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.clustering.spectral;
031
032import java.util.Iterator;
033
034import org.apache.log4j.Logger;
035import org.openimaj.ml.clustering.DataClusterer;
036import org.openimaj.ml.clustering.IndexClusters;
037import org.openimaj.ml.clustering.SpatialClusterer;
038import org.openimaj.ml.clustering.SpatialClusters;
039import org.openimaj.util.pair.DoubleObjectPair;
040import org.openimaj.util.pair.IndependentPair;
041
042import ch.akuhn.matrix.Vector;
043import ch.akuhn.matrix.Vector.Entry;
044import ch.akuhn.matrix.eigenvalues.Eigenvalues;
045
046/**
047 * For a given set of {@link Eigenvalues} perform the stages of spectral clustering
048 * which involve the selection of the best eigen values and the calling of an internal clustering
049 * algorithm
050 * @author Sina Samangooei (ss@ecs.soton.ac.uk)
051 */
052public class PreparedSpectralClustering implements DataClusterer<Eigenvalues, SpectralIndexedClusters>{
053        final static Logger logger = Logger.getLogger(PreparedSpectralClustering.class);
054        private SpectralClusteringConf<double[]> conf;
055
056        /**
057         * @param conf
058         */
059        public PreparedSpectralClustering(SpectralClusteringConf<double[]> conf) {
060                this.conf = conf;
061        }
062
063        @Override
064        public int[][] performClustering(Eigenvalues data) {
065                return cluster(data).clusters();
066        }
067
068        @Override
069        public SpectralIndexedClusters cluster(Eigenvalues eig) {
070                // Also normalise each row
071                IndependentPair<double[], double[][]> lowestCols = bestCols(eig);
072                // Using the eigenspace, cluster
073                return eigenspaceCluster(lowestCols);
074        }
075        
076        protected SpectralIndexedClusters eigenspaceCluster(IndependentPair<double[], double[][]> lowestCols) {
077                SpatialClusterer<? extends SpatialClusters<double[]>, double[]> clusterer = conf.internal.apply(lowestCols);
078                // Cluster the rows with the internal spatial clusterer
079                SpatialClusters<double[]> cluster = clusterer.cluster(lowestCols.getSecondObject());
080                // if the clusters contain the cluster indexes of the training examples use those
081                if(cluster instanceof IndexClusters){
082                        IndexClusters clusters = new IndexClusters(((IndexClusters)cluster).clusters());
083//                      logger.debug(clusters);
084                        return new SpectralIndexedClusters(clusters, lowestCols);
085                }
086                // Otherwise attempt to assign values to clusters
087                int[] clustered = cluster.defaultHardAssigner().assign(lowestCols.getSecondObject());
088                // done!
089                return new SpectralIndexedClusters(new IndexClusters(clustered),lowestCols);
090        }
091        
092        protected IndependentPair<double[], double[][]> bestCols(final Eigenvalues eig) {
093                
094                
095                int eigenVectorSelect = conf.eigenChooser.nEigenVectors(this.conf.laplacian.eigenIterator(eig), eig.getN());
096                int eigenVectorSkip = this.conf.skipEigenVectors;
097                logger.debug("Selected dimensions: " + eigenVectorSelect);
098                logger.debug("Skipping dimesions: " + eigenVectorSkip);
099                eigenVectorSelect -= eigenVectorSkip;
100
101
102                int nrows = eig.vector[0].size();
103                double[][] ret = new double[nrows][eigenVectorSelect];
104                double[] retSum = new double[nrows];
105                double[] eigvals = new double[eigenVectorSelect];
106                Iterator<DoubleObjectPair<Vector>> iterator = this.conf.laplacian.eigenIterator(eig);
107                // Skip a few at the beggining
108                for (int i = 0; i < eigenVectorSkip; i++) iterator.next();
109                int col = 0;
110                // Calculate U matrix (containing n smallests eigen valued columns)
111                for (; iterator.hasNext();) {
112                        DoubleObjectPair<Vector> v = iterator.next();
113                        eigvals[col] = v.first;
114                        
115                        for (Entry d : v.second.entries()) {
116                                double elColI = d.value;
117                                if(conf.eigenValueScale){
118                                        elColI *= Math.sqrt(v.first);
119                                }
120                                ret[d.index][col] = elColI;
121                                retSum[d.index] += elColI * elColI;
122                        }
123                        col++;
124                        if(col == eigenVectorSelect) break;
125                }
126
127                if(!conf.eigenValueScale){                      
128                        // normalise rows
129                        for (int i = 0; i < ret.length; i++) {
130                                double[] row = ret[i];
131                                for (int j = 0; j < row.length; j++) {
132                                        row[j] /= Math.sqrt(retSum[i]);
133                                }
134                        }
135                }
136
137                return IndependentPair.pair(eigvals, ret);
138        }
139
140}