_pass and h_total are valid and consistent histograms
if(TEfficiency::CheckConsistency(h_pass,h_total))
{
  pEff = new TEfficiency(h_pass,h_total);
  //this will attach the TEfficiency object to the current directory
  pEff->SetDirectory(gDirectory);
  //now all objects in gDirectory will be written to "myfile.root"
  pFile->Write();
}
In the case that you already have two filled histograms and you only want to plot them as a graph, you should rather use TGraphAsymmErrors::TGraphAsymmErrors(const TH1* pass,const TH1* total,Option_t* opt) to create a graph object.

III. Filling with events

You can fill the TEfficiency object by calling the Fill(Bool_t bPassed,Double_t x,Double_t y,Double_t z) method. The boolean flag "bPassed" indicates whether the current event is a good (both histograms are filled) or not (only fTotalHistogram is filled). The variables x,y and z determine the bin which is filled. For lower dimensions the z- or even the y-value may be omitted.

output of htmldoc/MACRO_TEfficiency_1.gif
{
   //canvas only needed for this documentation
   TCanvas* c1 = new TCanvas("example","",600,400);
   c1->SetFillStyle(1001);
   c1->SetFillColor(kWhite);

   //create one-dimensional TEfficiency object with fixed bin size
   TEfficiency* pEff = new TEfficiency("eff","my efficiency;x;#epsilon",20,0,10);
   TRandom3 rand3;

   bool bPassed;
   double x;
   for(int i=0; i<10000; ++i)
   {
      //simulate events with variable under investigation
      x = rand3.Uniform(10);
      //check selection: bPassed = DoesEventPassSelection(x)
      bPassed = rand3.Rndm() < TMath::Gaus(x,5,4);
      pEff->Fill(bPassed,x);
   }

   pEff->Draw("AP");

   //only for this documentation
   return c1;
}
   

You can also set the number of passed or total events for a bin directly by using the SetPassedEvents or SetTotalEvents method.

IV. Statistic options

The calculation of the estimated efficiency depends on the chosen statistic option. Let k denotes the number of passed events and N the number of total events.
Frequentist methods
The expectation value of the number of passed events is given by the true efficiency times the total number of events. One can estimate the efficiency by replacing the expected number of passed events by the observed number of passed events.

#LT k #GT = #epsilon #times N    #Rightarrow    #hat{#varepsilon} = #frac{k}{N}

Bayesian methods
In bayesian statistics a likelihood-function (how probable is it to get the observed data assuming a true efficiency) and a prior probability (what is the probability that a certain true efficiency is actually realised) are used to determine a posterior probability by using Bayes theorem. At the moment, only beta distributions (have 2 free parameters) are supported as prior probabilities.

P(#epsilon | k ; N) = #frac{1}{norm} #times P(k | #epsilon ; N) #times Prior(#epsilon)

By default the expectation value of this posterior distribution is used as estimator for the efficiency:

#hat{#varepsilon} = #frac{k + #alpha}{N + #alpha + #beta}

Optionally the mode can also be used as value for the estimated efficiency. This can be done by calling SetBit(kPosteriorMode) or SetPosteriorMode. In this case the estimated efficiency is:

#hat{#varepsilon} = #frac{k + #alpha -1}{N + #alpha + #beta - 2}

In the case of a uniform prior distribution, B(x,1,1), the posterior mode is k/n, equivalent to the frequentist estimate (the maximum likelihood value). The statistic options also specifiy which confidence interval is used for calculating the uncertainties of the efficiency. The following properties define the error calculation: In the following table the implemented confidence intervals are listed with their corresponding statistic option. For more details on the calculation, please have a look at the the mentioned functions.

implemented confidence intervals and their options
namestatistic optionfunctionkIsBayesianparameters
Clopper-PearsonkFCP ClopperPearson false
  • total events
  • passed events
  • confidence level
normal approximationkFNormal Normal false
  • total events
  • passed events
  • confidence level
WilsonkFWilson Wilson false
  • total events
  • passed events
  • confidence level
Agresti-CoullkFAC AgrestiCoull false
  • total events
  • passed events
  • confidence level
Feldman-CousinskFFC FeldmanCousins false
  • total events
  • passed events
  • confidence level
JeffreykBJeffrey Bayesian true
  • total events
  • passed events
  • confidence level
  • fBeta_alpha = 0.5
  • fBeta_beta = 0.5
Uniform priorkBUniform Bayesian true
  • total events
  • passed events
  • confidence level
  • fBeta_alpha = 1
  • fBeta_beta = 1
custom priorkBBayesian Bayesian true
  • total events
  • passed events
  • confidence level
  • fBeta_alpha
  • fBeta_beta

The following example demonstrates the effect of different statistic options and confidence levels.

output of htmldoc/MACRO_TEfficiency_11.gif
{
   //canvas only needed for the documentation
   TCanvas* c1 = new TCanvas("c1","",600,400);
   c1->Divide(2);
   c1->SetFillStyle(1001);
   c1->SetFillColor(kWhite);

   //create one-dimensional TEfficiency object with fixed bin size
   TEfficiency* pEff = new TEfficiency("eff","different confidence levels;x;#epsilon",20,0,10);
   TRandom3 rand3;

   bool bPassed;
   double x;
   for(int i=0; i<1000; ++i)
   {
      //simulate events with variable under investigation
      x = rand3.Uniform(10);
      //check selection: bPassed = DoesEventPassSelection(x)
      bPassed = rand3.Rndm() < TMath::Gaus(x,5,4);
      pEff->Fill(bPassed,x);
   }

   //set style attributes
   pEff->SetFillStyle(3004);
   pEff->SetFillColor(kRed);

   //copy current TEfficiency object and set new confidence level
   TEfficiency* pCopy = new TEfficiency(*pEff);
   pCopy->SetConfidenceLevel(0.90);

   //set style attributes
   pCopy->SetFillStyle(3005);
   pCopy->SetFillColor(kBlue);

   c1->cd(1);

   //add legend
   TLegend* leg1 = new TLegend(0.3,0.1,0.7,0.5);
   leg1->AddEntry(pEff,"95%","F");
   leg1->AddEntry(pCopy,"68.3%","F");

   pEff->Draw("A4");
   pCopy->Draw("same4");
   leg1->Draw("same");

   //use same confidence level but different statistic methods
   TEfficiency* pEff2 = new TEfficiency(*pEff);
   TEfficiency* pCopy2 = new TEfficiency(*pEff);

   pEff2->SetStatisticOption(TEfficiency::kFNormal);
   pCopy2->SetStatisticOption(TEfficiency::kFAC);

   pEff2->SetTitle("different statistic options;x;#epsilon");

   //set style attributes
   pCopy2->SetFillStyle(3005);
   pCopy2->SetFillColor(kBlue);

   c1->cd(2);

   //add legend
   TLegend* leg2 = new TLegend(0.3,0.1,0.7,0.5);
   leg2->AddEntry(pEff2,"kFNormal","F");
   leg2->AddEntry(pCopy2,"kFAC","F");

   pEff2->Draw("a4");
   pCopy2->Draw("same4");
   leg2->Draw("same");

   //only for this documentation
   c1->cd(0);
   return c1;
}
   

The prior probability of the efficiency in bayesian statistics can be given in terms of a beta distribution. The beta distribution has to positive shape parameters. The resulting priors for different combinations of these shape parameters are shown in the plot below.

output of htmldoc/MACRO_TEfficiency_13.gif
   {
      //canvas only needed for the documentation
      TCanvas* c1 = new TCanvas("c1","",600,400);
      c1->SetFillStyle(1001);
      c1->SetFillColor(kWhite);

      //create different beta distributions
      TF1* f1 = new TF1("f1","TMath::BetaDist(x,1,1)",0,1);
      f1->SetLineColor(kBlue);
      TF1* f2 = new TF1("f2","TMath::BetaDist(x,0.5,0.5)",0,1);
      f2->SetLineColor(kRed);
      TF1* f3 = new TF1("f3","TMath::BetaDist(x,1,5)",0,1);
      f3->SetLineColor(kGreen+3);
      f3->SetTitle("Beta distributions as priors;#epsilon;P(#epsilon)");
      TF1* f4 = new TF1("f4","TMath::BetaDist(x,4,3)",0,1);
      f4->SetLineColor(kViolet);

      //add legend
      TLegend* leg = new TLegend(0.25,0.5,0.85,0.89);
      leg->SetFillColor(kWhite);
      leg->SetFillStyle(1001);
      leg->AddEntry(f1,"a=1, b=1","L");
      leg->AddEntry(f2,"a=0.5, b=0.5","L");
      leg->AddEntry(f3,"a=1, b=5","L");
      leg->AddEntry(f4,"a=4, b=3","L");

      f3->Draw();
      f1->Draw("same");
      f2->Draw("Same");
      f4->Draw("same");
      leg->Draw("same");

      //only for this documentation
      return c1;
   }
   

IV.1 Coverage probabilities for different methods

The following pictures illustrate the actual coverage probability for the different values of the true efficiency and the total number of events when a confidence level of 95% is desired.

normal approximation

wilson

agresti coull

clopper pearson

uniform prior

jeffrey prior

The average (over all possible true efficiencies) coverage probability for different number of total events is shown in the next picture.

average coverage

V. Merging and combining TEfficiency objects

In many applications the efficiency should be calculated for an inhomogenous sample in the sense that it contains events with different weights. In order to be able to determine the correct overall efficiency, it is necessary to use for each subsample (= all events with the same weight) a different TEfficiency object. After finsihing your analysis you can then construct the overall efficiency with its uncertainty.
This procedure has the advantage that you can change the weight of one subsample easily without rerunning the whole analysis. On the other hand more efford is needed to handle several TEfficiency objects instead of one histogram. In the case of many different or even continuously distributed weights this approach becomes cumbersome. One possibility to overcome this problem is the usage of binned weights.

