RegularizedSVDFunction< MatType > Class Template Reference

The data is stored in a matrix of type MatType, so that this class can be used with both dense and sparse matrix types. More...

Public Member Functions
	RegularizedSVDFunction (const MatType &data, const size_t rank, const double lambda)
	Constructor for RegularizedSVDFunction class. More...
const arma::mat &	Dataset () const
	Return the dataset passed into the constructor. More...
double	Evaluate (const arma::mat &parameters) const
	Evaluates the cost function over all examples in the data. More...
double	Evaluate (const arma::mat &parameters, const size_t start, const size_t batchSize=1) const
	Evaluates the cost function for one training example. More...
const arma::mat &	GetInitialPoint () const
	Return the initial point for the optimization. More...
void	Gradient (const arma::mat &parameters, arma::mat &gradient) const
	Evaluates the full gradient of the cost function over all the training examples. More...
template < typename GradType >
void	Gradient (const arma::mat &parameters, const size_t start, GradType &gradient, const size_t batchSize=1) const
	Evaluates the gradient of the cost function over one training example. More...
double	Lambda () const
	Return the regularization parameters. More...
size_t	NumFunctions () const
	Return the number of training examples. Useful for SGD optimizer. More...
size_t	NumItems () const
	Return the number of items in the data. More...
size_t	NumUsers () const
	Return the number of users in the data. More...
size_t	Rank () const
	Return the rank used for the factorization. More...
void	Shuffle ()
	Shuffle the points in the dataset. More...

Detailed Description

template

<

typename

MatType

=

arma::mat

>

class mlpack::svd::RegularizedSVDFunction< MatType >

The data is stored in a matrix of type MatType, so that this class can be used with both dense and sparse matrix types.

Template Parameters

MatType

The matrix type of the dataset.

Definition at line 29 of file regularized_svd_function.hpp.

Constructor & Destructor Documentation

RegularizedSVDFunction()

RegularizedSVDFunction	(	const MatType &	data,
		const size_t	rank,
		const double	lambda
	)

Constructor for RegularizedSVDFunction class.

The constructor calculates the number of users and items in the passed data. It also randomly initializes the parameter values.

Parameters

data	Dataset for which SVD is calculated.
rank	Rank used for matrix factorization.
lambda	Regularization parameter used for optimization.

Member Function Documentation

Dataset()

const arma::mat& Dataset ( ) const

inline

Return the dataset passed into the constructor.

Definition at line 101 of file regularized_svd_function.hpp.

Evaluate() [1/2]

double Evaluate

(

const arma::mat &

parameters

)

const

Evaluates the cost function over all examples in the data.

Parameters

parameters

Parameters(user/item matrices) of the decomposition.

Evaluate() [2/2]

double Evaluate	(	const arma::mat &	parameters,
		const size_t	start,
		const size_t	batchSize = 1
	)		const

Evaluates the cost function for one training example.

Useful for the SGD optimizer abstraction which uses one training example at a time.

Parameters

parameters	Parameters(user/item matrices) of the decomposition.
start	First index of the training examples to be used.
batchSize	Size of batch to evaluate.

GetInitialPoint()

const arma::mat& GetInitialPoint ( ) const

inline

Return the initial point for the optimization.

Definition at line 98 of file regularized_svd_function.hpp.

Gradient() [1/2]

void Gradient	(	const arma::mat &	parameters,
		arma::mat &	gradient
	)		const

Evaluates the full gradient of the cost function over all the training examples.

Parameters

parameters	Parameters(user/item matrices) of the decomposition.
gradient	Calculated gradient for the parameters.

Gradient() [2/2]

void Gradient	(	const arma::mat &	parameters,
		const size_t	start,
		GradType &	gradient,
		const size_t	batchSize = 1
	)		const

Evaluates the gradient of the cost function over one training example.

This function is useful for optimizers like SGD. The type of the gradient parameter is a template argument to allow the computation of a sparse gradient.

Template Parameters

GradType

The type of the gradient out-param.

Parameters

parameters	Parameters(user/item matrices) of the decomposition.
start	The first index of the training examples to use.
gradient	Calculated gradient for the parameters.
batchSize	Size of batch to calculate gradient for.

Lambda()

double Lambda ( ) const

inline

Return the regularization parameters.

Definition at line 113 of file regularized_svd_function.hpp.

NumFunctions()

size_t NumFunctions ( ) const

inline

Return the number of training examples. Useful for SGD optimizer.

Definition at line 104 of file regularized_svd_function.hpp.

NumItems()

size_t NumItems ( ) const

inline

Return the number of items in the data.

Definition at line 110 of file regularized_svd_function.hpp.

NumUsers()

size_t NumUsers ( ) const

inline

Return the number of users in the data.

Definition at line 107 of file regularized_svd_function.hpp.

Rank()

size_t Rank ( ) const

inline

Return the rank used for the factorization.

Definition at line 116 of file regularized_svd_function.hpp.

Shuffle()

void Shuffle

(

)

Shuffle the points in the dataset.

This may be used by optimizers.

---

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

• /home/ryan/src/mlpack.org/_src/mlpack-git/src/mlpack/methods/regularized_svd/regularized_svd_function.hpp