npstat::AbsCGF1D Struct Reference

Inheritance diagram for npstat::AbsCGF1D:

Public Member Functions
virtual AbsCGF1D *	clone () const =0
virtual AbsCGF1D *	shiftAndScale (double mu, double sigma) const =0
virtual double	smin () const =0
virtual double	smax () const =0
double	operator() (double s) const
virtual double	derivative (unsigned order, double s) const =0
virtual double	saddlepoint (double x) const
virtual double	convexLimit (double slimit, double stepSize) const
double	standardizedCumulant (unsigned order, double s) const
double	saddlepointDensityApprox (unsigned order, double x) const
double	lugannaniRiceSFApprox (unsigned nTerms, double x) const
bool	operator== (const AbsCGF1D &r) const
bool	operator!= (const AbsCGF1D &r) const

Protected Member Functions
virtual bool	isEqual (const AbsCGF1D &) const =0

Member Function Documentation

clone()

virtual AbsCGF1D* npstat::AbsCGF1D::clone ( ) const

pure virtual

Virtual copy constructor

Implemented in npstat::SeriesCGF1D.

convexLimit()

virtual double npstat::AbsCGF1D::convexLimit ( double  slimit, double  stepSize  ) const

virtual

Search empirically for the maximum value of s for which this CGF remains convex. This function should search for the closest to 0 root of the CGF second derivative on the interval [0, slimit]. It should return 0.0 if no such root was found.

derivative()

virtual double npstat::AbsCGF1D::derivative ( unsigned  order, double  s  ) const

pure virtual

CGF derivatives. Get cum[order] by setting s to 0.

Implemented in npstat::SeriesCGF1D.

isEqual()

virtual bool npstat::AbsCGF1D::isEqual ( const AbsCGF1D &  ) const

protectedpure virtual

Comparison for equality. To be implemented by derived classes.

Implemented in npstat::SeriesCGF1D.

lugannaniRiceSFApprox()

double npstat::AbsCGF1D::lugannaniRiceSFApprox	(	unsigned	nTerms,
		double	x
	)		const

Lugannani-Rice approximation for the survival function. The approximation employs a series which looks like c0/N^{1/2} + c1/N^{3/2} + c2/N^{5/2} + ... The parameter "nTerms" specifies the number of terms in the series to use and, thereby, the approximation order.

operator!=()

bool npstat::AbsCGF1D::operator!= ( const AbsCGF1D &  r ) const

inline

Logical negation of operator==

operator()()

double npstat::AbsCGF1D::operator() ( double  s ) const

inline

CGF value

operator==()

bool npstat::AbsCGF1D::operator== ( const AbsCGF1D &  r ) const

inline

Derived classes should not implement "operator==", implement "isEqual" instead

saddlepoint()

virtual double npstat::AbsCGF1D::saddlepoint ( double  x ) const

virtual

Solution of the saddlepoint equation CGF'(s) == x

saddlepointDensityApprox()

double npstat::AbsCGF1D::saddlepointDensityApprox	(	unsigned	order,
		double	x
	)		const

Saddlepoint density approximation. Parameter "order" is the approximation order (e.g., order = 2 means O(N^-2)). This parameter must be positive. The raw, unnormalized approximation is returned.

shiftAndScale()

virtual AbsCGF1D* npstat::AbsCGF1D::shiftAndScale ( double  mu, double  sigma  ) const

pure virtual

CGF for the variable (X - mu)/sigma. Created on the heap and must be later deleted by the user.

Implemented in npstat::SeriesCGF1D.

smax()

virtual double npstat::AbsCGF1D::smax ( ) const

pure virtual

Supremum of the support

Implemented in npstat::SeriesCGF1D.

smin()

virtual double npstat::AbsCGF1D::smin ( ) const

pure virtual

Infimum of the support

Implemented in npstat::SeriesCGF1D.

standardizedCumulant()

double npstat::AbsCGF1D::standardizedCumulant	(	unsigned	order,
		double	s
	)		const

Standardized cumulant for s-tilted density

---

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

• stat/AbsCGF1D.hh