Docs | All | Real | Category | PDF | DataSet | Plot | Container | Misc | Aux | User

class RooHistError

    private:

                   RooHistError()
            Bool_t getPoissonIntervalCalc(Int_t n, Double_t& mu1, Double_t& mu2, Double_t nSigma = 1) const
          Double_t seek(const RooAbsFunc& f, Double_t startAt, Double_t step, Double_t value) const

    public:

                                 RooHistError(const RooHistError&)
                         virtual ~RooHistError()
                  static TClass* Class()
              static RooAbsFunc* createBinomialSum(Int_t n, Int_t m)
              static RooAbsFunc* createPoissonSum(Int_t n)
                          Bool_t getBinomialInterval(Int_t n, Int_t m, Double_t& a1, Double_t& a2, Double_t nSigma = 1) const
                          Bool_t getInterval(const RooAbsFunc* Qu, const RooAbsFunc* Ql, Double_t pointEstimate, Double_t stepSize, Double_t& lo, Double_t& hi, Double_t nSigma) const
                          Bool_t getPoissonInterval(Int_t n, Double_t& mu1, Double_t& mu2, Double_t nSigma = 1) const
      static const RooHistError& instance() const
                 virtual TClass* IsA() const
                   RooHistError& operator=(const RooHistError&)
                    virtual void ShowMembers(TMemberInspector& insp, char* parent)
                    virtual void Streamer(TBuffer& b)
                            void StreamerNVirtual(TBuffer& b)

Data Members

    private:

      Double_t _poissonLoLUT[1000]  
      Double_t _poissonHiLUT[1000]  

Class Description

 RooHistError is a singleton class used to calculate the error bars
 for each bin of a RooHist object. Errors are calculated by integrating
 a specified area of a Poisson or Binomail error distribution.

 Return a reference to a singleton object that is created the
 first time this method is called. Only one object will be
 constructed per ROOT session.

 Construct our singleton object.

 Use lookup table for most common cases

 Calculate a confidence interval for the expected number of events given n
 observed (unweighted) events. The interval will contain the same probability
 as nSigma of a Gaussian. Uses a central interval unless this does not enclose
 the point estimate n (ie, for small n) in which case the interval is adjusted
 to start at n.

 sanity checks

 Calculate a confidence interval using the cummulative functions provided.
 The interval will be "central" when both cummulative functions are provided,
 unless this would exclude the pointEstimate, in which case a one-sided interval
 pinned at the point estimate is returned instead.

 Scan f(x)-value until it changes sign. Start at the specified point and take constant
 steps of the specified size. Give up after 1000 steps.

Inline Functions

                 void ~RooHistError()
          RooAbsFunc* createPoissonSum(Int_t n)
          RooAbsFunc* createBinomialSum(Int_t n, Int_t m)
              TClass* Class()
              TClass* IsA() const
                 void ShowMembers(TMemberInspector& insp, char* parent)
                 void Streamer(TBuffer& b)
                 void StreamerNVirtual(TBuffer& b)
         RooHistError RooHistError(const RooHistError&)
        RooHistError& operator=(const RooHistError&)

Page maintained by Wouter Verkerke and David Kirkby