Factorization Class Reference

Decompose tensors (CP, Tucker) with non-negative constraints or coupled matrix. More...

Inheritance diagram for Factorization:

Public Member Functions
void	Parafac (int rank, int fully_or_partially_observable, char *in_tensor_path, char *out_factor_path, const GPU_MODE _gpu_mode=ANY)
	CANDECOMP/PARAFAC factorization. More...
void	ParafacNN (int rank, int fully_or_partially_observable, char *in_tensor_path, char *out_factor_path, const GPU_MODE _gpu_mode=ANY)
	Nonegative CANDECOMP/PARAFAC factorization. More...
void	Tucker (int rank, int fully_or_partially_observable, char *in_tensor_path, char *out_factor_path, const GPU_MODE _gpu_mode=ANY)
	Tucker factorization. More...
void	TuckerNN (int rank, int fully_or_partially_observable, char *in_tensor_path, char *out_factor_path, const GPU_MODE _gpu_mode=ANY)
	Nonnegative Tucker factorization. More...
void	CoupledMatrixTensorFactorization (int rank, int coupled_mode, char *in_tensor_path, char *in_matrix_path, char *out_factor_path, const GPU_MODE _gpu_mode=ANY)
	Coupled Matrix and Tensor factorization. More...

Detailed Description

Decompose tensors (CP, Tucker) with non-negative constraints or coupled matrix.

Member Function Documentation

CoupledMatrixTensorFactorization()

void Factorization::CoupledMatrixTensorFactorization	(	int	rank,
		int	cmode,
		char *	in_tensor_path,
		char *	in_matrix_path,
		char *	out_factor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Coupled Matrix and Tensor factorization.

Parameters

rank	size of rank
cmode	couple mode of matrix and tensor
in_tensor_path	input tensor path
in_matrix_path	input matrix path
out_factor_path	output factor matrices and a core tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Parafac()

void Factorization::Parafac	(	int	rank,
		int	fully_or_partially_observable,
		char *	in_tensor_path,
		char *	out_factor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

CANDECOMP/PARAFAC factorization.

Parameters

rank	size of rank
fully_or_partially_observable	fully observable tensor or partially observable tensor
in_tensor_path	input tensor path
out_factor_path	output factor matrices path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

ParafacNN()

void Factorization::ParafacNN	(	int	rank,
		int	fully_or_partially_observable,
		char *	in_tensor_path,
		char *	out_factor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Nonegative CANDECOMP/PARAFAC factorization.

Parameters

rank	size of rank
fully_or_partially_observable	fully observable tensor or partially observable tensor
in_tensor_path	input tensor path
out_factor_path	output factor matrices path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Tucker()

void Factorization::Tucker	(	int	rank,
		int	fully_or_partially_observable,
		char *	in_tensor_path,
		char *	out_factor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Tucker factorization.

Parameters

rank	size of rank
fully_or_partially_observable	0 for fully observable and 1 for partially observable
in_tensor_path	input tensor path
out_factor_path	output factor matrices and a core tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

TuckerNN()

void Factorization::TuckerNN	(	int	rank,
		int	fully_or_partially_observable,
		char *	in_tensor_path,
		char *	out_factor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Nonnegative Tucker factorization.

Parameters

rank	size of rank
fully_or_partially_observable	0 for fully observable and 1 for partially observable
in_tensor_path	input tensor path
out_factor_path	output factor matrices and a core tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None