Flood::NormalizedSquaredError Class Reference

#include <NormalizedSquaredError.h>

Inheritance diagram for Flood::NormalizedSquaredError:

List of all members.

Public Member Functions

NormalizedSquaredError (void)

NormalizedSquaredError (MultilayerPerceptron *)

NormalizedSquaredError (InputTargetDataSet *)

NormalizedSquaredError (MultilayerPerceptron *, InputTargetDataSet *)

virtual ~NormalizedSquaredError (void)

InputTargetDataSet * get_input_target_data_set_pointer (void)

void set (void)

void set (MultilayerPerceptron *)

void set (InputTargetDataSet *)

void set (MultilayerPerceptron *, InputTargetDataSet *)

void set_input_target_data_set_pointer (InputTargetDataSet *)

double calculate_training_normalization_coefficient (void)

double calculate_validation_normalization_coefficient (void)

double calculate_objective (void)

double calculate_validation_error (void)

Vector< double > calculate_output_errors (const Vector< Vector< double > > &, const Vector< double > &)

Vector< Vector< double > > calculate_hidden_errors (const Vector< Vector< double > > &, const Vector< double > &)

Vector< double > calculate_hidden_layers_error_gradient (const Vector< double > &, const Vector< Vector< double > > &, const Vector< Vector< double > > &)

Vector< double > calculate_output_layer_error_gradient (const Vector< Vector< double > > &, const Vector< double > &)

Vector< double > calculate_objective_gradient (void)

Detailed Description

This class represents the normalized squared error objective functional of a multilayer perceptron. This objective functional is used in data modeling problems. If it has a value of unity then the neural network predicting the data "in the mean" while a value of zero means perfect prediction of data.

Definition at line 30 of file NormalizedSquaredError.h.

Constructor & Destructor Documentation

Flood::NormalizedSquaredError::NormalizedSquaredError ( void ) [explicit]

Default constructor. It creates a normalized squared error objective functional object not associated to any multilayer perceptron and not measured on any input-target data set. It also initializes all the rest of class members to their default values.

Definition at line 39 of file NormalizedSquaredError.cpp.

Flood::NormalizedSquaredError::NormalizedSquaredError ( MultilayerPerceptron * new_multilayer_perceptron_pointer ) [explicit]

Multilayer perceptron constructor. It creates a normalized squared error objective functional object associated to a multilayer perceptron object but not measured on any input-target data set. It also initializes all the rest of class members to their default values.

Parameters:

new_multilayer_perceptron_pointer

Pointer to a multilayer perceptron object.

Definition at line 52 of file NormalizedSquaredError.cpp.

Flood::NormalizedSquaredError::NormalizedSquaredError ( InputTargetDataSet * new_input_target_data_set_pointer ) [explicit]

Multilayer perceptron constructor. It creates a normalized squared error objective functional object not associated to any multilayer perceptron but to be measured on an input-target data set object. It also initializes all the rest of class members to their default values.

Parameters:

new_input_target_data_set_pointer

Pointer to an input-target data set object.

Definition at line 66 of file NormalizedSquaredError.cpp.

Flood::NormalizedSquaredError::NormalizedSquaredError	(	MultilayerPerceptron *	new_multilayer_perceptron_pointer,
		InputTargetDataSet *	new_input_target_data_set_pointer
	)			`[explicit]`

General constructor. It creates a normalized squared error objective functional object associated to a multilayer perceptron and measured on an input-target data set. It also initializes all the rest of class members to their default values.

Parameters:

	new_multilayer_perceptron_pointer	Pointer to a multilayer perceptron object.
	new_input_target_data_set_pointer	Pointer to an input-target data set object.

Definition at line 81 of file NormalizedSquaredError.cpp.

Flood::NormalizedSquaredError::~NormalizedSquaredError ( void ) [virtual]

Destructor.

Definition at line 92 of file NormalizedSquaredError.cpp.

Member Function Documentation

Vector< Vector< double > > Flood::NormalizedSquaredError::calculate_hidden_errors	(	const Vector< Vector< double > > &	forward_propagation_derivative,
		const Vector< double > &	output_errors
	)

This method returns the hidden errors to be used in the back-propagation algorithm for calculating the objective gradient.

Parameters:

	forward_propagation_derivative	Forward propagation derivative.
	output_errors	Output errors.

Definition at line 532 of file NormalizedSquaredError.cpp.

Vector< double > Flood::NormalizedSquaredError::calculate_hidden_layers_error_gradient	(	const Vector< double > &	input,
		const Vector< Vector< double > > &	forward_propagation_derivative,
		const Vector< Vector< double > > &	hidden_errors
	)

This method returns the hidden layers error gradient to be used in the back-propagation algorithm for calculating the objective gradient.

Parameters:

	input	Input values.
	forward_propagation_derivative	Forward propagation derivative.
	hidden_errors	Hidden errors.

Definition at line 666 of file NormalizedSquaredError.cpp.

double Flood::NormalizedSquaredError::calculate_objective ( void ) [virtual]

This method returns the evaluation value of a multilayer perceptron according to the normalized squared error on an input-target data set.

Implements Flood::ObjectiveFunctional.

Definition at line 257 of file NormalizedSquaredError.cpp.

