IncrementalSGD< T > Class Template Reference

This class presents generic IncrementalSGD and two kind of sgd that extends this class which are TrueSGD and ReservoirSGD. More...

#include <incremental_sgd.h>

Inheritance diagram for IncrementalSGD< T >:

List of all members.

Public Member Functions
	IncrementalSGD (int dim, double l)
	Constructor that initializes IncrementalSgd.
void	resetModel ()
void	setModel (model m)
model &	getModel ()
	Returns model.
virtual bool	addExample (int y, T ex)=0
	There are many algorithms that do not train when we an add example the return value indicates whether or not the model has changed as a result of this update.
virtual	~IncrementalSGD ()
	Desctructor.
bool	classifyExample (T f)
	Classifies given feature vector of the entity with the current model and returns class of the example.
Static Public Member Functions
static double	classifyExample (struct model &m, T x)
	Classifies given feature vector of the entity with the given model and returns distance of the entity to the model.
Protected Member Functions
void	oneStep (double y, T ex)
	This takes one step with this example.
Protected Attributes
double	lambda
	step size
model	_m
	model

---

Detailed Description

template<class T>
class IncrementalSGD< T >

This class presents generic IncrementalSGD and two kind of sgd that extends this class which are TrueSGD and ReservoirSGD.

---

Constructor & Destructor Documentation

template<class T>

IncrementalSGD< T >::IncrementalSGD	(	int	dim,
		double	l
	)

Constructor that initializes IncrementalSgd.

Parameters:

[in]	dim	dimension of the model
[in]	l	lambda

template<class T>

virtual IncrementalSGD< T >::~IncrementalSGD

(

)

[inline, virtual]

Desctructor.

---

Member Function Documentation

template<class T>

virtual bool IncrementalSGD< T >::addExample	(	int	y,
		T	ex
	)			[pure virtual]

There are many algorithms that do not train when we an add example the return value indicates whether or not the model has changed as a result of this update.

Parameters:

[in]	y	class of the given example
[in]	ex	feature vector of the given example

Returns:

whether model changed or not

Implemented in TrueSGD< T >, and ReservoirSGD< T >.

template<class T>

static double IncrementalSGD< T >::classifyExample	(	struct model &	m,
		T	x
	)			[inline, static]

Classifies given feature vector of the entity with the given model and returns distance of the entity to the model.

Parameters:

[in]	m	model to classify the entity
[in]	x	feature vector of the entity

Returns:

distance of the entity to the model

template<class T>

bool IncrementalSGD< T >::classifyExample

(

T

f

)

Classifies given feature vector of the entity with the current model and returns class of the example.

Parameters:

[in]

f

feature vector of the entity

Returns:

class of the example

template<class T>

struct model& IncrementalSGD< T >::getModel

(

)

[inline]

Returns model.

Returns:

model

template<class T>

void IncrementalSGD< T >::oneStep	(	double	y,
		T	ex
	)			[protected]

This takes one step with this example.

Parameters:

[in]	y	class of the example
[in]	ex	feature vector of the example

template<class T>

void IncrementalSGD< T >::resetModel

(

)

template<class T>

void IncrementalSGD< T >::setModel

(

model

m

)

[inline]

---

Member Data Documentation

template<class T>

struct model IncrementalSGD< T >::_m [protected]

model

template<class T>

double IncrementalSGD< T >::lambda [protected]

step size

---

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

• incremental_sgd.h
• incremental_sgd.hxx

---