Example
In high particle physics weights arises from the fact that you want to normalise your results to a certain reference value. A very common formula for calculating weights is

w = #frac{#sigma L}{N_{gen} #epsilon_{trig}} - #sigma ... cross section

The reason for different weights can therefore be:
  • different processes
  • other integrated luminosity
  • varying trigger efficiency
  • different sample sizes
  • ...
  • or even combination of them
Depending on the actual meaning of different weights in your case, you should either merge or combine them to get the overall efficiency.

V.1 When should I use merging?

If the weights are artificial and do not represent real alternative hypotheses, you should merge the different TEfficiency objects. That means especially for the bayesian case that the prior probability should be the same for all merged TEfficiency objects. The merging can be done by invoking one of the following operations:
  • eff1.Add(eff2)
  • eff1 += eff2
  • eff = eff1 + eff2
The result of the merging is stored in the TEfficiency object which is marked bold above. The contents of the internal histograms of both TEfficiency objects are added and a new weight is assigned. The statistic options are not changed.
   #frac{1}{w_{new}} = #frac{1}{w_{1}} + #frac{1}{w_{2}}

Example:
If you use two samples with different numbers of generated events for the same process and you want to normalise both to the same integrated luminosity and trigger efficiency, the different weights then arise just from the fact that you have different numbers of events. The TEfficiency objects should be merged because the samples do not represent true alternatives. You expect the same result as if you would have a big sample with all events in it.

w_{1} = #frac{#sigma L}{#epsilon N_{1}}, w_{2} = #frac{#sigma L}{#epsilon N_{2}} #Rightarrow w_{new} = #frac{#sigma L}{#epsilon (N_{1} + N_{2})} = #frac{1}{#frac{1}{w_{1}} + #frac{1}{w_{2}}}

V.2 When should I use combining?

You should combine TEfficiency objects whenever the weights represent alternatives processes for the efficiency. As the combination of two TEfficiency objects is not always consistent with the representation by two internal histograms, the result is not stored in a TEfficiency object but a TGraphAsymmErrors is returned which shows the estimated combined efficiency and its uncertainty for each bin. At the moment the combination method Combine only supports combination of 1-dimensional efficiencies in a bayesian approach.
For calculating the combined efficiency and its uncertainty for each bin only Bayesian statistics is used. No frequentists methods are presently supported for computing the combined efficiency and its confidence interval. In the case of the Bayesian statistics a combined posterior is constructed taking into account the weight of each TEfficiency object. The same prior is used for all the TEfficiency objects.
P_{comb}(#epsilon | {w_{i}}, {k_{i}} , {N_{i}}) = #frac{1}{norm} #prod_{i}{L(k_{i} | N_{i}, #epsilon)}^{w_{i}} #Pi( #epsilon )
Example:
If you use cuts to select electrons which can originate from two different processes, you can determine the selection efficiency for each process. The overall selection efficiency is then the combined efficiency. The weights to be used in the combination should be the probability that an electron comes from the corresponding process.
p_{1} = #frac{#sigma_{1}}{#sigma_{1} + #sigma_{2}} = #frac{N_{1}w_{1}}{N_{1}w_{1} + N_{2}w_{2}}

VI. Further operations

VI.1 Information about the internal histograms

The methods GetPassedHistogram and GetTotalHistogram return a constant pointer to the internal histograms. They can be used to obtain information about the internal histograms (e.g. the binning, number of passed / total events in a bin, mean values...). One can obtain a clone of the internal histograms by calling GetCopyPassedHisto or GetCopyTotalHisto. The returned histograms are completely independent from the current TEfficiency object. By default, they are not attached to a directory to avoid the duplication of data and the user is responsible for deleting them.
Example:
//open a root file which contains a TEfficiency object
TFile* pFile = new TFile("myfile.root","update");
//get TEfficiency object with name "my_eff"
TEfficiency* pEff = (TEfficiency*)pFile->Get("my_eff");
//get clone of total histogram
TH1* clone = pEff->GetCopyTotalHisto();
//change clone...
//save changes of clone directly
clone->Write();
//or append it to the current directoy and write the file
//clone->SetDirectory(gDirectory);
//pFile->Wrtie();
//delete histogram object
delete clone;
clone = 0;
It is also possible to set the internal total or passed histogram by using the methods SetPassedHistogram or SetTotalHistogram. In order to ensure the validity of the TEfficiency object, the consistency of the new histogram and the stored histogram is checked. It sometimes might be impossible to change the histograms in a consistent way. Therefore one can force the replacement by passing the option "f". Then the user has to ensure that the other internal histogram is replaced as well and that the TEfficiency object is in a valid state.

VI.2 Fitting

The efficiency can be fitted using the Fit function which uses internally the TBinomialEfficiencyFitter::Fit method. As this method is using a maximum-likelihood-fit, it is necessary to initialise the given fit function with reasonable start values. The resulting fit function is attached to the list of associated functions and will be drawn automatically during the next Draw command. The list of associated function can be modified by using the pointer returned by GetListOfFunctions.

output of htmldoc/MACRO_TEfficiency_25.gif
{
   //canvas only needed for this documentation
   TCanvas* c1 = new TCanvas("example","",600,400);
   c1->SetFillStyle(1001);
   c1->SetFillColor(kWhite);

   //create one-dimensional TEfficiency object with fixed bin size
   TEfficiency* pEff = new TEfficiency("eff","my efficiency;x;#epsilon",20,0,10);
   TRandom3 rand3;

   bool bPassed;
   double x;
   for(int i=0; i<10000; ++i)
   {
      //simulate events with variable under investigation
      x = rand3.Uniform(10);
      //check selection: bPassed = DoesEventPassSelection(x)
      bPassed = rand3.Rndm() < TMath::Gaus(x,5,4);
      pEff->Fill(bPassed,x);
   }

   //create a function for fitting and do the fit
   TF1* f1 = new TF1("f1","gaus",0,10);
   f1->SetParameters(1,5,2);
   pEff->Fit(f1);

   //create a threshold function
   TF1* f2 = new TF1("thres","0.8",0,10);
   f2->SetLineColor(kRed);
   //add it to the list of functions
   //use add first because the parameters of the last function will be displayed
   pEff->GetListOfFunctions()->AddFirst(f2);

   pEff->Draw("AP");

   //only for this documentation
   return c1;
}
   

VI.3 Draw a TEfficiency object

A TEfficiency object can be drawn by calling the usual Draw method. At the moment drawing is only supported for 1- and 2-dimensional TEfficiency objects. In the 1-dimensional case you can use the same options as for the
TGraphAsymmErrors::Draw method. For 2-dimensional TEfficiency objects you can pass the same options as for a TH2::Draw object.