Vector< double > Flood::NormalizedSquaredError::calculate_objective_gradient ( void ) [virtual]

This method returns the the normalized squared error function gradient of a multilayer perceptron on an input-target data set. It uses the error back-propagation method.

Reimplemented from Flood::ObjectiveFunctional.

Definition at line 868 of file NormalizedSquaredError.cpp.

Vector< double > Flood::NormalizedSquaredError::calculate_output_errors	(	const Vector< Vector< double > > &	forward_propagation_derivative,
		const Vector< double > &	target
	)

This method returns the output errors to be used in the back-propagation algorithm for calculating the objective gradient.

Parameters:

	forward_propagation_derivative	Forward propagation derivative.
	target	Target values.

Definition at line 355 of file NormalizedSquaredError.cpp.

Vector< double > Flood::NormalizedSquaredError::calculate_output_layer_error_gradient	(	const Vector< Vector< double > > &	forward_propagation_derivative,
		const Vector< double > &	output_errors
	)

This method returns the output layers error gradient to be used in the back-propagation algorithm for calculating the objective gradient.

Parameters:

	forward_propagation_derivative	Forward propagation derivative.
	output_errors	Output errors.

Definition at line 790 of file NormalizedSquaredError.cpp.

double Flood::NormalizedSquaredError::calculate_training_normalization_coefficient ( void )

This method returns the normalization coefficient to be used for the evaluation of the error. This is measured on the training instances of the input-target data set.

Definition at line 171 of file NormalizedSquaredError.cpp.

double Flood::NormalizedSquaredError::calculate_validation_normalization_coefficient ( void )

This method returns the normalization coefficient to be used for the evaluation of the validation error. This is therefore measured on the validation instances of the input-target data set.

Definition at line 214 of file NormalizedSquaredError.cpp.

InputTargetDataSet* Flood::NormalizedSquaredError::get_input_target_data_set_pointer ( void ) [inline]

This method returns a pointer to the input-target data set object on which the objective functional is measured.

Definition at line 62 of file NormalizedSquaredError.h.

void Flood::NormalizedSquaredError::set	(	MultilayerPerceptron *	new_multilayer_perceptron_pointer,
		InputTargetDataSet *	new_input_target_data_set_pointer
	)

This method sets new multilayer perceptron and input-target data set pointers. It also initializes all the rest of class members to their default values.

Parameters:

	new_multilayer_perceptron_pointer	Pointer to a multilayer perceptron object.
	new_input_target_data_set_pointer	Pointer to an input-target data set object.

Definition at line 147 of file NormalizedSquaredError.cpp.

void Flood::NormalizedSquaredError::set ( InputTargetDataSet * new_input_target_data_set_pointer )

This method sets the multilayer perceptron pointer to null, and sets a new input-target data set pointer. It also initializes all the rest of class members to their default values.

Parameters:

new_input_target_data_set_pointer

Pointer to an input-target data set object.

Definition at line 132 of file NormalizedSquaredError.cpp.

void Flood::NormalizedSquaredError::set ( MultilayerPerceptron * new_multilayer_perceptron_pointer )

This method sets a new multilayer perceptron pointer, and sets the input-target data set pointer to NULL. It also initializes all the rest of class members to their default values.

Parameters:

new_multilayer_perceptron_pointer

Pointer to a multilayer perceptron object.

Definition at line 118 of file NormalizedSquaredError.cpp.

void Flood::NormalizedSquaredError::set ( void )

This method sets the multilayer perceptron and input-target data set object pointers to NULL. It also initializes all the rest of class members to their default values.

Definition at line 104 of file NormalizedSquaredError.cpp.

void Flood::NormalizedSquaredError::set_input_target_data_set_pointer ( InputTargetDataSet * new_input_target_data_set_pointer )

This method sets a pointer to an input-data set object on which the objective functional is to be measured.

Parameters:

new_input_target_data_set_pointer

Pointer to an input-target data set object.

Definition at line 160 of file NormalizedSquaredError.cpp.

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


Public Member Functions
	NormalizedSquaredError (void)
	NormalizedSquaredError (MultilayerPerceptron *)
	NormalizedSquaredError (InputTargetDataSet *)
	NormalizedSquaredError (MultilayerPerceptron , InputTargetDataSet )
virtual	~NormalizedSquaredError (void)
InputTargetDataSet *	get_input_target_data_set_pointer (void)
void	set (void)
void	set (MultilayerPerceptron *)
void	set (InputTargetDataSet *)
void	set (MultilayerPerceptron , InputTargetDataSet )
void	set_input_target_data_set_pointer (InputTargetDataSet *)
double	calculate_training_normalization_coefficient (void)
double	calculate_validation_normalization_coefficient (void)
double	calculate_objective (void)
double	calculate_validation_error (void)
Vector< double >	calculate_output_errors (const Vector< Vector< double > > &, const Vector< double > &)
Vector< Vector< double > >	calculate_hidden_errors (const Vector< Vector< double > > &, const Vector< double > &)
Vector< double >	calculate_hidden_layers_error_gradient (const Vector< double > &, const Vector< Vector< double > > &, const Vector< Vector< double > > &)
Vector< double >	calculate_output_layer_error_gradient (const Vector< Vector< double > > &, const Vector< double > &)
Vector< double >	calculate_objective_gradient (void)