mlpack::amf::SimpleResidueTermination Class Reference

This class implements a simple residue-based termination policy. More...

#include <simple_residue_termination.hpp>

Public Member Functions
	SimpleResidueTermination (const double minResidue=1e-5, const size_t maxIterations=10000)
	Construct the SimpleResidueTermination object with the given minimum residue (or the default) and the given maximum number of iterations (or the default). More...
template<typename MatType >
void	Initialize (const MatType &V)
	Initializes the termination policy before stating the factorization. More...
bool	IsConverged (arma::mat &W, arma::mat &H)
	Check if termination criterion is met. More...
const double &	Index () const
	Get current value of residue.
const size_t &	Iteration () const
	Get current iteration count.
const size_t &	MaxIterations () const
	Access max iteration count.
size_t &	MaxIterations ()
const double &	MinResidue () const
	Access minimum residue value.
double &	MinResidue ()

Public Attributes
double	minResidue
	residue threshold
size_t	maxIterations
	iteration threshold
double	residue
	current value of residue
size_t	iteration
	current iteration count
double	normOld
	norm of previous iteration
size_t	nm

Detailed Description

This class implements a simple residue-based termination policy.

The termination decision depends on two factors: the value of the residue (the difference between the norm of WH this iteration and the previous iteration), and the number of iterations. If the current value of residue drops below the threshold or the number of iterations goes above the iteration limit, IsConverged() will return true. This class is meant for use with the AMF (alternating matrix factorization) class.

See also

AMF

Constructor & Destructor Documentation

SimpleResidueTermination()

mlpack::amf::SimpleResidueTermination::SimpleResidueTermination ( const double  minResidue = 1e-5, const size_t  maxIterations = 10000  )

inline

Construct the SimpleResidueTermination object with the given minimum residue (or the default) and the given maximum number of iterations (or the default).

0 indicates no iteration limit.

Parameters

minResidue	Minimum residue for termination.
maxIterations	Maximum number of iterations.

Member Function Documentation

Initialize()

template<typename MatType >

void mlpack::amf::SimpleResidueTermination::Initialize ( const MatType &  V )

inline

Initializes the termination policy before stating the factorization.

Parameters

V	Input matrix being factorized.

IsConverged()

bool mlpack::amf::SimpleResidueTermination::IsConverged ( arma::mat &  W, arma::mat &  H  )

inline

Check if termination criterion is met.

Parameters

W	Basis matrix of output.
H	Encoding matrix of output.

---

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

• src/mlpack/methods/amf/termination_policies/simple_residue_termination.hpp