VII. TEfficiency class



Function Members (Methods)

public:
virtual~TEfficiency()
voidTObject::AbstractMethod(const char* method) const
voidAdd(const TEfficiency& rEff)
static Double_tAgrestiCoull(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
virtual voidTObject::AppendPad(Option_t* option = "")
static Double_tBayesian(Int_t total, Int_t passed, Double_t level, Double_t alpha, Double_t beta, Bool_t bUpper, Bool_t bShortest = false)
static Double_tBetaCentralInterval(Double_t level, Double_t alpha, Double_t beta, Bool_t bUpper)
static Double_tBetaMean(Double_t alpha, Double_t beta)
static Double_tBetaMode(Double_t alpha, Double_t beta)
static Bool_tBetaShortestInterval(Double_t level, Double_t alpha, Double_t beta, Double_t& lower, Double_t& upper)
virtual voidTObject::Browse(TBrowser* b)
static Bool_tCheckBinning(const TH1& pass, const TH1& total)
static Bool_tCheckConsistency(const TH1& pass, const TH1& total, Option_t* opt = "")
static Bool_tCheckEntries(const TH1& pass, const TH1& total, Option_t* opt = "")
static TClass*Class()
virtual const char*TObject::ClassName() const
virtual voidTNamed::Clear(Option_t* option = "")
virtual TObject*TNamed::Clone(const char* newname = "") const
static Double_tClopperPearson(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
static TGraphAsymmErrors*Combine(TCollection* pList, Option_t* opt = "", Int_t n = 0, const Double_t* w = 0)
static Double_tCombine(Double_t& up, Double_t& low, Int_t n, const Int_t* pass, const Int_t* total, Double_t alpha, Double_t beta, Double_t level = 0.68300000000000005, const Double_t* w = 0, Option_t* opt = "")
virtual Int_tTNamed::Compare(const TObject* obj) const
virtual voidTNamed::Copy(TObject& named) const
TGraphAsymmErrors*CreateGraph(Option_t* opt = "") const
TH2*CreateHistogram(Option_t* opt = "") const
virtual voidTObject::Delete(Option_t* option = "")MENU
Int_tTAttLine::DistancetoLine(Int_t px, Int_t py, Double_t xp1, Double_t yp1, Double_t xp2, Double_t yp2)
virtual Int_tDistancetoPrimitive(Int_t px, Int_t py)
virtual voidDraw(Option_t* opt = "")
virtual voidTObject::DrawClass() constMENU
virtual TObject*TObject::DrawClone(Option_t* option = "") constMENU
virtual voidTObject::Dump() constMENU
virtual voidTObject::Error(const char* method, const char* msgfmt) const
virtual voidTObject::Execute(const char* method, const char* params, Int_t* error = 0)
virtual voidTObject::Execute(TMethod* method, TObjArray* params, Int_t* error = 0)
virtual voidExecuteEvent(Int_t event, Int_t px, Int_t py)
virtual voidTObject::Fatal(const char* method, const char* msgfmt) const
static Double_tFeldmanCousins(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
static Bool_tFeldmanCousinsInterval(Int_t total, Int_t passed, Double_t level, Double_t& lower, Double_t& upper)
voidFill(Bool_t bPassed, Double_t x, Double_t y = 0, Double_t z = 0)
virtual voidTNamed::FillBuffer(char*& buffer)
voidFillWeighted(Bool_t bPassed, Double_t weight, Double_t x, Double_t y = 0, Double_t z = 0)
Int_tFindFixBin(Double_t x, Double_t y = 0, Double_t z = 0) const
virtual TObject*TObject::FindObject(const char* name) const
virtual TObject*TObject::FindObject(const TObject* obj) const
Int_tFit(TF1* f1, Option_t* opt = "")
Double_tGetBetaAlpha(Int_t bin = -1) const
Double_tGetBetaBeta(Int_t bin = -1) const
Double_tGetConfidenceLevel() const
TH1*GetCopyPassedHisto() const
TH1*GetCopyTotalHisto() const
Int_tGetDimension() const
TDirectory*GetDirectory() const
virtual Option_t*TObject::GetDrawOption() const
static Long_tTObject::GetDtorOnly()
Double_tGetEfficiency(Int_t bin) const
Double_tGetEfficiencyErrorLow(Int_t bin) const
Double_tGetEfficiencyErrorUp(Int_t bin) const
virtual Color_tTAttFill::GetFillColor() const
virtual Style_tTAttFill::GetFillStyle() const
Int_tGetGlobalBin(Int_t binx, Int_t biny = 0, Int_t binz = 0) const
virtual const char*TObject::GetIconName() const
virtual Color_tTAttLine::GetLineColor() const
virtual Style_tTAttLine::GetLineStyle() const
virtual Width_tTAttLine::GetLineWidth() const
TList*GetListOfFunctions()
virtual Color_tTAttMarker::GetMarkerColor() const
virtual Size_tTAttMarker::GetMarkerSize() const
virtual Style_tTAttMarker::GetMarkerStyle() const
virtual const char*TNamed::GetName() const
virtual char*TObject::GetObjectInfo(Int_t px, Int_t py) const
static Bool_tTObject::GetObjectStat()
virtual Option_t*TObject::GetOption() const
TGraphAsymmErrors*GetPaintedGraph() const
TH2*GetPaintedHistogram() const
const TH1*GetPassedHistogram() const
TEfficiency::EStatOptionGetStatisticOption() const
virtual const char*TNamed::GetTitle() const
const TH1*GetTotalHistogram() const
virtual UInt_tTObject::GetUniqueID() const
Double_tGetWeight() const
virtual Bool_tTObject::HandleTimer(TTimer* timer)
virtual ULong_tTNamed::Hash() const
virtual voidTObject::Info(const char* method, const char* msgfmt) const
virtual Bool_tTObject::InheritsFrom(const char* classname) const
virtual Bool_tTObject::InheritsFrom(const TClass* cl) const
virtual voidTObject::Inspect() constMENU
voidTObject::InvertBit(UInt_t f)
virtual TClass*IsA() const
virtual Bool_tTObject::IsEqual(const TObject* obj) const
virtual Bool_tTObject::IsFolder() const
Bool_tTObject::IsOnHeap() const
virtual Bool_tTNamed::IsSortable() const
virtual Bool_tTAttFill::IsTransparent() const
Bool_tTObject::IsZombie() const
virtual voidTNamed::ls(Option_t* option = "") const
voidTObject::MayNotUse(const char* method) const
Long64_tMerge(TCollection* list)
virtual voidTAttLine::Modify()
static Double_tNormal(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
virtual Bool_tTObject::Notify()
voidTObject::Obsolete(const char* method, const char* asOfVers, const char* removedFromVers) const
voidTObject::operator delete(void* ptr)
voidTObject::operator delete(void* ptr, void* vp)
voidTObject::operator delete[](void* ptr)
voidTObject::operator delete[](void* ptr, void* vp)
void*TObject::operator new(size_t sz)
void*TObject::operator new(size_t sz, void* vp)
void*TObject::operator new[](size_t sz)
void*TObject::operator new[](size_t sz, void* vp)
TEfficiency&operator+=(const TEfficiency& rhs)
TEfficiency&operator=(const TEfficiency& rhs)
virtual voidPaint(Option_t* opt)
virtual voidTObject::Pop()
virtual voidTNamed::Print(Option_t* option = "") const
virtual Int_tTObject::Read(const char* name)
virtual voidTObject::RecursiveRemove(TObject* obj)
virtual voidTAttFill::ResetAttFill(Option_t* option = "")
virtual voidTAttLine::ResetAttLine(Option_t* option = "")
virtual voidTAttMarker::ResetAttMarker(Option_t* toption = "")
voidTObject::ResetBit(UInt_t f)
virtual voidTObject::SaveAs(const char* filename = "", Option_t* option = "") constMENU
virtual voidTAttFill::SaveFillAttributes(ostream& out, const char* name, Int_t coldef = 1, Int_t stydef = 1001)
virtual voidTAttLine::SaveLineAttributes(ostream& out, const char* name, Int_t coldef = 1, Int_t stydef = 1, Int_t widdef = 1)
virtual voidTAttMarker::SaveMarkerAttributes(ostream& out, const char* name, Int_t coldef = 1, Int_t stydef = 1, Int_t sizdef = 1)
virtual voidSavePrimitive(ostream& out, Option_t* opt = "")
voidSetBetaAlpha(Double_t alpha)
voidSetBetaBeta(Double_t beta)
voidSetBetaBinParameters(Int_t bin, Double_t alpha, Double_t beta)
Bool_tSetBins(Int_t nx, const Double_t* xBins)
Bool_tSetBins(Int_t nx, Double_t xmin, Double_t xmax)
Bool_tSetBins(Int_t nx, const Double_t* xBins, Int_t ny, const Double_t* yBins)
Bool_tSetBins(Int_t nx, Double_t xmin, Double_t xmax, Int_t ny, Double_t ymin, Double_t ymax)
Bool_tSetBins(Int_t nx, const Double_t* xBins, Int_t ny, const Double_t* yBins, Int_t nz, const Double_t* zBins)
Bool_tSetBins(Int_t nx, Double_t xmin, Double_t xmax, Int_t ny, Double_t ymin, Double_t ymax, Int_t nz, Double_t zmin, Double_t zmax)
voidTObject::SetBit(UInt_t f)
voidTObject::SetBit(UInt_t f, Bool_t set)
voidSetCentralInterval(Bool_t on = true)
voidSetConfidenceLevel(Double_t level)
voidSetDirectory(TDirectory* dir)
virtual voidTObject::SetDrawOption(Option_t* option = "")MENU
static voidTObject::SetDtorOnly(void* obj)
virtual voidTAttFill::SetFillAttributes()MENU
virtual voidTAttFill::SetFillColor(Color_t fcolor)
virtual voidTAttFill::SetFillColorAlpha(Color_t fcolor, Float_t falpha)
virtual voidTAttFill::SetFillStyle(Style_t fstyle)
virtual voidTAttLine::SetLineAttributes()MENU
virtual voidTAttLine::SetLineColor(Color_t lcolor)
virtual voidTAttLine::SetLineColorAlpha(Color_t lcolor, Float_t lalpha)
virtual voidTAttLine::SetLineStyle(Style_t lstyle)
virtual voidTAttLine::SetLineWidth(Width_t lwidth)
virtual voidTAttMarker::SetMarkerAttributes()MENU
virtual voidTAttMarker::SetMarkerColor(Color_t mcolor = 1)
virtual voidTAttMarker::SetMarkerColorAlpha(Color_t mcolor, Float_t malpha)
virtual voidTAttMarker::SetMarkerSize(Size_t msize = 1)
virtual voidTAttMarker::SetMarkerStyle(Style_t mstyle = 1)
virtual voidSetName(const char* name)
virtual voidTNamed::SetNameTitle(const char* name, const char* title)
static voidTObject::SetObjectStat(Bool_t stat)
Bool_tSetPassedEvents(Int_t bin, Int_t events)
Bool_tSetPassedHistogram(const TH1& rPassed, Option_t* opt)
voidSetPosteriorAverage(Bool_t on = true)
voidSetPosteriorMode(Bool_t on = true)
voidSetShortestInterval(Bool_t on = true)
voidSetStatisticOption(TEfficiency::EStatOption option)
virtual voidSetTitle(const char* title)
Bool_tSetTotalEvents(Int_t bin, Int_t events)
Bool_tSetTotalHistogram(const TH1& rTotal, Option_t* opt)
virtual voidTObject::SetUniqueID(UInt_t uid)
voidSetUseWeightedEvents()
voidSetWeight(Double_t weight)
virtual voidShowMembers(TMemberInspector& insp) const
virtual Int_tTNamed::Sizeof() const
virtual voidStreamer(TBuffer&)
voidStreamerNVirtual(TBuffer& ClassDef_StreamerNVirtual_b)
virtual voidTObject::SysError(const char* method, const char* msgfmt) const
TEfficiency()
TEfficiency(const TEfficiency& heff)
TEfficiency(const TH1& passed, const TH1& total)
TEfficiency(const char* name, const char* title, Int_t nbins, const Double_t* xbins)
TEfficiency(const char* name, const char* title, Int_t nbins, Double_t xlow, Double_t xup)
TEfficiency(const char* name, const char* title, Int_t nbinsx, const Double_t* xbins, Int_t nbinsy, const Double_t* ybins)
TEfficiency(const char* name, const char* title, Int_t nbinsx, Double_t xlow, Double_t xup, Int_t nbinsy, Double_t ylow, Double_t yup)
TEfficiency(const char* name, const char* title, Int_t nbinsx, const Double_t* xbins, Int_t nbinsy, const Double_t* ybins, Int_t nbinsz, const Double_t* zbins)
TEfficiency(const char* name, const char* title, Int_t nbinsx, Double_t xlow, Double_t xup, Int_t nbinsy, Double_t ylow, Double_t yup, Int_t nbinsz, Double_t zlow, Double_t zup)
Bool_tTObject::TestBit(UInt_t f) const
Int_tTObject::TestBits(UInt_t f) const
virtual voidTObject::UseCurrentStyle()
Bool_tUsesBayesianStat() const
Bool_tUsesCentralInterval() const
Bool_tUsesPosteriorAverage() const
Bool_tUsesPosteriorMode() const
Bool_tUsesShortestInterval() const
Bool_tUsesWeights() const
virtual voidTObject::Warning(const char* method, const char* msgfmt) const
static Double_tWilson(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
virtual Int_tTObject::Write(const char* name = 0, Int_t option = 0, Int_t bufsize = 0)
virtual Int_tTObject::Write(const char* name = 0, Int_t option = 0, Int_t bufsize = 0) const
protected:
voidBuild(const char* name, const char* title)
virtual voidTObject::DoError(int level, const char* location, const char* fmt, va_list va) const
voidFillGraph(TGraphAsymmErrors* graph, Option_t* opt) const
voidFillHistogram(TH2* h2) const
voidTObject::MakeZombie()

Data Members

public:
static TEfficiency::EStatOptionkBBayesian
static TEfficiency::EStatOptionkBJeffrey
static TEfficiency::EStatOptionkBUniform
static TObject::(anonymous)TObject::kBitMask
static TObject::EStatusBitsTObject::kCanDelete
static TObject::EStatusBitsTObject::kCannotPick
static TEfficiency::EStatOptionkFAC
static TEfficiency::EStatOptionkFCP
static TEfficiency::EStatOptionkFFC
static TEfficiency::EStatOptionkFNormal
static TEfficiency::EStatOptionkFWilson
static TObject::EStatusBitsTObject::kHasUUID
static TObject::EStatusBitsTObject::kInvalidObject
static TObject::(anonymous)TObject::kIsOnHeap
static TObject::EStatusBitsTObject::kIsReferenced
static TObject::EStatusBitsTObject::kMustCleanup
static TObject::EStatusBitsTObject::kNoContextMenu
static TObject::(anonymous)TObject::kNotDeleted
static TObject::EStatusBitsTObject::kObjInCanvas
static TObject::(anonymous)TObject::kOverwrite
static TObject::(anonymous)TObject::kSingleKey
static TObject::(anonymous)TObject::kWriteDelete
static TObject::(anonymous)TObject::kZombie
protected:
Double_tfBeta_alphaglobal parameter for prior beta distribution (default = 1)
Double_tfBeta_betaglobal parameter for prior beta distribution (default = 1)
vector<pair<Double_t,Double_t> >fBeta_bin_paramsparameter for prior beta distribution different bin by bin
Double_t(*)(Int_t,Int_t,Double_t,Bool_t)fBoundary!pointer to a method calculating the boundaries of confidence intervals
Double_tfConfLevelconfidence level (default = 0.683, 1 sigma)
TDirectory*fDirectory!pointer to directory holding this TEfficiency object
Color_tTAttFill::fFillColorfill area color
Style_tTAttFill::fFillStylefill area style
TList*fFunctions->pointer to list of functions
Color_tTAttLine::fLineColorline color
Style_tTAttLine::fLineStyleline style
Width_tTAttLine::fLineWidthline width
Color_tTAttMarker::fMarkerColorMarker color index
Size_tTAttMarker::fMarkerSizeMarker size
Style_tTAttMarker::fMarkerStyleMarker style
TStringTNamed::fNameobject identifier
TGraphAsymmErrors*fPaintGraph!temporary graph for painting
TH2*fPaintHisto!temporary histogram for painting
TH1*fPassedHistogramhistogram for events which passed certain criteria
TEfficiency::EStatOptionfStatisticOptiondefines how the confidence intervals are determined
TStringTNamed::fTitleobject title
TH1*fTotalHistogramhistogram for total number of events
Double_tfWeightweight for all events (default = 1)
static TEfficiency::(anonymous)kIsBayesian
static TEfficiency::(anonymous)kPosteriorMode
static TEfficiency::(anonymous)kShortestInterval
static TEfficiency::(anonymous)kUseBinPrior
static TEfficiency::(anonymous)kUseWeights

Class Charts

Class Charts

Function documentation

TEfficiency()
default constructor

should not be used explicitly
TEfficiency(const TH1& passed, const TH1& total)
constructor using two existing histograms as input

Input: passed - contains the events fullfilling some criteria
       total  - contains all investigated events

Notes: - both histograms have to fullfill the conditions of CheckConsistency (with option 'w')
       - dimension of the resulating efficiency object depends
         on the dimension of the given histograms
       - Clones of both histograms are stored internally
       - The function SetName(total.GetName() + "_clone") is called to set
         the names of the new object and the internal histograms..
       - The created TEfficiency object is NOT appended to a directory. It
         will not be written to disk during the next TFile::Write() command
         in order to prevent duplication of data. If you want to save this
         TEfficiency object anyway, you can either append it to a
         directory by calling SetDirectory(TDirectory*) or write it
         explicitly to disk by calling Write().
TEfficiency(const char* name, const char* title, Int_t nbins, const Double_t* xbins)
create 1-dimensional TEfficiency object with variable bin size

constructor creates two new and empty histograms with a given binning

 Input: name   - the common part of the name for both histograms (no blanks)
                 fTotalHistogram has name: name + "_total"
                 fPassedHistogram has name: name + "_passed"
        title  - the common part of the title for both histogram
                 fTotalHistogram has title: title + " (total)"
                 fPassedHistogram has title: title + " (passed)"
                 It is possible to label the axis by passing a title with
                 the following format: "title;xlabel;ylabel".
        nbins  - number of bins on the x-axis
        xbins  - array of length (nbins + 1) with low-edges for each bin
                 xbins[nbinsx] ... lower edge for overflow bin
TEfficiency(const char* name, const char* title, Int_t nbins, Double_t xlow, Double_t xup)
create 1-dimensional TEfficiency object with fixed bins isze

