keel.Algorithms.Neural_Networks.NNEP_Common.initiators

Class RandomInitiator

java.lang.Object

keel.Algorithms.Neural_Networks.NNEP_Common.initiators.PureLayerInitiator

keel.Algorithms.Neural_Networks.NNEP_Common.initiators.RandomInitiator

All Implemented Interfaces:

IInitiator

Direct Known Subclasses:

FullRandomInitiator

public class RandomInitiator
extends PureLayerInitiator

Random initiator both conections and weights

Since:

JDK1.5

Version:

0.1

Author:

Written by Pedro Antonio Gutierrez Penya, Aaron Ruiz Mora (University of Cordoba) 17/07/2007

Field Summary

Fields

Modifier and Type	Field and Description
protected double	significativeWeight Minimum absolute value of the new weights

Fields inherited from class keel.Algorithms.Neural_Networks.NNEP_Common.initiators.PureLayerInitiator

context, randGen, species

Constructor Summary

Constructors

Constructor and Description
RandomInitiator() Empty constructor

Method Summary

Modifier and Type	Method and Description
Link[]	createLinks(LinkedLayer linkedLayer, ILayer<? extends INeuron> previousLayer, LinkedNeuron newNeuron) Create all the links of a neural net
double	getSignificativeWeight() Access to significative weight value
void	initiateWeights(LinkedLayer linkedLayer) Initiate the weights of all the links of a neural net
void	prepareInitiation() Prepare initiation process
void	setSignificativeWeight(double significativeWeight) Sets the significative weight value

Methods inherited from class keel.Algorithms.Neural_Networks.NNEP_Common.initiators.PureLayerInitiator

contextualize, initiate, setRandGen

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

significativeWeight

protected double significativeWeight

Minimum absolute value of the new weights

Constructor Detail

RandomInitiator

public RandomInitiator()

Empty constructor

Method Detail

getSignificativeWeight

public double getSignificativeWeight()

Access to significative weight value

Returns:

double Significative weight value

setSignificativeWeight

public void setSignificativeWeight(double significativeWeight)

Sets the significative weight value

Parameters:

significativeWeight - New significative weight value

prepareInitiation

public void prepareInitiation()

Prepare initiation process

Specified by:

prepareInitiation in class PureLayerInitiator

createLinks

public Link[] createLinks(LinkedLayer linkedLayer,
                          ILayer<? extends INeuron> previousLayer,
                          LinkedNeuron newNeuron)

Create all the links of a neural net

Specified by:

createLinks in class PureLayerInitiator

Parameters:

linkedLayer - Linked layer where create the links

previousLayer - LinkedLayer the neurons are going to be connected

newNeuron - New neuron to create its links

Returns:

Random generator

initiateWeights

public void initiateWeights(LinkedLayer linkedLayer)

Initiate the weights of all the links of a neural net

Specified by:

initiateWeights in class PureLayerInitiator

Parameters:

linkedLayer - Linked layer where initiate the weights