org.ai_heuristic.algo.cluster.som.network

Class KohonenNeuron

java.lang.Object

org.ai_heuristic.algo.cluster.som.network.DefaultNeuron

org.ai_heuristic.algo.cluster.som.network.KohonenNeuron

All Implemented Interfaces:

NeuronModel

Direct Known Subclasses:

KohonenNeuronWithTired

public class KohonenNeuron
extends DefaultNeuron

Updated version of the source code for the kohonen self-organising map at http://jknnl.sourceforge.net/ Some classes or packages may have been moved or changed. See above for the copyright notices.

Author:

Kieran Greer.

Field Summary

Fields

Modifier and Type	Field and Description
protected FunctionMetric	distanceFunction Distance function

Fields inherited from class org.ai_heuristic.algo.cluster.som.network.DefaultNeuron

activationFunction, learnFunction, weight

Constructor Summary

Constructors

Constructor and Description
KohonenNeuron(double[] weightArray, FunctionMetric distanceFunction, FunctionLearn learnFunction, FunctionActivate activationFunction) Creates new neuron with weight specified by array and specified activation function.
KohonenNeuron(int weightNumber, double[] maxWeight, FunctionMetric distanceFunction, FunctionLearn learnFunction, FunctionActivate activationFunction) Creates new neuron with specified numbers of weight.

Method Summary

Modifier and Type	Method and Description
void	changeNeuronWeight(double[] vector, double factor, double influence) Change neuron weights for specified neuron number,factor and input data vector
FunctionMetric	getDistanceFunction() Return reference to distance function
FunctionLearn	getLearnFunction() Return reference to learn function
double	getValue(double[] inputVector) Return value of the neuron after activation.
boolean	isBetter(double value1, double value2) Return true if value2 is better than value1, as determined by the measurements of the distance evaluation function, or a direct comparison if there is not one.
void	setDistanceFunction(FunctionMetric distanceFunction) Set distance function
void	setLearnFunction(FunctionLearn learnFunction) Set learn function
java.lang.String	toString() Returns a string representation of the NeuronModel.

Methods inherited from class org.ai_heuristic.algo.cluster.som.network.DefaultNeuron

changeNeuronWeight, getActivationFunction, getWeight, setActivationFunction, setWeight, valueTimesWeights

Methods inherited from class java.lang.Object

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

Field Detail

distanceFunction

protected FunctionMetric distanceFunction

Distance function

Constructor Detail

KohonenNeuron

public KohonenNeuron(int weightNumber,
                     double[] maxWeight,
                     FunctionMetric distanceFunction,
                     FunctionLearn learnFunction,
                     FunctionActivate activationFunction)

Creates new neuron with specified numbers of weight. Value of the weight is random where max value is defined by maxWeight. Activation function is defined by activationFunction parameter.

Parameters:

weightNumber - numbers of weight

maxWeight - max value of the weight. If null or different number of entries to weightNumber, then start with random value between 0 and 1.

distanceFunction - distance function.

learnFunction - reference to learning function

activationFunction - activation function.

KohonenNeuron

public KohonenNeuron(double[] weightArray,
                     FunctionMetric distanceFunction,
                     FunctionLearn learnFunction,
                     FunctionActivate activationFunction)

Creates new neuron with weight specified by array and specified activation function. Numbers of weights are the same as length of the array.

Parameters:

weightArray - array of the weight

distanceFunction - distance function.

learnFunction - reference to learning function

activationFunction - activation function.

Method Detail

getValue

public double getValue(double[] inputVector)
                throws java.lang.Exception

Return value of the neuron after activation. if activation function is not set, function return sum of the multiplication of vector v_i and weight w_i for each i (numbers of weight and length of the input vector)

Specified by:

getValue in interface NeuronModel

Overrides:

getValue in class DefaultNeuron

Parameters:

inputVector - input vector for neuron

Returns:

return

Throws:

java.lang.Exception - any error.

changeNeuronWeight

public void changeNeuronWeight(double[] vector,
                               double factor,
                               double influence)
                        throws java.lang.Exception

Change neuron weights for specified neuron number,factor and input data vector

Specified by:

changeNeuronWeight in interface NeuronModel

Overrides:

changeNeuronWeight in class DefaultNeuron

Parameters:

vector - input data vector.

factor - the iteration number or factor.

influence - cells in neighbourhood have a reduced influence.

Throws:

java.lang.Exception - any error.

isBetter

public boolean isBetter(double value1,
                        double value2)
                 throws java.lang.Exception

Return true if value2 is better than value1, as determined by the measurements of the distance evaluation function, or a direct comparison if there is not one.

Parameters:

value1 - the first value type.

value2 - the second value type.

Returns:

true if value2 is better than value1, as determined by this evaluation function.

Throws:

java.lang.Exception - any error.

toString

public java.lang.String toString()

Returns a string representation of the NeuronModel. NeuronModel is describe by its weights [w_1,w_2,...]

Overrides:

toString in class java.lang.Object

Returns:

a string representation of the NeuronModel

getDistanceFunction

public FunctionMetric getDistanceFunction()

Return reference to distance function

Returns:

reference to distance function

setDistanceFunction

public void setDistanceFunction(FunctionMetric distanceFunction)

Set distance function

Parameters:

distanceFunction - reference to distance function

getLearnFunction

public FunctionLearn getLearnFunction()

Return reference to learn function

Returns:

reference to learn function

setLearnFunction

public void setLearnFunction(FunctionLearn learnFunction)

Set learn function

Parameters:

learnFunction - reference to learn function