constructor creates two new and empty histograms with a fixed binning

 Input: name   - the common part of the name for both histograms(no blanks)
                 fTotalHistogram has name: name + "_total"
                 fPassedHistogram has name: name + "_passed"
        title  - the common part of the title for both histogram
                 fTotalHistogram has title: title + " (total)"
                 fPassedHistogram has title: title + " (passed)"
                 It is possible to label the axis by passing a title with
                 the following format: "title;xlabel;ylabel".
        nbinsx - number of bins on the x-axis
        xlow   - lower edge of first bin
        xup    - upper edge of last bin
TEfficiency(const char* name, const char* title, Int_t nbinsx, Double_t xlow, Double_t xup, Int_t nbinsy, Double_t ylow, Double_t yup)
create 2-dimensional TEfficiency object with fixed bin size

constructor creates two new and empty histograms with a fixed binning

 Input: name   - the common part of the name for both histograms(no blanks)
                 fTotalHistogram has name: name + "_total"
                 fPassedHistogram has name: name + "_passed"
        title  - the common part of the title for both histogram
                 fTotalHistogram has title: title + " (total)"
                 fPassedHistogram has title: title + " (passed)"
                 It is possible to label the axis by passing a title with
                 the following format: "title;xlabel;ylabel;zlabel".
        nbinsx - number of bins on the x-axis
        xlow   - lower edge of first x-bin
        xup    - upper edge of last x-bin
        nbinsy - number of bins on the y-axis
        ylow   - lower edge of first y-bin
        yup    - upper edge of last y-bin
TEfficiency(const char* name, const char* title, Int_t nbinsx, const Double_t* xbins, Int_t nbinsy, const Double_t* ybins)
create 2-dimensional TEfficiency object with variable bin size

constructor creates two new and empty histograms with a given binning

 Input: name   - the common part of the name for both histograms(no blanks)
                 fTotalHistogram has name: name + "_total"
                 fPassedHistogram has name: name + "_passed"
        title  - the common part of the title for both histogram
                 fTotalHistogram has title: title + " (total)"
                 fPassedHistogram has title: title + " (passed)"
                 It is possible to label the axis by passing a title with
                 the following format: "title;xlabel;ylabel;zlabel".
        nbinsx - number of bins on the x-axis
        xbins  - array of length (nbins + 1) with low-edges for each bin
                 xbins[nbinsx] ... lower edge for overflow x-bin
        nbinsy - number of bins on the y-axis
        ybins  - array of length (nbins + 1) with low-edges for each bin
                 ybins[nbinsy] ... lower edge for overflow y-bin
TEfficiency(const char* name, const char* title, Int_t nbinsx, Double_t xlow, Double_t xup, Int_t nbinsy, Double_t ylow, Double_t yup, Int_t nbinsz, Double_t zlow, Double_t zup)
create 3-dimensional TEfficiency object with fixed bin size

constructor creates two new and empty histograms with a fixed binning

 Input: name   - the common part of the name for both histograms(no blanks)
                 fTotalHistogram has name: name + "_total"
                 fPassedHistogram has name: name + "_passed"
        title  - the common part of the title for both histogram
                 fTotalHistogram has title: title + " (total)"
                 fPassedHistogram has title: title + " (passed)"
                 It is possible to label the axis by passing a title with
                 the following format: "title;xlabel;ylabel;zlabel".
        nbinsx - number of bins on the x-axis
        xlow   - lower edge of first x-bin
        xup    - upper edge of last x-bin
        nbinsy - number of bins on the y-axis
        ylow   - lower edge of first y-bin
        yup    - upper edge of last y-bin
        nbinsz - number of bins on the z-axis
        zlow   - lower edge of first z-bin
        zup    - upper edge of last z-bin
TEfficiency(const char* name, const char* title, Int_t nbinsx, const Double_t* xbins, Int_t nbinsy, const Double_t* ybins, Int_t nbinsz, const Double_t* zbins)
create 3-dimensional TEfficiency object with variable bin size

constructor creates two new and empty histograms with a given binning

 Input: name   - the common part of the name for both histograms(no blanks)
                 fTotalHistogram has name: name + "_total"
                 fPassedHistogram has name: name + "_passed"
        title  - the common part of the title for both histogram
                 fTotalHistogram has title: title + " (total)"
                 fPassedHistogram has title: title + " (passed)"
                 It is possible to label the axis by passing a title with
                 the following format: "title;xlabel;ylabel;zlabel".
        nbinsx - number of bins on the x-axis
        xbins  - array of length (nbins + 1) with low-edges for each bin
                 xbins[nbinsx] ... lower edge for overflow x-bin
        nbinsy - number of bins on the y-axis
        ybins  - array of length (nbins + 1) with low-edges for each bin
                 xbins[nbinsx] ... lower edge for overflow y-bin
        nbinsz - number of bins on the z-axis
        zbins  - array of length (nbins + 1) with low-edges for each bin
                 xbins[nbinsx] ... lower edge for overflow z-bin
TEfficiency(const TEfficiency& heff)
copy constructor

The list of associated objects (e.g. fitted functions) is not copied.

Note: - SetName(rEff.GetName() + "_copy") is called to set the names of the
        object and the histograms.
      - The titles are set by calling SetTitle("[copy] " + rEff.GetTitle()).
      - The copied TEfficiency object is NOT appended to a directory. It
        will not be written to disk during the next TFile::Write() command
        in order to prevent duplication of data. If you want to save this
        TEfficiency object anyway, you can either append it to a directory
        by calling SetDirectory(TDirectory*) or write it explicitly to disk
        by calling Write().
~TEfficiency()
default destructor
Double_t AgrestiCoull(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
calculates the boundaries for the frequentist Agresti-Coull interval

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned

calculation:

#alpha = 1 - #frac{level}{2}
Double_t FeldmanCousins(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
calculates the boundaries for the frequentist Feldman-Cousins interval

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned


Bool_t FeldmanCousinsInterval(Int_t total, Int_t passed, Double_t level, Double_t& lower, Double_t& upper)
calculates the interval boundaries using the frequentist methods of Feldman-Cousins

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
Output:
       - lower :  lower boundary returned on exit
       - upper :  lower boundary returned on exit

Return a flag with the status of the calculation

 Calculation:
 The Feldman-Cousins is a frequentist method where the interval is estimated using a Neyman construction where the ordering
 is based on the likelihood ratio:

LR =  #frac{Binomial(k | N, #epsilon)}{Binomial(k | N, #hat{#epsilon} ) }
See G. J. Feldman and R. D. Cousins, Phys. Rev. D57 (1998) 3873
 and   R. D. Cousins, K. E. Hymes, J. Tucker, Nuclear Instruments and Methods in Physics Research A 612 (2010) 388

 Implemented using classes developed by Jordan Tucker and Luca Lista
 See File hist/hist/src/TEfficiencyHelper.h

Double_t Bayesian(Int_t total, Int_t passed, Double_t level, Double_t alpha, Double_t beta, Bool_t bUpper, Bool_t bShortest = false)
calculates the boundaries for a Bayesian confidence interval (shortest or central interval depending on the option)

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
       - alpha : shape parameter > 0 for the prior distribution (fBeta_alpha)
       - beta  : shape parameter > 0 for the prior distribution (fBeta_beta)
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned

Note: In the case central confidence interval is calculated.
      when passed = 0 (or passed = total) the lower (or upper)
      interval values will be larger than 0 (or smaller than 1).

Calculation:

The posterior probability in bayesian statistics is given by:
P(#varepsilon |k,N) #propto L(#varepsilon|k,N) #times Prior(#varepsilon)
As an efficiency can be interpreted as probability of a positive outcome of
a Bernoullli trial the likelihood function is given by the binomial
distribution:
L(#varepsilon|k,N) = Binomial(N,k) #varepsilon ^{k} (1 - #varepsilon)^{N-k}
At the moment only beta distributions are supported as prior probabilities
of the efficiency (#scale[0.8]{B(#alpha,#beta)} is the beta function):
Prior(#varepsilon) = #frac{1}{B(#alpha,#beta)} #varepsilon ^{#alpha - 1} (1 - #varepsilon)^{#beta - 1}
The posterior probability is therefore again given by a beta distribution:
P(#varepsilon |k,N) #propto #varepsilon ^{k + #alpha - 1} (1 - #varepsilon)^{N - k + #beta - 1}
In case of central intervals
the lower boundary for the equal-tailed confidence interval is given by the
inverse cumulative (= quantile) function for the quantile #frac{1 - level}{2}.
The upper boundary for the equal-tailed confidence interval is given by the
inverse cumulative (= quantile) function for the quantile #frac{1 + level}{2}.
Hence it is the solution #varepsilon of the following equation:
I_{#varepsilon}(k + #alpha,N - k + #beta) = #frac{1}{norm} #int_{0}^{#varepsilon} dt t^{k + #alpha - 1} (1 - t)^{N - k + #beta - 1} =  #frac{1 #pm level}{2}
 In the case of shortest interval the minimum interval aorund the mode is found by minimizing the length of all intervals whith the
 given probability content. See TEfficiency::BetaShortestInterval
Double_t BetaCentralInterval(Double_t level, Double_t alpha, Double_t beta, Bool_t bUpper)
calculates the boundaries for a central confidence interval for a Beta distribution

