A.I.M.S algorithms


AimsMLPNeuron Class Reference

The class that describes MLP neuron. More...

#include <aims/neuralnet/multilayerperceptron.h>

Collaboration diagram for AimsMLPNeuron:

Public Member Functions

Constructor and Destructor
 AimsMLPNeuron (int ninputs)
 Constructor. More...
 
virtual ~AimsMLPNeuron ()
 Destructor deletes allocated data. More...
 
Methods
int nInputs () const
 Return the number of inputs. More...
 
float output () const
 Return the actual output. More...
 
float activation () const
 Return the activation. More...
 
float error () const
 Return the error. More...
 
float deltaW (int i) const
 Return the variation of weights. More...
 
float weight (int i) const
 Return the ieme weight. More...
 
void setWeight (const AimsData< float > &weight)
 Set the weights from weight vector. More...
 
void forward (const AimsData< float > &input)
 Forward propagation. More...
 
void evaluateError (float target)
 Error evaluation. More...
 
void backPropagation (const AimsData< float > &nexterror, const AimsData< float > &nextweight)
 Gradient back propagation. More...
 
void adjust (const AimsData< float > &input, float learningrate, float momentum)
 Modification of the weights. More...
 

Protected Attributes

Data
int _nInputs
 Number of inputs. More...
 
float _output
 Floating output. More...
 
float _activation
 Activation value (before entering sigmoid function) More...
 
float _error
 Error between desired and actual value. More...
 
AimsData< float > * _deltaW
 Weight variation data. More...
 
AimsData< float > * _weight
 Weight data. More...
 

Detailed Description

The class that describes MLP neuron.

Definition at line 47 of file multilayerperceptron.h.

Constructor & Destructor Documentation

AimsMLPNeuron::AimsMLPNeuron ( int  ninputs)

Constructor.

Parameters
ninputsnumber of inputs of the neuron
virtual AimsMLPNeuron::~AimsMLPNeuron ( )
virtual

Destructor deletes allocated data.

Member Function Documentation

float AimsMLPNeuron::activation ( ) const

Return the activation.

void AimsMLPNeuron::adjust ( const AimsData< float > &  input,
float  learningrate,
float  momentum 
)

Modification of the weights.

void AimsMLPNeuron::backPropagation ( const AimsData< float > &  nexterror,
const AimsData< float > &  nextweight 
)

Gradient back propagation.

float AimsMLPNeuron::deltaW ( int  i) const

Return the variation of weights.

float AimsMLPNeuron::error ( ) const

Return the error.

void AimsMLPNeuron::evaluateError ( float  target)

Error evaluation.

void AimsMLPNeuron::forward ( const AimsData< float > &  input)

Forward propagation.

int AimsMLPNeuron::nInputs ( ) const

Return the number of inputs.

float AimsMLPNeuron::output ( ) const

Return the actual output.

void AimsMLPNeuron::setWeight ( const AimsData< float > &  weight)

Set the weights from weight vector.

float AimsMLPNeuron::weight ( int  i) const

Return the ieme weight.

Member Data Documentation

float AimsMLPNeuron::_activation
protected

Activation value (before entering sigmoid function)

Definition at line 56 of file multilayerperceptron.h.

AimsData<float>* AimsMLPNeuron::_deltaW
protected

Weight variation data.

Definition at line 60 of file multilayerperceptron.h.

float AimsMLPNeuron::_error
protected

Error between desired and actual value.

Definition at line 58 of file multilayerperceptron.h.

int AimsMLPNeuron::_nInputs
protected

Number of inputs.

Definition at line 52 of file multilayerperceptron.h.

float AimsMLPNeuron::_output
protected

Floating output.

Definition at line 54 of file multilayerperceptron.h.

AimsData<float>* AimsMLPNeuron::_weight
protected

Weight data.

Definition at line 62 of file multilayerperceptron.h.


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