#include <LocalPolyFilter1D.hh>

Inheritance diagram for npstat::LocalPolyFilter1D:

Public Member Functions
	LocalPolyFilter1D (const double *taper, unsigned maxDegree, const AbsFilter1DBuilder &filterBuilder, unsigned dataLen)

	LocalPolyFilter1D (const LocalPolyFilter1D &)

LocalPolyFilter1D &	operator= (const LocalPolyFilter1D &)

bool	operator== (const LocalPolyFilter1D &r) const

bool	operator!= (const LocalPolyFilter1D &r) const

double	taper (unsigned degree) const

unsigned	maxDegree () const

unsigned	dataLen () const

const std::vector< double > &	getBandwidthFactors () const

double	selfContribution (unsigned binNumber) const

const PolyFilter1D &	getFilter (unsigned binNumber) const

template<typename Tin , typename Tout >
void	filter (const Tin in, unsigned dataLen, Tout out) const

template<typename Tin , typename Tout >
void	convolve (const Tin in, unsigned dataLen, Tout out) const

Matrix< double >	getFilterMatrix () const

CPP11_auto_ptr< LocalPolyFilter1D >	doublyStochasticFilter (double tolerance, unsigned maxIterations) const

CPP11_auto_ptr< LocalPolyFilter1D >	eigenGroomedFilter () const

virtual gs::ClassId	classId () const

virtual bool	write (std::ostream &os) const

Static Public Member Functions
static const char *	classname ()

static unsigned	version ()

static LocalPolyFilter1D *	read (const gs::ClassId &id, std::istream &in)

Detailed Description

This class performs local polynomial filtering in one dimension

Constructor & Destructor Documentation

◆ LocalPolyFilter1D()

npstat::LocalPolyFilter1D::LocalPolyFilter1D	(	const double *	taper,
		unsigned	maxDegree,
		const AbsFilter1DBuilder &	filterBuilder,
		unsigned	dataLen
	)

Main constructor. The arguments are as follows:

taper – Damping factors for each polynomial degree (starting with the 0th order term). This can be NULL in which case it is assumed that all factors are 1.

maxDegree – Maximum degree of the polynomials. The length of the "taper" array (if not NULL) must be equal to maxDegree + 1. Note that, far away from the boundaries (where the situation is symmetric) the same filter will be produced using the same taper with an even degree N and with an odd degree N+1. Near the boundaries the filter coefficients will, of course, differ in these two cases.

filterBuilder – An instance of a class which actually builds the filters when requested. This builder is used only inside the constructor.

dataLen – The length of the data arrays which will be used with this filter (this info is needed in order to take into account the boundary effects).

Member Function Documentation

◆ convolve()

template<typename Tin , typename Tout >

void npstat::LocalPolyFilter1D::convolve	(	const Tin *	in,
		unsigned	dataLen,
		Tout *	out
	)		const

A diffent filtering method in which the shapes of the kernels are determined by the positions of the "sources" (i.e., sample points) instead of the positions at which the density (or response) is estimated.

◆ dataLen()

unsigned npstat::LocalPolyFilter1D::dataLen ( ) const

inlinevirtual

Length of the data array

Implements npstat::AbsPolyFilter1D.

◆ doublyStochasticFilter()

CPP11_auto_ptr<LocalPolyFilter1D> npstat::LocalPolyFilter1D::doublyStochasticFilter	(	double	tolerance,
		unsigned	maxIterations
	)		const

Generate a doubly stochastic filter out of this one. Such filters are useful for sequential copula smoothing. NULL pointer will be returned in case the requested margin tolerance is positive and can not be reached within the number of iterations allowed.

◆ eigenGroomedFilter()

CPP11_auto_ptr<LocalPolyFilter1D> npstat::LocalPolyFilter1D::eigenGroomedFilter ( ) const

An experimental filter with an adjusted eigenspectrum

◆ filter()

template<typename Tin , typename Tout >

void npstat::LocalPolyFilter1D::filter	(	const Tin *	in,
		unsigned	dataLen,
		Tout *	out
	)		const

This method performs the filtering. "dataLen", which is the length of both "in" and "out" arrays, must be the same as the one in the constructor.

◆ getFilter()

const PolyFilter1D& npstat::LocalPolyFilter1D::getFilter ( unsigned binNumber ) const

Get the filter coefficients for the given bin

◆ getFilterMatrix()

Matrix<double> npstat::LocalPolyFilter1D::getFilterMatrix ( ) const

Generate the complete (non-sparse) representation of the filter. It will be a generalized stochastic matrix (each row sums to 1).

◆ selfContribution()

double npstat::LocalPolyFilter1D::selfContribution ( unsigned binNumber ) const

virtual

Self contribution needed for cross-validation