Input: - level : confidence level
       -    a  : parameter > 0 for the beta distribution (for a posterior is passed + prior_alpha
       -    b  : parameter > 0 for the beta distribution (for a posterior is (total-passed) + prior_beta
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned

Bool_t BetaShortestInterval(Double_t level, Double_t alpha, Double_t beta, Double_t& lower, Double_t& upper)
calculates the boundaries for a shortest confidence interval for a Beta  distribution

Input: - level : confidence level
       -    a  : parameter > 0 for the beta distribution (for a posterior is passed + prior_alpha
       -    b  : parameter > 0 for the beta distribution (for a posterior is (total-passed) + prior_beta
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned


The lower/upper boundary are then obtained by finding the shortest interval of the beta distribbution
 contained the desired probability level.
 The length of all possible intervals is minimized in order to find the shortest one
Double_t BetaMean(Double_t alpha, Double_t beta)
 compute the mean (average) of the beta distribution

Input:    a  : parameter > 0 for the beta distribution (for a posterior is passed + prior_alpha
          b  : parameter > 0 for the beta distribution (for a posterior is (total-passed) + prior_beta

Double_t BetaMode(Double_t alpha, Double_t beta)
 compute the mode of the beta distribution

Input:    a  : parameter > 0 for the beta distribution (for a posterior is passed + prior_alpha
          b  : parameter > 0 for the beta distribution (for a posterior is (total-passed) + prior_beta

 note the mode is defined for a Beta(a,b) only if (a,b)>1 (a = passed+alpha; b = total-passed+beta)
 return then the following in case (a,b) < 1:
  if (a==b) return 0.5 (it is really undefined)
  if (a < b) return 0;
  if (a > b) return 1;
void Build(const char* name, const char* title)
building standard data structure of a TEfficiency object

Notes: - calls: SetName(name), SetTitle(title)
       - set the statistic option to the default (kFCP)
       - appends this object to the current directory
         SetDirectory(gDirectory)
Bool_t CheckBinning(const TH1& pass, const TH1& total)
checks binning for each axis

It is assumed that the passed histograms have the same dimension.
Bool_t CheckConsistency(const TH1& pass, const TH1& total, Option_t* opt = "")
checks the consistence of the given histograms

The histograms are considered as consistent if:
- both have the same dimension
- both have the same binning
- pass.GetBinContent(i) <= total.GetBinContent(i) for each bin i

Option: - w: The check for unit weights is skipped and therefore histograms
             filled with weights are accepted.
Bool_t CheckEntries(const TH1& pass, const TH1& total, Option_t* opt = "")
checks whether bin contents are compatible with binomial statistics

The following inequality has to be valid for each bin i:
 total.GetBinContent(i) >= pass.GetBinContent(i)

and the histogram have to be filled with unit weights.

Option: - w: Do not check for unit weights -> accept histograms filled with
             weights

Note: - It is assumed that both histograms have the same dimension and
        binning.
TGraphAsymmErrors * CreateGraph(Option_t* opt = "") const
 Create the graph used be painted (for dim=1 TEfficiency)
 The return object is managed by the caller
void FillGraph(TGraphAsymmErrors* graph, Option_t* opt) const
 Fill the graph to be painted with information from TEfficiency
 Internal metyhod called by TEfficiency::Paint or TEfficiency::CreateGraph
TH2 * CreateHistogram(Option_t* opt = "") const
 Create the histogram used to be painted (for dim=2 TEfficiency)
 The return object is managed by the caller
void FillHistogram(TH2* h2) const
 Fill the 2d histogram to be painted with information from TEfficiency 2D
 Internal metyhod called by TEfficiency::Paint or TEfficiency::CreatePaintingGraph
Double_t ClopperPearson(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
calculates the boundaries for the frequentist Clopper-Pearson interval

This interval is recommended by the PDG.

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned

calculation:

The lower boundary of the Clopper-Pearson interval is the "exact" inversion
of the test:

P(x #geq passed; total) = #frac{1 - level}{2}
The lower boundary is therfore given by the #frac{1 - level}{2} quantile
of the beta distribution.

The upper boundary of the Clopper-Pearson interval is the "exact" inversion
of the test:

P(x #leq passed; total) = #frac{1 - level}{2}
The upper boundary is therfore given by the #frac{1 + level}{2} quantile
of the beta distribution.

Note: The connection between the binomial distribution and the regularized
      incomplete beta function I_{#varepsilon}(#alpha,#beta) has been used.
Double_t Combine(Double_t& up, Double_t& low, Int_t n, const Int_t* pass, const Int_t* total, Double_t alpha, Double_t beta, Double_t level = 0.68300000000000005, const Double_t* w = 0, Option_t* opt = "")
calculates the combined efficiency and its uncertainties

This method does a bayesian combination of the given samples.

Input:
- up     : contains the upper limit of the confidence interval afterwards
- low    : contains the lower limit of the confidence interval afterwards
- n      : number of samples which are combined
- pass   : array of length n containing the number of passed events
- total  : array of length n containing the corresponding numbers of total
           events
- alpha  : shape parameters for the beta distribution as prior
- beta   : shape parameters for the beta distribution as prior
- level  : desired confidence level
- w      : weights for each sample; if not given, all samples get the weight 1
           The weights do not need to be normalized, since they are internally renormalized
           to the number of effective entries.
- options:

 + mode : The mode is returned instead of the mean of the posterior as best value
          When using the mode the shortest interval is also computed instead of the central one
 + shortest: compute shortest interval (done by default if mode option is set)
 + central: compute central interval (done by default if mode option is NOT set)


Calculation:
  1. The combined posterior distributions is calculated from the Bayes theorem assuming a common prior Beta distribution. It is easy to proof that the combined posterior is then:
  2. Begin_Latex(separator='=',align='rl') P_{comb}(#epsilon |{w_{i}}; {k_{i}}; {N_{i}}) = B(#epsilon, #sum_{i}{ w_{i} k_{i}} + #alpha, #sum_{i}{ w_{i}(n_{i}-k_{i})}+#beta) w_{i} = weight for each sample renormalized to the effective entries w^{'}_{i} = w_{i} #frac{ #sum_{i} {w_{i} } } { #sum_{i} {w_{i}^{2} } } End_Latex Begin_Html
  3. The estimated efficiency is the mode (or the mean) of the obtained posterior distribution
  4. 
    
  5. The boundaries of the confidence interval for a confidence level (1 - a) are given by the a/2 and 1-a/2 quantiles of the resulting cumulative distribution.
Example (uniform prior distribution):

output of htmldoc/MACRO_TEfficiency_Combine_2_2.gif
{
  TCanvas* c1 = new TCanvas("c1","",600,800);
  c1->Divide(1,2);
  c1->SetFillStyle(1001);
  c1->SetFillColor(kWhite);
  TF1* p1 = new TF1("p1","TMath::BetaDist(x,19,9)",0,1);
  TF1* p2 = new TF1("p2","TMath::BetaDist(x,4,8)",0,1);
  TF1* comb = new TF1("comb2","TMath::BetaDist(x,[0],[1])",0,1);
  double nrm = 1./(0.6*0.6+0.4*0.4); // weight normalization
  double a = 0.6*18.0 + 0.4*3.0 + 1.0;  // new alpha parameter of combined beta dist.
  double b = 0.6*10+0.4*7+1.0;  // new beta parameter of combined beta dist.
  comb->SetParameters(nrm*a ,nrm *b );
  TF1* const1 = new TF1("const1","0.05",0,1);
  TF1* const2 = new TF1("const2","0.95",0,1);
  p1->SetLineColor(kRed);
  p1->SetTitle("combined posteriors;#epsilon;P(#epsilon|k,N)");
  p2->SetLineColor(kBlue);
  comb->SetLineColor(kGreen+2);
  TLegend* leg1 = new TLegend(0.12,0.65,0.5,0.85);
  leg1->AddEntry(p1,"k1 = 18, N1 = 26","l");
  leg1->AddEntry(p2,"k2 = 3, N2 = 10","l");
  leg1->AddEntry(comb,"combined: p1 = 0.6, p2=0.4","l");
  c1->cd(1);
  comb->Draw();
  p1->Draw("same");
  p2->Draw("same");
  leg1->Draw("same");
  c1->cd(2);
  const1->SetLineWidth(1);
  const2->SetLineWidth(1);
  TGraph* gr = (TGraph*)comb->DrawIntegral();
  gr->SetTitle("cumulative function of combined posterior with boundaries for cl = 95%;#epsilon;CDF");
  const1->Draw("same");
  const2->Draw("same");
  c1->cd(0);
  return c1;
}
TGraphAsymmErrors* Combine(TCollection* pList, Option_t* opt = "", Int_t n = 0, const Double_t* w = 0)
combines a list of 1-dimensional TEfficiency objects

A TGraphAsymmErrors object is returned which contains the estimated
efficiency and its uncertainty for each bin.
If the combination fails, a zero pointer is returned.

At the moment the combining is only implemented for bayesian statistics.

Input:
- pList  : list containing TEfficiency objects which should be combined
           only one-dimensional efficiencies are taken into account
- options
 + s     : strict combining; only TEfficiency objects with the same beta
           prior and the flag kIsBayesian == true are combined
           If not specified the prior parameter of the first TEfficiency object is used
 + v     : verbose mode; print information about combining
 + cl=x  : set confidence level (0 < cl < 1). If not specified, the
           confidence level of the first TEfficiency object is used.
 + mode    Use mode of combined posterior as estimated value for the efficiency
 + shortest: compute shortest interval (done by default if mode option is set)
 + central: compute central interval (done by default if mode option is NOT set)

- n      : number of weights (has to be the number of one-dimensional
           TEfficiency objects in pList)
           If no weights are passed, the internal weights GetWeight() of
           the given TEfficiency objects are used.
- w      : array of length n with weights for each TEfficiency object in
           pList (w[0] correspond to pList->First ... w[n-1] -> pList->Last)
           The weights do not have to be normalised.

For each bin the calculation is done by the Combine(double&, double& ...) method.
Int_t DistancetoPrimitive(Int_t px, Int_t py)
 Compute distance from point px,py to a graph.

  Compute the closest distance of approach from point px,py to this line.
  The distance is computed in pixels units.

 Forward the call to the painted graph
void Draw(Option_t* opt = "")
draws the current TEfficiency object

options:
- 1-dimensional case: same options as TGraphAsymmErrors::Draw()
    but as default "AP" is used
- 2-dimensional case: same options as TH2::Draw()
- 3-dimensional case: not yet supported

 specific TEfficiency drawing options:
 - E0 - plot bins where the total number of passed events is zero
      (the error interval will be [0,1] )
void ExecuteEvent(Int_t event, Int_t px, Int_t py)
 Execute action corresponding to one event.

  This member function is called when the drawn class is clicked with the locator
  If Left button clicked on one of the line end points, this point
     follows the cursor until button is released.

  if Middle button clicked, the line is moved parallel to itself
     until the button is released.
 Forward the call to the underlying graph
void Fill(Bool_t bPassed, Double_t x, Double_t y = 0, Double_t z = 0)
This function is used for filling the two histograms.

Input: bPassed - flag whether the current event passed the selection
                 true: both histograms are filled
                 false: only the total histogram is filled
       x       - x value
       y       - y value (use default=0 for 1-D efficiencies)
       z       - z value (use default=0 for 2-D or 1-D efficiencies)
void FillWeighted(Bool_t bPassed, Double_t weight, Double_t x, Double_t y = 0, Double_t z = 0)
This function is used for filling the two histograms with a weight.

Input: bPassed - flag whether the current event passed the selection
                 true: both histograms are filled
                 false: only the total histogram is filled
       weight  - weight for the event
       x       - x value
       y       - y value (use default=0 for 1-D efficiencies)
       z       - z value (use default=0 for 2-D or 1-D efficiencies)

Note: - this function will call SetUseWeightedEvents if it was not called by the user before
Int_t FindFixBin(Double_t x, Double_t y = 0, Double_t z = 0) const
returns the global bin number containing the given values

Note: - values which belong to dimensions higher than the current dimension
        of the TEfficiency object are ignored (i.e. for 1-dimensional
        efficiencies only the x-value is considered)
Int_t Fit(TF1* f1, Option_t* opt = "")
fits the efficiency using the TBinomialEfficiencyFitter class

The resulting fit function is added to the list of associated functions.

Options: - "+": previous fitted functions in the list are kept, by default
                all functions in the list are deleted
         - for more fitting options see TBinomialEfficiencyFitter::Fit
TH1* GetCopyPassedHisto() const
returns a cloned version of fPassedHistogram

Notes: - The histogram is filled with unit weights. You might want to scale
         it with the global weight GetWeight().
       - The returned object is owned by the user who has to care about the
         deletion of the new TH1 object.
       - This histogram is by default NOT attached to the current directory
         to avoid duplication of data. If you want to store it automatically
         during the next TFile::Write() command, you have to attach it to
         the corresponding directory.

  TFile* pFile = new TFile("passed.root","update");
  TEfficiency* pEff = (TEfficiency*)gDirectory->Get("my_eff");
  TH1* copy = pEff->GetCopyPassedHisto();
  copy->SetDirectory(gDirectory);
  pFile->Write();
TH1* GetCopyTotalHisto() const
returns a cloned version of fTotalHistogram

Notes: - The histogram is filled with unit weights. You might want to scale
         it with the global weight GetWeight().
       - The returned object is owned by the user who has to care about the
         deletion of the new TH1 object.
       - This histogram is by default NOT attached to the current directory
         to avoid duplication of data. If you want to store it automatically
         during the next TFile::Write() command, you have to attach it to
         the corresponding directory.

  TFile* pFile = new TFile("total.root","update");
  TEfficiency* pEff = (TEfficiency*)gDirectory->Get("my_eff");
  TH1* copy = pEff->GetCopyTotalHisto();
  copy->SetDirectory(gDirectory);
  pFile->Write();
Int_t GetDimension() const
returns the dimension of the current TEfficiency object
Double_t GetEfficiency(Int_t bin) const
returns the efficiency in the given global bin

Note: - The estimated efficiency depends on the chosen statistic option:
        for frequentist ones:
        #hat{#varepsilon} = #frac{passed}{total}
        for bayesian ones the expectation value of the resulting posterior
        distribution is returned:
        #hat{#varepsilon} = #frac{passed + #alpha}{total + #alpha + #beta}
        If the bit kPosteriorMode is set (or the method TEfficiency::UsePosteriorMode() has been called ) the
        mode (most probable value) of the posterior is returned:
        #hat{#varepsilon} = #frac{passed + #alpha -1}{total + #alpha + #beta -2}

      - If the denominator is equal to 0, an efficiency of 0 is returned.
      - When  passed + #alpha < 1 or total - passed + #beta < 1 the above
        formula for the mode is not valid. In these cases values the estimated efficiency is 0 or 1.
Double_t GetEfficiencyErrorLow(Int_t bin) const
returns the lower error on the efficiency in the given global bin

The result depends on the current confidence level fConfLevel and the
chosen statistic option fStatisticOption. See SetStatisticOption(Int_t) for
more details.

Note: If the histograms are filled with weights, only bayesian methods and the
      normal approximation are supported.
Double_t GetEfficiencyErrorUp(Int_t bin) const
returns the upper error on the efficiency in the given global bin

The result depends on the current confidence level fConfLevel and the
chosen statistic option fStatisticOption. See SetStatisticOption(Int_t) for
more details.

Note: If the histograms are filled with weights, only bayesian methods and the
      normal approximation are supported.
Int_t GetGlobalBin(Int_t binx, Int_t biny = 0, Int_t binz = 0) const
returns the global bin number which can be used as argument for the
following functions:

 - GetEfficiency(bin), GetEfficiencyErrorLow(bin), GetEfficiencyErrorUp(bin)
 - SetPassedEvents(bin), SetTotalEvents(bin)

see TH1::GetBin() for conventions on numbering bins
Long64_t Merge(TCollection* list)
merges the TEfficiency objects in the given list to the given
TEfficiency object using the operator+=(TEfficiency&)

The merged result is stored in the current object. The statistic options and
the confidence level are taken from the current object.

This function should be used when all TEfficiency objects correspond to
the same process.

The new weight is set according to:
#frac{1}{w_{new}} = #sum_{i} \frac{1}{w_{i}}
Double_t Normal(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
returns the confidence limits for the efficiency supposing that the
efficiency follows a normal distribution with the rms below

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned

calculation:

#hat{#varepsilon} = #frac{passed}{total}
TEfficiency& operator+=(const TEfficiency& rhs)
adds the histograms of another TEfficiency object to current histograms

The statistic options and the confidence level remain unchanged.

fTotalHistogram += rhs.fTotalHistogram;
fPassedHistogram += rhs.fPassedHistogram;

calculates a new weight:
current weight of this TEfficiency object = w_{1}
weight of rhs = w_{2}
w_{new} = \frac{w_{1} \times w_{2}}{w_{1} + w_{2}}
TEfficiency& operator=(const TEfficiency& rhs)
assignment operator

The histograms, statistic option, confidence level, weight and paint styles
of rhs are copied to the this TEfficiency object.

Note: - The list of associated functions is not copied. After this
        operation the list of associated functions is empty.
void Paint(Option_t* opt)
paints this TEfficiency object

For details on the possible option see Draw(Option_t*)

 Note for 1D classes
 In 1D the TEfficiency uses a TGraphAsymmErrors for drawing
 The TGraph is created only the first time Paint is used. The user can manipulate the
 TGraph via the method TEfficiency::GetPaintedGraph()
 The TGraph creates behing an histogram for the axis. The histogram is created also only the first time.
 If the axis needs to be updated because in the meantime the class changed use this trick
 which will trigger a re-calculation of the axis of the graph
 TEfficiency::GetPaintedGraph()->Set(0)

 Note that in order to access the painted graph via GetPaintedGraph() you need either to call Paint or better
 gPad->Update();

void SavePrimitive(ostream& out, Option_t* opt = "")
have histograms fixed bins along each axis?
void SetBetaAlpha(Double_t alpha)
sets the shape parameter \alpha

The prior probability of the efficiency is given by the beta distribution:

f(\varepsilon;\alpha;\beta) = \frac{1}{B(\alpha,\beta)} \varepsilon^{\alpha-1} (1 - \varepsilon)^{\beta-1}

Note: - both shape parameters have to be positive (i.e. > 0)
void SetBetaBeta(Double_t beta)
sets the shape parameter \beta

The prior probability of the efficiency is given by the beta distribution:

f(\varepsilon;\alpha,\beta) = \frac{1}{B(\alpha,\beta)} \varepsilon^{\alpha-1} (1 - \varepsilon)^{\beta-1}

Note: - both shape parameters have to be positive (i.e. > 0)
void SetBetaBinParameters(Int_t bin, Double_t alpha, Double_t beta)
sets different  shape parameter \alpha and \beta
 for the prior distribution for each bin. By default the global parameter are used if they are not set
 for the specific bin
The prior probability of the efficiency is given by the beta distribution:

f(\varepsilon;\alpha;\beta) = \frac{1}{B(\alpha,\beta)} \varepsilon^{\alpha-1} (1 - \varepsilon)^{\beta-1}

Note: - both shape parameters have to be positive (i.e. > 0)
      - bin gives the global bin number (cf. GetGlobalBin)
Bool_t SetBins(Int_t nx, Double_t xmin, Double_t xmax)
 set the bins for the underlined passed and total histograms
 If the class have been already filled the previous contents will be lost
Bool_t SetBins(Int_t nx, const Double_t* xBins)
 set the bins for the underlined passed and total histograms
 If the class have been already filled the previous contents will be lost
Bool_t SetBins(Int_t nx, Double_t xmin, Double_t xmax, Int_t ny, Double_t ymin, Double_t ymax)
 set the bins for the underlined passed and total histograms
 If the class have been already filled the previous contents will be lost
Bool_t SetBins(Int_t nx, const Double_t* xBins, Int_t ny, const Double_t* yBins)
 set the bins for the underlined passed and total histograms
 If the class have been already filled the previous contents will be lost
Bool_t SetBins(Int_t nx, Double_t xmin, Double_t xmax, Int_t ny, Double_t ymin, Double_t ymax, Int_t nz, Double_t zmin, Double_t zmax)
 set the bins for the underlined passed and total histograms
 If the class have been already filled the previous contents will be lost
Bool_t SetBins(Int_t nx, const Double_t* xBins, Int_t ny, const Double_t* yBins, Int_t nz, const Double_t* zBins)
 set the bins for the underlined passed and total histograms
 If the class have been already filled the previous contents will be lost
void SetConfidenceLevel(Double_t level)
sets the confidence level (0 < level < 1)
 The default value is 1-sigma :~ 0.683
void SetDirectory(TDirectory* dir)
sets the directory holding this TEfficiency object

A reference to this TEfficiency object is removed from the current
directory (if it exists) and a new reference to this TEfficiency object is
added to the given directory.

Notes: - If the given directory is 0, the TEfficiency object does not
         belong to any directory and will not be written to file during the
         next TFile::Write() command.
void SetName(const char* name)
sets the name

Note: The names of the internal histograms are set to "name + _total" and
      "name + _passed" respectively.
Bool_t SetPassedEvents(Int_t bin, Int_t events)
sets the number of passed events in the given global bin

returns "true" if the number of passed events has been updated
otherwise "false" ist returned

Note: - requires: 0 <= events <= fTotalHistogram->GetBinContent(bin)
Bool_t SetPassedHistogram(const TH1& rPassed, Option_t* opt)
sets the histogram containing the passed events

The given histogram is cloned and stored internally as histogram containing
the passed events. The given histogram has to be consistent with the current
fTotalHistogram (see CheckConsistency(const TH1&,const TH1&)).
The method returns whether the fPassedHistogram has been replaced (true) or
not (false).

Note: The list of associated functions fFunctions is cleared.

Option: - "f": force the replacement without checking the consistency
               This can lead to inconsistent histograms and useless results
               or unexpected behaviour. But sometimes it might be the only
               way to change the histograms. If you use this option, you
               should ensure that the fTotalHistogram is replaced by a
               consistent one (with respect to rPassed) as well.
void SetStatisticOption(TEfficiency::EStatOption option)
sets the statistic option which affects the calculation of the confidence interval

Options:
- kFCP (=0)(default): using the Clopper-Pearson interval (recommended by PDG)
                      sets kIsBayesian = false
                      see also ClopperPearson
- kFNormal   (=1)   : using the normal approximation
                      sets kIsBayesian = false
                      see also Normal
- kFWilson   (=2)   : using the Wilson interval
                      sets kIsBayesian = false
                      see also Wilson
- kFAC       (=3)   : using the Agresti-Coull interval
                      sets kIsBayesian = false
                      see also AgrestiCoull
- kFFC       (=4)   : using the Feldman-Cousins frequentist method
                      sets kIsBayesian = false
                      see also FeldmanCousins
- kBJeffrey  (=5)   : using the Jeffrey interval
                      sets kIsBayesian = true, fBeta_alpha = 0.5 and fBeta_beta = 0.5
                      see also Bayesian
- kBUniform  (=6)   : using a uniform prior
                      sets kIsBayesian = true, fBeta_alpha = 1 and fBeta_beta = 1
                      see also Bayesian
- kBBayesian (=7)   : using a custom prior defined by fBeta_alpha and fBeta_beta
                      sets kIsBayesian = true
                      see also Bayesian
void SetTitle(const char* title)
sets the title

Notes: - The titles of the internal histograms are set to "title + (total)"
         or "title + (passed)" respectively.
       - It is possible to label the axis of the histograms as usual (see
         TH1::SetTitle).

Example: Setting the title to "My Efficiency" and label the axis

pEff->SetTitle("My Efficiency;x label;eff");
Bool_t SetTotalEvents(Int_t bin, Int_t events)
sets the number of total events in the given global bin

returns "true" if the number of total events has been updated
otherwise "false" ist returned

Note: - requires: fPassedHistogram->GetBinContent(bin) <= events
Bool_t SetTotalHistogram(const TH1& rTotal, Option_t* opt)
sets the histogram containing all events

The given histogram is cloned and stored internally as histogram containing
all events. The given histogram has to be consistent with the current
fPassedHistogram (see CheckConsistency(const TH1&,const TH1&)).
The method returns whether the fTotalHistogram has been replaced (true) or
not (false).

Note: The list of associated functions fFunctions is cleared.

Option: - "f": force the replacement without checking the consistency
               This can lead to inconsistent histograms and useless results
               or unexpected behaviour. But sometimes it might be the only
               way to change the histograms. If you use this option, you
               should ensure that the fPassedHistogram is replaced by a
               consistent one (with respect to rTotal) as well.
void SetWeight(Double_t weight)
sets the global weight for this TEfficiency object

Note: - weight has to be positive ( > 0)
Double_t Wilson(Int_t total, Int_t passed, Double_t level, Bool_t bUpper)
calculates the boundaries for the frequentist Wilson interval

Input: - total : number of total events
       - passed: 0 <= number of passed events <= total
       - level : confidence level
       - bUpper: true  - upper boundary is returned
                 false - lower boundary is returned

calculation:

#alpha = 1 - #frac{level}{2}
void Add(const TEfficiency& rEff)
{*this += rEff;}
Double_t GetBetaAlpha(Int_t bin = -1) const
 use trick of -1 to return global parameters
{return (fBeta_bin_params.size() > (UInt_t)bin) ? fBeta_bin_params[bin].first : fBeta_alpha;}
Double_t GetBetaBeta(Int_t bin = -1) const
{return (fBeta_bin_params.size() > (UInt_t)bin) ? fBeta_bin_params[bin].second : fBeta_beta;}
TDirectory* GetDirectory() const
{return fDirectory;}
TH2* GetPaintedHistogram() const
{ return fPaintHisto; }
const TH1* GetPassedHistogram() const
EStatOption GetStatisticOption() const
const TH1* GetTotalHistogram() const
{return fTotalHistogram;}
Double_t GetWeight() const
{return fWeight;}
void SetPosteriorAverage(Bool_t on = true)