lti::sammonsMapping::parameters Class Reference

the parameters for the class sammonsMapping More...

#include <ltiSammonsMapping.h>

Inheritance diagram for lti::sammonsMapping::parameters:

[legend]

Collaboration diagram for lti::sammonsMapping::parameters:

[legend]

List of all members.

Public Types
enum	eInit { Random, PCA }
enum	eSearch { Steepest, Gradient, Momentum }
Public Member Functions
	parameters ()
	parameters (const parameters &other)
	~parameters ()
const char *	getTypeName () const
parameters &	copy (const parameters &other)
parameters &	operator= (const parameters &other)
virtual functor::parameters *	clone () const
virtual bool	write (ioHandler &handler, const bool complete=true) const
virtual bool	read (ioHandler &handler, const bool complete=true)
Public Attributes
int	dimensions
int	steps
double	errorThresh
double	alpha
eInit	initType
dvector	initBox
eSearch	searchType
double	mu

Detailed Description

the parameters for the class sammonsMapping

Member Enumeration Documentation

enum lti::sammonsMapping::parameters::eInit

Initialization for the lower dimensional image space.

The Random version uses initBox for the size of the hyper cube the values are drawn from. The PCA version was suggested by Sammon for real-world problems. It uses the projection of the data points into the image space as initialization.

Enumerator:

Random	choose initial values randomly from hyper cube initBox
PCA	initial values are projection of the data by PCA

enum lti::sammonsMapping::parameters::eSearch

Different methods for seeking the minimum of the error:.

Steepest uses steepest descent, i.e. the gradient is divided by the norm of the second derivative.
Gradient uses regular gradient descent
Momentum uses gradient descent with momentum. The momentum is set by the parameter mu.

Enumerator:

Steepest	steepest descent
Gradient	gradient descent
Momentum	gradient descent with momentum

Constructor & Destructor Documentation

lti::sammonsMapping::parameters::parameters ( )

default constructor

Reimplemented from lti::functor::parameters.

lti::sammonsMapping::parameters::parameters ( const parameters & other )

copy constructor

Parameters:

other

the parameters object to be copied

lti::sammonsMapping::parameters::~parameters ( ) [virtual]

destructor

Reimplemented from lti::functor::parameters.

Member Function Documentation

virtual functor::parameters* lti::sammonsMapping::parameters::clone ( ) const [virtual]

returns a pointer to a clone of the parameters

Implements lti::functor::parameters.

parameters& lti::sammonsMapping::parameters::copy ( const parameters & other )

copy the contents of a parameters object

Parameters:

other

the parameters object to be copied

Returns:: a reference to this parameters object

const char* lti::sammonsMapping::parameters::getTypeName ( ) const [virtual]

returns name of this type

Reimplemented from lti::functor::parameters.

parameters& lti::sammonsMapping::parameters::operator= ( const parameters & other )

copy the contents of a parameters object

Parameters:

other

the parameters object to be copied

Returns:: a reference to this parameters object

virtual bool lti::sammonsMapping::parameters::read	(	ioHandler &	handler,
		const bool	complete = `true`
	)			`[virtual]`

read the parameters from the given ioHandler

Parameters:

	handler	the ioHandler to be used
	complete	if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written.

Returns:: true if write was successful

Reimplemented from lti::functor::parameters.

virtual bool lti::sammonsMapping::parameters::write	(	ioHandler &	handler,
		const bool	complete = `true`
	)			const `[virtual]`

write the parameters in the given ioHandler

Parameters:

	handler	the ioHandler to be used
	complete	if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written.

Returns:: true if write was successful

Reimplemented from lti::functor::parameters.

Member Data Documentation

double lti::sammonsMapping::parameters::alpha

A 'learn' rate or step size.

According to Kohonen it should be somewhere between 0.3 and 0.4. The default is 0.35

int lti::sammonsMapping::parameters::dimensions

The number of dimensions of the output space.

Since Sammon's Mapping is mostly used for visualization the dafault is 2.

double lti::sammonsMapping::parameters::errorThresh

Threshold for the error of the mapping.

By default not used, thus 0.

dvector lti::sammonsMapping::parameters::initBox

Hyper cube random init values are chosen from in case initType is Random.

Each dimension ranges from 0 to the given value. Take care to choose the same dimesionality for initBox as the value of dimensions. Default is a unit cube.

eInit lti::sammonsMapping::parameters::initType

Sets the initType.

Default is PCA.

double lti::sammonsMapping::parameters::mu

The momentum.

Used if searchType is set to Momentum. Default is 0.1

eSearch lti::sammonsMapping::parameters::searchType

Sets the type of search method used to minimize the error.

Default is Steepest.

int lti::sammonsMapping::parameters::steps

Number of steps.

Default is 200.