npstat is hosted by Hepforge, IPPP Durham
NPStat  5.10.0
npsi::MinuitLOrPEBgCVFcn1D< Numeric, NumIn > Class Template Reference
Inheritance diagram for npsi::MinuitLOrPEBgCVFcn1D< Numeric, NumIn >:

Public Member Functions

 MinuitLOrPEBgCVFcn1D (const npstat::HistoND< Numeric > &histo, npstat::AbsSymbetaFilterProvider &fbuilder, const npstat::AbsDistribution1D &signal, const double signalFraction, const unsigned nIntegrationPoints, const NumIn *initialApproximation, const unsigned lenApproximation, const int symmetricBetaPower, const double minimumBgFeatureSize, const double effectiveNumBgEvents, const double convergenceEpsilon, const unsigned maxIterations, const npstat::BoundaryHandling &bm, const bool useLeastSquaresCV, const bool verbose, const bool bandwidthIsAbsolute, const unsigned cvmode, const double regularizationParameter, const double minlog, const double up)
 
unsigned long callCount () const
 
npstat::AbsSymbetaFilterProvidergetFilterProvider () const
 
double getActualBandwidth (const double bwParameter, const double maxDegree) const
 
virtual double operator() (const std::vector< double > &x) const
 
double Up () const
 

Constructor & Destructor Documentation

◆ MinuitLOrPEBgCVFcn1D()

template<class Numeric , class NumIn = double>
npsi::MinuitLOrPEBgCVFcn1D< Numeric, NumIn >::MinuitLOrPEBgCVFcn1D ( const npstat::HistoND< Numeric > &  histo,
npstat::AbsSymbetaFilterProvider fbuilder,
const npstat::AbsDistribution1D signal,
const double  signalFraction,
const unsigned  nIntegrationPoints,
const NumIn *  initialApproximation,
const unsigned  lenApproximation,
const int  symmetricBetaPower,
const double  minimumBgFeatureSize,
const double  effectiveNumBgEvents,
const double  convergenceEpsilon,
const unsigned  maxIterations,
const npstat::BoundaryHandling bm,
const bool  useLeastSquaresCV,
const bool  verbose,
const bool  bandwidthIsAbsolute,
const unsigned  cvmode,
const double  regularizationParameter,
const double  minlog,
const double  up 
)
inline

Most arguments of this constructor are the same as the arguments of the npstat::lorpeBackground1D function. See the description of that function for more detail. The two arguments not of this kind are:

up – This parameter regulates Minuit convergence. See Minuit manual for details.

minlog – Infrequently, LOrPE creates a density estimate which is exactly 0 for some histogram bin which is not empty; the value of this parameter limits the contribution of such bins into the overall likelihood.

An internal copy of "initialApproximation" will be made if such an approximation is provided (initialApproximation can also be specified as NULL to use the uniform background density as a starting point for iterations).

This class will not assume ownership of any references or pointers

Member Function Documentation

◆ operator()()

template<class Numeric , class NumIn = double>
virtual double npsi::MinuitLOrPEBgCVFcn1D< Numeric, NumIn >::operator() ( const std::vector< double > &  x) const
inlinevirtual

This method returns either the negative log likelihood or the cross validation approximation of MISE.

It is assumed that bandwidth is the first parameter while degree of the LOrPE polynomial is the second.


The documentation for this class was generated from the following file: