w2img::LinearRBF::KMeansStrategy Class Reference

Inheritance diagram for w2img::LinearRBF::KMeansStrategy:

List of all members.

---

Detailed Description

Uses k-means clustering to choose the best k centers based on the training data.

Public Member Functions
virtual	~KMeansStrategy ()
virtual void	update (const std::vector< LinearRBF::Input > &inputs, const std::vector< double > &outputs, double &influence_radius, std::vector< LinearRBF::Input > &centers)
	sub-classes implement this method to update the two pieces of data.
	KMeansStrategy (size_t k_in)
	how many centers are preferred?
void	update (const std::vector< LinearRBF::Input > &inputs, const std::vector< double > &outputs)
	update based on new training data.
double	getInfluenceRadius () const
const std::vector< LinearRBF::Input > &	getCenters () const
Static Public Member Functions
static double	findDistanceSquared (const std::vector< double > &a, const std::vector< double > &b)
static double	findAverageDistance (const std::vector< LinearRBF::Input > &)
Protected Member Functions
void	init (const LinearRBF::Input &minbounds, const LinearRBF::Input &maxbounds, const std::vector< size_t > &number)
Protected Attributes
std::vector< size_t >	number
double	influenceRadius
std::vector< LinearRBF::Input >	centers

---

Constructor & Destructor Documentation

virtual w2img::LinearRBF::KMeansStrategy::~KMeansStrategy

(

)

[inline, virtual]

w2img::LinearRBF::KMeansStrategy::KMeansStrategy

(

size_t

k_in

)

[inline]

how many centers are preferred?

---

Member Function Documentation

static double w2img::LinearRBF::Strategy::findAverageDistance

(

const std::vector< LinearRBF::Input > &

)

[static, inherited]

static double w2img::LinearRBF::Strategy::findDistanceSquared	(	const std::vector< double > &	a,
		const std::vector< double > &	b
	)			[static, inherited]

const std::vector<LinearRBF::Input>& w2img::LinearRBF::Strategy::getCenters

(

)

const [inline, inherited]

double w2img::LinearRBF::Strategy::getInfluenceRadius

(

)

const [inline, inherited]

void w2img::LinearRBF::UniformStrategy::init	(	const LinearRBF::Input &	minbounds,
		const LinearRBF::Input &	maxbounds,
		const std::vector< size_t > &	number
	)			[protected, inherited]

void w2img::LinearRBF::Strategy::update	(	const std::vector< LinearRBF::Input > &	inputs,
		const std::vector< double > &	outputs
	)			[inherited]

update based on new training data.

virtual void w2img::LinearRBF::KMeansStrategy::update	(	const std::vector< LinearRBF::Input > &	inputs,
		const std::vector< double > &	outputs,
		double &	influence_radius,
		std::vector< LinearRBF::Input > &	centers
	)			[virtual]

sub-classes implement this method to update the two pieces of data.

Reimplemented from w2img::LinearRBF::DataDrivenUniformStrategy.

---

Member Data Documentation

std::vector<LinearRBF::Input> w2img::LinearRBF::Strategy::centers [protected, inherited]

double w2img::LinearRBF::Strategy::influenceRadius [protected, inherited]

std::vector<size_t> w2img::LinearRBF::DataDrivenUniformStrategy::number [protected, inherited]

---