001/**
002 * Copyright (c) 2011, The University of Southampton and the individual contributors.
003 * All rights reserved.
004 *
005 * Redistribution and use in source and binary forms, with or without modification,
006 * are permitted provided that the following conditions are met:
007 *
008 *   *  Redistributions of source code must retain the above copyright notice,
009 *      this list of conditions and the following disclaimer.
010 *
011 *   *  Redistributions in binary form must reproduce the above copyright notice,
012 *      this list of conditions and the following disclaimer in the documentation
013 *      and/or other materials provided with the distribution.
014 *
015 *   *  Neither the name of the University of Southampton nor the names of its
016 *      contributors may be used to endorse or promote products derived from this
017 *      software without specific prior written permission.
018 *
019 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
020 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
021 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
022 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
023 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
024 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
025 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
026 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
027 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
028 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
029 */
030package org.openimaj.ml.linear.learner.perceptron;
031
032import java.util.ArrayList;
033import java.util.Arrays;
034import java.util.HashMap;
035import java.util.List;
036import java.util.Map;
037
038import org.openimaj.ml.linear.kernel.VectorKernel;
039import org.openimaj.util.pair.IndependentPair;
040
041import ch.akuhn.matrix.Matrix;
042
043/**
044 * An implementation of a simple {@link KernelPerceptron} which works with
045 * {@link Matrix} inputs and is binary.
046 * @author Sina Samangooei (ss@ecs.soton.ac.uk)
047 */
048public class MatrixKernelPerceptron extends KernelPerceptron<double[], PerceptronClass>{
049
050        class WrappedDouble{
051                private double[] d;
052
053                public WrappedDouble(double[] d) {
054                        this.d = d;
055                }
056                @Override
057                public boolean equals(Object obj) {
058                        if(obj instanceof WrappedDouble){
059                                WrappedDouble that = (WrappedDouble) obj;
060                                return Arrays.equals(d, that.d);
061                        }
062                        return false;
063                }
064                
065                 @Override
066                public int hashCode() {
067                        return Arrays.hashCode(d);
068                }
069        }
070        protected List<double[]> supports = new ArrayList<double[]>();
071        protected List<Double> weights = new ArrayList<Double>();
072        
073        Map<WrappedDouble,Integer> index = new HashMap<WrappedDouble, Integer>();
074        
075        
076        /**
077         * @param k the kernel
078         */
079        public MatrixKernelPerceptron(VectorKernel k) {
080                super(k);
081        }
082        
083        public double[] correct(double[] in) {
084                return in.clone();
085        }
086        
087        protected double mapping(double[] in){
088                double ret = getBias();
089                in = correct(in);
090                for (int i = 0; i < supports.size(); i++) {
091                        double alpha = this.weights.get(i);
092                        double[] x_i = correct(this.supports.get(i));
093                        ret += alpha * kernel.apply(IndependentPair.pair(x_i, in));
094                        
095                }
096                return ret;
097        }
098        
099        @Override
100        public PerceptronClass predict(double[] x) {
101                return PerceptronClass.fromSign(Math.signum(mapping(x)));
102        }
103
104        @Override
105        public void update(double[] xt, PerceptronClass yt, PerceptronClass yt_prime) {
106                WrappedDouble d = new WrappedDouble(xt);
107                double updateAmount = this.getUpdateRate() * (double) yt.v();
108                if(!this.index.containsKey(d)){
109                        this.index.put(d, this.supports.size());
110                        this.supports.add(xt);
111                        this.weights.add(updateAmount);
112                } else {
113                        int index = this.index.get(d);
114                        this.weights.set(index, this.weights.get(index) + updateAmount);
115                }
116        }
117
118        double getUpdateRate() {
119                return 1;
120        }
121
122        @Override
123        public List<double[]> getSupports() {
124                return this.supports;
125        }
126
127        @Override
128        public List<Double> getWeights() {
129                return this.weights;
130        }
131
132        @Override
133        public double getBias() {
134                double bias = 0;
135                for (double d : this.weights) {
136                        bias += d;
137                }
138                return bias;
139        }
140        
141        
142        
143
144}