#include <kernel.h>

Inheritance diagram for Kernel:

Public Member Functions
	Kernel (void)

	Kernel (double stdev)

virtual	~Kernel (void)

double	operator() (double x) const

virtual double	value (double x) const

double	scale (void) const

virtual void	setScale (double scale)

virtual double	mean (void) const

virtual double	stdev (void) const

virtual void	setStdev (double stdev)

virtual double	max (void) const

virtual double	left (void) const

virtual double	right (void) const

Detailed Description

Base class of all kernel functions.

Author: Jan Benda

Version: 1.0 A Kernel is a function y=f(x) which is concentrated between x=left() and x=right(), i.e. f(x) is zero (or close to zero) for x less than left() and for x greater than right(). The integral over the kernel is unity:
$\int_{-\infty}^{+\infty} f(x) dx = 1$
The width of the kernel can be controlled by a scale factor s, $f(x) \rightarrow f(x/s)$ , via setScale() or by setting the standard deviation of the kernel via setStdev(). The actual scale factor is returned by scale() and the corresponding standard deviation by stdev(). Note that the value returned by stdev() usually differs from scale(). The mean of the kernel is mean(). The maximum value of the kernel is max().

Constructor & Destructor Documentation

Kernel ( void )

inline

Constructs a default kernel with its scale factor set to one.

Kernel ( double stdev )

inline

Constructs a kernel with standard deviation stdev.

References Kernel::setStdev().

virtual ~Kernel ( void )

inlinevirtual

Destructor.

Member Function Documentation

double operator() ( double x ) const

inline

Returns the value of the kernel at position x. This operator is NOT virtual in order to speed up computation when using templates. If you need a virtual function as an interface, implement value().

References Kernel::value().

Referenced by Kernel::value().

double value ( double x ) const

virtual

Returns the value of the kernel at position x. This function IS virtual and might slow things down. If you need a NON virtual function as an interface, implement operator().

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, TriangularKernel, and RectKernel.

References Kernel::operator()().

Referenced by EventData::addCyclicRate(), EventData::addRate(), Kernel::max(), Kernel::mean(), Kernel::operator()(), EventData::rate(), and Kernel::stdev().

double scale ( void ) const

inline

Return the scale factor of the kernel (the width).

Referenced by GammaKernel::GammaKernel(), GaussKernel::GaussKernel(), GaussKernel::left(), GammaKernel::mean(), RectKernel::operator()(), TriangularKernel::operator()(), GaussKernel::operator()(), GammaKernel::operator()(), GaussKernel::right(), Kernel::setScale(), RectKernel::setScale(), GaussKernel::setScale(), GammaKernel::setScale(), RectKernel::stdev(), TriangularKernel::stdev(), GaussKernel::stdev(), and GammaKernel::stdev().

void setScale ( double scale )

virtual

Set the scale factor (width) of the kernel to scale.

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, TriangularKernel, and RectKernel.

References Kernel::scale().

Referenced by RectKernel::setScale(), TriangularKernel::setScale(), GaussKernel::setScale(), GammaKernel::setScale(), Kernel::setStdev(), and EpanechnikovKernel::setStdev().

double mean ( void ) const

virtual

Return the mean of the kernel. This default implementation calculates the mean numerically from 400 points between left() and right().

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, TriangularKernel, and RectKernel.

References Kernel::left(), Kernel::right(), and Kernel::value().

double stdev ( void ) const

virtual

Return the standard deviation of the kernel, which does not have to equal the scale(). This default implementation calculates the standard deviation numerically from 400 points between left() and right().

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, TriangularKernel, and RectKernel.

References Kernel::left(), Kernel::right(), relacs::sqrt(), and Kernel::value().

void setStdev ( double stdev )

virtual

Set the standard deviation of the kernel to stdev. This default implementation sets the scale() equal to stdev.

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, TriangularKernel, and RectKernel.

References Kernel::setScale().

Referenced by Kernel::Kernel().

double max ( void ) const

virtual

Return the maximum value of the kernel. This default implementation searches for the maximum of 400 computed function values between left() and right().

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, TriangularKernel, and RectKernel.

References Kernel::left(), Kernel::right(), and Kernel::value().

double left ( void ) const

virtual

The leftmost extension of the kernel. For x values less than this value, the value of the kernel is negligible. Returns minus half the scale factor.

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, and TriangularKernel.

Referenced by EventData::addCyclicRate(), EventData::addRate(), Kernel::max(), Kernel::mean(), EventData::rate(), and Kernel::stdev().

double right ( void ) const

virtual

The rightmost extension of the kernel. For x values greater than this value, the value of the kernel is negligible. Returns half the scale factor.

Reimplemented in GammaKernel, GaussKernel, EpanechnikovKernel, and TriangularKernel.

Referenced by EventData::addCyclicRate(), EventData::addRate(), Kernel::max(), Kernel::mean(), EventData::rate(), and Kernel::stdev().

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

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation