/**
* Copyright 2010 Neuroph Project http://neuroph.sourceforge.net
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.neuroph.core;
import java.io.Serializable;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.Random;
import java.util.Vector;
import org.neuroph.core.input.InputFunction;
import org.neuroph.core.transfer.Step;
import org.neuroph.core.transfer.TransferFunction;
/**
*<pre>
* Basic general neuron model according to McCulloch-Pitts neuron model.
* Different neuron models can be created by using different input and transfer functions for instances of this class,
* or by deriving from this class. The neuron is basic processing element of neural network.
* This class implements the following behaviour:
*
* output = transferFunction( inputFunction(inputConnections) )
*</pre>
*
* @see InputFunction
* @see TransferFunction
* @author Zoran Sevarac <sevarac@gmail.com>
*/
public class Neuron implements Serializable {
/**
* The class fingerprint that is set to indicate serialization
* compatibility with a previous version of the class
*/
private static final long serialVersionUID = 3L;
/**
* Parent layer for this neuron
*/
protected Layer parentLayer;
/**
* Collection of neuron's input connections (connections to this neuron)
*/
protected List <Connection> inputConnections = new ArrayList<Connection>();
/**
* Collection of neuron's output connections (connections from this to other
* neurons)
*/
protected List <Connection> outConnections = new ArrayList<Connection>();
/**
* Total net input for this neuron. Represents total input for this neuron
* received from input function.
*/
protected transient double netInput = 0;
/**
* Neuron output
*/
protected transient double output = 0;
/**
* Local error for this neuron
*/
protected transient double error = 0;
/**
* Input function for this neuron
*/
protected InputFunction inputFunction;
/**
* Transfer function for this neuron
*/
protected TransferFunction transferFunction;
/**
* Creates an instance of Neuron with the weighted sum, input function
* and Step transfer function. This is the original McCulloch-Pitts
* neuron model.
*/
public Neuron() {
this.inputFunction = new InputFunction();
this.transferFunction = new Step();
}
/**
* Creates an instance of Neuron with the specified input and transfer functions.
*
* @param inputFunction
* input function for this neuron
* @param transferFunction
* transfer function for this neuron
*/
public Neuron(InputFunction inputFunction, TransferFunction transferFunction) {
this.inputFunction = inputFunction;
this.transferFunction = transferFunction;
}
/**
* Calculates neuron's output
*/
public void calculate() {
//why fo we need this? for input neurons!!!
if (this.hasInputConnections()) {
this.netInput = this.inputFunction.getOutput(this.inputConnections);
}
this.output = this.transferFunction.getOutput(this.netInput);
}
/**
* Sets input and output activation levels to zero
*/
public void reset() {
this.setInput(0d);
this.setOutput(0d);
}
/**
* Sets neuron's input
*
* @param input
* input value to set
*/
public void setInput(double input) {
this.netInput = input;
}
/**
* Returns total net input
*
* @return total net input
*/
public double getNetInput() {
return this.netInput;
}
/**
* Returns neuron's output
*
* @return neuron output
*/
public double getOutput() {
return this.output;
}
/**
* Returns true if there are input connections for this neuron, false
* otherwise
*
* @return true if there is input connection, false otherwise
*/
public boolean hasInputConnections() {
return (!this.inputConnections.isEmpty());
}
/**
* Returns Iterator interface for accessing input connections
*
* @return iterator interface for accessing input connections
*/
public Iterator<Connection> getInputsIterator() {
return this.inputConnections.iterator();
}
/**
* Adds the specified input connection
*
* @param connection
* input connection to add
*/
public void addInputConnection(Connection connection) {
this.inputConnections.add(connection);
Neuron fromNeuron = connection.getFromNeuron();
fromNeuron.addOutputConnection(connection);
}
/**
* Adds input connection from specified neuron
*
* @param fromNeuron
* neuron to connect from
*/
public void addInputConnection(Neuron fromNeuron) {
Connection connection = new Connection(fromNeuron, this);
this.addInputConnection(connection);
}
/**
* Adds input connection with the given weight, from given neuron
*
* @param fromNeuron
* neuron to connect from
* @param weightVal
* connection weight value
*
*/
