/*
* Encog(tm) Core v3.3 - Java Version
* http://www.heatonresearch.com/encog/
* https://github.com/encog/encog-java-core
* Copyright 2008-2014 Heaton Research, Inc.
*
* 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.
*
* For more information on Heaton Research copyrights, licenses
* and trademarks visit:
* http://www.heatonresearch.com/copyright
*/
package org.encog.neural.som;
import org.encog.mathutil.matrices.Matrix;
import org.encog.ml.BasicML;
import org.encog.ml.MLClassification;
import org.encog.ml.MLError;
import org.encog.ml.MLResettable;
import org.encog.ml.data.MLData;
import org.encog.ml.data.MLDataPair;
import org.encog.ml.data.MLDataSet;
import org.encog.neural.NeuralNetworkError;
import org.encog.neural.som.training.basic.BestMatchingUnit;
import org.encog.util.EngineArray;
/**
* A self organizing map neural network.
*
*/
public class SOM extends BasicML implements MLClassification, MLResettable,
MLError {
/**
* Serial id.
*/
private static final long serialVersionUID = 1L;
/**
* Do not allow patterns to go below this very small number.
*/
public static final double VERYSMALL = 1.E-30;
/**
* The weights of the output neurons base on the input from the input
* neurons.
*/
private Matrix weights;
/**
* Default constructor.
*/
public SOM() {
}
/**
* The constructor.
*
* @param inputCount
* Number of input neurons
* @param outputCount
* Number of output neurons
*/
public SOM(final int inputCount, final int outputCount) {
this.weights = new Matrix(outputCount, inputCount);
}
/**
* {@inheritDoc}
*/
@Override
public double calculateError(final MLDataSet data) {
final BestMatchingUnit bmu = new BestMatchingUnit(this);
bmu.reset();
// Determine the BMU for each training element.
for (final MLDataPair pair : data) {
final MLData input = pair.getInput();
bmu.calculateBMU(input);
}
// update the error
return bmu.getWorstDistance() / 100.0;
}
/**
* {@inheritDoc}
*/
@Override
public int classify(final MLData input) {
if (input.size() > getInputCount()) {
throw new NeuralNetworkError(
"Can't classify SOM with input size of " + getInputCount()
+ " with input data of count " + input.size());
}
double[][] m = this.weights.getData();
double[] inputData = input.getData();
double minDist = Double.POSITIVE_INFINITY;
int result = -1;
for (int i = 0; i < getOutputCount(); i++) {
double dist = EngineArray.euclideanDistance(inputData, m[i]);
if (dist < minDist) {
minDist = dist;
result = i;
}
}
return result;
}
/**
* {@inheritDoc}
*/
@Override
public int getInputCount() {
return this.weights.getCols();
}
/**
* {@inheritDoc}
*/
@Override
public int getOutputCount() {
return this.weights.getRows();
}
/**
* @return the weights
*/
public Matrix getWeights() {
return this.weights;
}
/**
* {@inheritDoc}
*/
@Override
public void reset() {
this.weights.randomize(-1, 1);
}
/**
* {@inheritDoc}
*/
@Override
public void reset(final int seed) {
reset();
}
/**
* @param weights
* the weights to set
*/
public void setWeights(final Matrix weights) {
this.weights = weights;
}
/**
* {@inheritDoc}
*/
@Override
public void updateProperties() {
// unneeded
}
/**
* An alias for the classify method, kept for compatibility
* with earlier versions of Encog.
*
* @param input
* The input pattern.
* @return The winning neuron.
*/
public int winner(final MLData input) {
return classify(input);
}
}