SvmSgd Class Reference

Class for maintaining svm model. More...

#include <sgd_model.h>

List of all members.

Public Member Functions
	SvmSgd (int dim, double lambda)
	SvmSgd constructor.
void	scale ()
	Scales model with wscale.
model &	get_model ()
	Retrieves model learnt.
void	test_example (char *newfeature)
	Tests given example in svm format and check whether prediction is correct or not.
double	getAccuracy ()
	Returns accuracy on test data.
bool	isExampleCorrectlyClassified (std::string newfeature)
	Checks whether given example is classified correctly or not.
Private Attributes
int	_dim
model	_current
int	ex
	number of predictions
int	match
	number of correct prediction during testing
Incremental_SGD *	sgd
	Updates old model.

---

Detailed Description

Class for maintaining svm model.

---

Constructor & Destructor Documentation

SvmSgd::SvmSgd	(	int	dim,
		double	lambda
	)

SvmSgd constructor.

Parameters:

[in]	dim	dimension of the model
[in]	lambda	lambda of the sgd learning

---

Member Function Documentation

struct model& SvmSgd::get_model

(

)

[inline]

Retrieves model learnt.

Returns:

model learnt

double SvmSgd::getAccuracy

(

)

Returns accuracy on test data.

Returns:

accuracy on test data

bool SvmSgd::isExampleCorrectlyClassified

(

std::string

newfeature

)

Checks whether given example is classified correctly or not.

Returns:

whether example is correctly classified by the current model or not

void SvmSgd::scale

(

)

Scales model with wscale.

void SvmSgd::test_example

(

char *

featurechar

)

Tests given example in svm format and check whether prediction is correct or not.

Parameters:

[in]

featurechar

feature vector of the entity in svm format

---

Member Data Documentation

model SvmSgd::_current [private]

int SvmSgd::_dim [private]

int SvmSgd::ex [private]

number of predictions

int SvmSgd::match [private]

number of correct prediction during testing

Incremental_SGD* SvmSgd::sgd [private]

Updates old model.

---

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

• sgd_model.h
• sgd_model.cxx

---