itpp::Fast_ICA Class Reference
[Fast Independent Component Analysis]

Fast_ICA Fast Independent Component Analysis (Fast ICA)

The software is based upon original FastICA for Matlab from A. Hyvarinen. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis. IEEE Transactions on Neural Networks, 10(3), pp. 626-634, 1999. More...

#include <fastica.h>

List of all members.

Public Member Functions

Fast_ICA (mat ma_mixed_sig)

Constructor.

void separate (void)

Explicit launch of main FastICA function.

void set_approach (int in_approach)

Set approach : FICA_APPROACH_DEFL or FICA_APPROACH_SYMM (default).

void set_nrof_independent_components (int in_nrIC)

Set number of independent components to separate.

void set_non_linearity (int in_g)

Set non-linearity.

void set_fine_tune (bool in_finetune)

Set fine tuning.

void set_a1 (double fl_a1)

Set $a_1$ parameter.

void set_a2 (double fl_a2)

Set $a_2$ parameter.

void set_mu (double fl_mu)

Set $\mu$ parameter.

void set_epsilon (double fl_epsilon)

Set convergence parameter $\epsilon$ .

void set_sample_size (double fl_sampleSize)

Set sample size.

void set_stabilization (bool in_stabilization)

Set stabilization mode true or off.

void set_max_num_iterations (int in_maxNumIterations)

Set maximum number of iterations.

void set_max_fine_tune (int in_maxFineTune)

Set maximum number of iterations for fine tuning.

void set_first_eig (int in_firstEig)

Set first eigenvalue index to take into account.

void set_last_eig (int in_lastEig)

Set last eigenvalue index to take into account.

void set_pca_only (bool in_PCAonly)

If true, only perform Principal Component Analysis (default = false).

void set_init_guess (mat ma_initGuess)

Set initial guess matrix instead of random (default).

mat get_mixing_matrix ()

Get mixing matrix.

mat get_separating_matrix ()

Get separating matrix.

mat get_independent_components ()

Get separated signals.

int get_nrof_independent_components ()

Get number of independent components.

mat get_principal_eigenvectors ()

Get nrIC first columns of the de-whitening matrix.

mat get_whitening_matrix ()

Get the whitening matrix.

mat get_dewhitening_matrix ()

Get the de-whitening matrix.

mat get_white_sig ()

Get whitened signals.

Detailed Description

Fast_ICA Fast Independent Component Analysis (Fast ICA)

The software is based upon original FastICA for Matlab from A. Hyvarinen. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis. IEEE Transactions on Neural Networks, 10(3), pp. 626-634, 1999.

Example:

FastICA fastica(sources);
fastica.set_nrof_independent_components(sources.rows());
fastica.set_non_linearity(  FICA_NONLIN_TANH );
fastica.set_approach( FICA_APPROACH_DEFL );
fastica.separate();
mat ICs = fastica.get_independent_components();

Constructor & Destructor Documentation

itpp::Fast_ICA::Fast_ICA ( mat ma_mixed_sig )

Constructor.

Construct a Fast_ICA object with mixed signals to separate.

Parameters:

ma_mixed_sig

(Input) Mixed signals to separate

References FICA_APPROACH_SYMM, FICA_INIT_RAND, and FICA_NONLIN_POW3.

Member Function Documentation

mat itpp::Fast_ICA::get_dewhitening_matrix ( )

Get the de-whitening matrix.

Return dewhitening matrix.

Returns:: Dewhitening matrix

mat itpp::Fast_ICA::get_independent_components ( )

Get separated signals.

Return separated signals (Independent Components).

Returns:: ICs

References it_warning, and itpp::zeros().

mat itpp::Fast_ICA::get_mixing_matrix ( )

Get mixing matrix.

Return mixing matrix.

Returns:: Mixing matrix

References it_warning, and itpp::zeros().

int itpp::Fast_ICA::get_nrof_independent_components ( )

Get number of independent components.

Return number of ICs.

Returns:: Number of ICs

mat itpp::Fast_ICA::get_principal_eigenvectors ( )

Get nrIC first columns of the de-whitening matrix.

Return principal eigenvectors.

Returns:: Principal eigenvectors

mat itpp::Fast_ICA::get_separating_matrix ( )

Get separating matrix.

Return separating matrix.

Returns:: Separating matrix

References it_warning, and itpp::zeros().

mat itpp::Fast_ICA::get_white_sig ( )

Get whitened signals.

Return whitened signals.

Returns:: Whitened signals

mat itpp::Fast_ICA::get_whitening_matrix ( )

Get the whitening matrix.

Return whitening matrix.

Returns:: Whitening matrix

void itpp::Fast_ICA::separate ( void )

Explicit launch of main FastICA function.

Explicit launch of the Fast_ICA algorithm.

References itpp::diag(), FICA_INIT_RAND, itpp::max_index(), itpp::to_ivec(), and itpp::zeros().

void itpp::Fast_ICA::set_a1 ( double fl_a1 )

Set $a_1$ parameter.

Set internal parameter $a_1$ of Fast_ICA (See reference paper).

Parameters:

fl_a1

(Input) Parameter $a_1$

from reference paper

void itpp::Fast_ICA::set_a2 ( double fl_a2 )

Set $a_2$ parameter.

Set internal parameter $a_2$ of Fast_ICA (See reference paper).

Parameters:

fl_a2

(Input) Parameter $a_2$

from reference paper

void itpp::Fast_ICA::set_approach ( int in_approach )

Set approach : FICA_APPROACH_DEFL or FICA_APPROACH_SYMM (default).

Set approach to use : FICA_APPROACH_SYMM (symmetric) or FICA_APPROACH_DEFL (deflation). The symmetric approach computes all ICs at a time, whereas the deflation approach computes them one by one.

Parameters:

in_approach

(Input) Type of approach to use

References FICA_APPROACH_DEFL.

void itpp::Fast_ICA::set_epsilon ( double fl_epsilon )

Set convergence parameter $\epsilon$ .

Set $\epsilon$ parameter for convergence precision.

Parameters:

fl_epsilon

(Input) $\epsilon$ is convergence precision

void itpp::Fast_ICA::set_fine_tune ( bool in_finetune )

Set fine tuning.

Set fine tuning true or false.

Parameters:

in_finetune

(Input) Boolean (true or false)

void itpp::Fast_ICA::set_first_eig ( int in_firstEig )

Set first eigenvalue index to take into account.

Parameters:

in_firstEig

(Input) First eigenvalue index to take into account

void itpp::Fast_ICA::set_init_guess ( mat ma_initGuess )

Set initial guess matrix instead of random (default).

Set initial matrix instead of random matrix.

Parameters:

ma_initGuess

(Input) Initial guess matrix

References FICA_INIT_GUESS.

void itpp::Fast_ICA::set_last_eig ( int in_lastEig )

Set last eigenvalue index to take into account.

Parameters:

in_lastEig

(Input) Last eigenvalue index to take into account

void itpp::Fast_ICA::set_max_fine_tune ( int in_maxFineTune )

Set maximum number of iterations for fine tuning.

Set maximum numberr of iterations for fine tuning.

Parameters:

in_maxFineTune

(Input) Maximum number of iterations for fine tuning stage

void itpp::Fast_ICA::set_max_num_iterations ( int in_maxNumIterations )

Set maximum number of iterations.

Set maximum number of iterations for Fast_ICA.

Parameters:

in_maxNumIterations

(Input) Maximum number of iterations to go through

void itpp::Fast_ICA::set_mu ( double fl_mu )

Set $\mu$ parameter.

Set internal parameter $\mu$ of Fast_ICA (See reference paper).

Parameters:

fl_mu

(Input) Parameter $\mu$ from reference paper

void itpp::Fast_ICA::set_non_linearity ( int in_g )

Set non-linearity.

Set non-linearity to use : FICA_NONLIN_POW3 (default), FICA_NONLIN_TANH, FICA_NONLIN_GAUSS, FICA_NONLIN_SKEW

Parameters:

in_g

(Input) Non-linearity. Can be selected from FICA_NONLIN_POW3, FICA_NONLIN_TANH, FICA_NONLIN_GAUSS or FICA_NONLIN_SKEW

void itpp::Fast_ICA::set_nrof_independent_components ( int in_nrIC )

Set number of independent components to separate.

Set the number of ICs to compute.

Parameters:

in_nrIC

(Input) Number of ICs to compute

void itpp::Fast_ICA::set_pca_only ( bool in_PCAonly )

If true, only perform Principal Component Analysis (default = false).

Wether to perform PCA only or PCA+ICA.

Parameters:

in_PCAonly

(Input) True = PCA only, false = PCA+ICA (default)

void itpp::Fast_ICA::set_sample_size ( double fl_sampleSize )

Set sample size.

Set the percentage of samples to take into account at every iteration.

Parameters:

fl_sampleSize

(Input) Percentage of data to take into account at every iteration

void itpp::Fast_ICA::set_stabilization ( bool in_stabilization )

Set stabilization mode true or off.

Set stabilization mode.

Parameters:

in_stabilization

(Input) Set stabilization true or false

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


Public Member Functions
	Fast_ICA (mat ma_mixed_sig)
	Constructor.
void	separate (void)
	Explicit launch of main FastICA function.
void	set_approach (int in_approach)
	Set approach : FICA_APPROACH_DEFL or FICA_APPROACH_SYMM (default).
void	set_nrof_independent_components (int in_nrIC)
	Set number of independent components to separate.
void	set_non_linearity (int in_g)
	Set non-linearity.
void	set_fine_tune (bool in_finetune)
	Set fine tuning.
void	set_a1 (double fl_a1)
	Set parameter.
void	set_a2 (double fl_a2)
	Set parameter.
void	set_mu (double fl_mu)
	Set $\mu$ parameter.
void	set_epsilon (double fl_epsilon)
	Set convergence parameter $\epsilon$ .
void	set_sample_size (double fl_sampleSize)
	Set sample size.
void	set_stabilization (bool in_stabilization)
	Set stabilization mode true or off.
void	set_max_num_iterations (int in_maxNumIterations)
	Set maximum number of iterations.
void	set_max_fine_tune (int in_maxFineTune)
	Set maximum number of iterations for fine tuning.
void	set_first_eig (int in_firstEig)
	Set first eigenvalue index to take into account.
void	set_last_eig (int in_lastEig)
	Set last eigenvalue index to take into account.
void	set_pca_only (bool in_PCAonly)
	If true, only perform Principal Component Analysis (default = false).
void	set_init_guess (mat ma_initGuess)
	Set initial guess matrix instead of random (default).
mat	get_mixing_matrix ()
	Get mixing matrix.
mat	get_separating_matrix ()
	Get separating matrix.
mat	get_independent_components ()
	Get separated signals.
int	get_nrof_independent_components ()
	Get number of independent components.
mat	get_principal_eigenvectors ()
	Get nrIC first columns of the de-whitening matrix.
mat	get_whitening_matrix ()
	Get the whitening matrix.
mat	get_dewhitening_matrix ()
	Get the de-whitening matrix.
mat	get_white_sig ()
	Get whitened signals.

itpp::Fast_ICA Class Reference [Fast Independent Component Analysis]

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

itpp::Fast_ICA Class Reference
[Fast Independent Component Analysis]