public void addInputConnection(Neuron fromNeuron, double weightVal) {
Connection connection = new Connection(fromNeuron, this, weightVal);
this.addInputConnection(connection);
}
/**
* Adds the specified output connection
*
* @param connection output connection to add
*/
protected void addOutputConnection(Connection connection) {
this.outConnections.add(connection);
}
/**
* Returns input connections for this neuron as Vector collection
*
* @return input connections of this neuron
*/
public List<Connection> getInputConnections() {
return inputConnections;
}
/**
* Returns output connections from this neuron
*
* @return output connections from this neuron
*/
public List<Connection> getOutConnections() {
return outConnections;
}
/**
* Removes input connection which is connected to specified neuron
*
* @param fromNeuron
* neuron which is connected as input
*/
public void removeInputConnectionFrom(Neuron fromNeuron) {
for(Connection connection : this.inputConnections) {
if (connection.getFromNeuron() == fromNeuron) {
this.inputConnections.remove(connection);
return;
}
}
}
/**
* Gets input connection from the specified neuron * @param fromNeuron
* neuron connected to this neuron as input
*/
public Connection getConnectionFrom(Neuron fromNeuron) {
for(Connection connection : this.inputConnections) {
if (connection.getFromNeuron() == fromNeuron)
return connection;
}
return null;
}
/**
* Sets input function
*
* @param inputFunction
* input function for this neuron
*/
public void setInputFunction(InputFunction inputFunction) {
this.inputFunction = inputFunction;
}
/**
* Sets transfer function
*
* @param transferFunction
* transfer function for this neuron
*/
public void setTransferFunction(TransferFunction transferFunction) {
this.transferFunction = transferFunction;
}
/**
* Returns input function
*
* @return input function
*/
public InputFunction getInputFunction() {
return this.inputFunction;
}
/**
* Returns transfer function
*
* @return transfer function
*/
public TransferFunction getTransferFunction() {
return this.transferFunction;
}
/**
* Sets reference to parent layer for this neuron (layer in which the neuron
* is located)
*
* @param parent
* reference on layer in which the cell is located
*/
public void setParentLayer(Layer parent) {
this.parentLayer = parent;
}
/**
* Returns reference to parent layer for this neuron
*
* @return parent layer for this neuron
*/
public Layer getParentLayer() {
return this.parentLayer;
}
/**
* Returns weights vector of input connections
*
* @return weights vector of input connections
*/
public Vector<Weight> getWeightsVector() {
Vector<Weight> weights = new Vector<Weight>();
for(Connection connection : this.inputConnections) {
Weight weight = connection.getWeight();
weights.addElement(weight);
}
return weights;
}
/**
* Returns error for this neuron. This is used by supervised learing rules.
*
* @return error for this neuron which is set by learning rule
*/
public double getError() {
return error;
}
/**
* Sets error for this neuron. This is used by supervised learing rules.
*
* @param error
* neuron error
*/
public void setError(double error) {
this.error = error;
}
/**
* Sets this neuron output
*
* @param output
* value to set
*/
public void setOutput(double output) {
this.output = output;
}
/**
* Randomize all input weights
*/
public void randomizeInputWeights() {
for(Connection connection : this.inputConnections) {
connection.getWeight().randomize();
}
}
/**
* Randomize all input weights within specified value range
*/
public void randomizeInputWeights(double minWeight, double maxWeight) {
for(Connection connection : this.inputConnections) {
connection.getWeight().randomize(minWeight, maxWeight);
}
}
/**
* Initialize weights for all input connections to specified value
* @param value the weight value
*/
public void initializeWeights(double value) {
for(Connection connection : this.inputConnections) {
connection.getWeight().setValue(value);
}
}
/**
* Initialize weights for all input connections to using random number generator
* @param generator the random number generator
*/
public void initializeWeights(Random generator) {
for(Connection connection : this.inputConnections) {
connection.getWeight().setValue(generator.nextDouble());
}
}
/**
* Initialize weights for all input connections with random value within specified interval
* @param min
* @param max
*/
public void initializeWeights(double min, double max) {
for(Connection connection : this.inputConnections) {
connection.getWeight().randomize(min, max);
}
}
}