Examples with Cluster de.lmu.ifi.dbs.elki.data.Cluster used on opensource projects

Example 16 with Cluster

use of de.lmu.ifi.dbs.elki.data.Cluster in project elki by elki-project.

the class DOC method makeCluster.

/**
 * Utility method to create a subspace cluster from a list of DBIDs and the
 * relevant attributes.
 *
 * @param relation to compute a centroid.
 * @param C the cluster points.
 * @param D the relevant dimensions.
 * @return an object representing the subspace cluster.
 */
protected Cluster<SubspaceModel> makeCluster(Relation<V> relation, DBIDs C, long[] D) {
    // copy, also to lose distance values!
    DBIDs ids = DBIDUtil.newHashSet(C);
    Cluster<SubspaceModel> cluster = new Cluster<>(ids);
    cluster.setModel(new SubspaceModel(new Subspace(D), Centroid.make(relation, ids).getArrayRef()));
    return cluster;
}

Example 17 with Cluster

use of de.lmu.ifi.dbs.elki.data.Cluster in project elki by elki-project.

the class PROCLUS method run.

/**
 * Performs the PROCLUS algorithm on the given database.
 *
 * @param database Database to process
 * @param relation Relation to process
 */
public Clustering<SubspaceModel> run(Database database, Relation<V> relation) {
    if (RelationUtil.dimensionality(relation) < l) {
        throw new IllegalStateException("Dimensionality of data < parameter l! (" + RelationUtil.dimensionality(relation) + " < " + l + ")");
    }
    DistanceQuery<V> distFunc = database.getDistanceQuery(relation, SquaredEuclideanDistanceFunction.STATIC);
    RangeQuery<V> rangeQuery = database.getRangeQuery(distFunc);
    final Random random = rnd.getSingleThreadedRandom();
    // initialization phase
    if (LOG.isVerbose()) {
        LOG.verbose("1. Initialization phase...");
    }
    int sampleSize = Math.min(relation.size(), k_i * k);
    DBIDs sampleSet = DBIDUtil.randomSample(relation.getDBIDs(), sampleSize, random);
    int medoidSize = Math.min(relation.size(), m_i * k);
    ArrayDBIDs medoids = greedy(distFunc, sampleSet, medoidSize, random);
    if (LOG.isDebugging()) {
        LOG.debugFine(// 
        new StringBuilder().append("sampleSize ").append(sampleSize).append('\n').append("sampleSet ").append(sampleSet).append(// 
        '\n').append("medoidSize ").append(medoidSize).append(// 
        '\n').append("m ").append(medoids).toString());
    }
    // iterative phase
    if (LOG.isVerbose()) {
        LOG.verbose("2. Iterative phase...");
    }
    double bestObjective = Double.POSITIVE_INFINITY;
    ArrayDBIDs m_best = null;
    DBIDs m_bad = null;
    ArrayDBIDs m_current = initialSet(medoids, k, random);
    if (LOG.isDebugging()) {
        LOG.debugFine(new StringBuilder().append("m_c ").append(m_current).toString());
    }
    IndefiniteProgress cprogress = LOG.isVerbose() ? new IndefiniteProgress("Current number of clusters:", LOG) : null;
    ArrayList<PROCLUSCluster> clusters = null;
    int loops = 0;
    while (loops < 10) {
        long[][] dimensions = findDimensions(m_current, relation, distFunc, rangeQuery);
        clusters = assignPoints(m_current, dimensions, relation);
        double objectiveFunction = evaluateClusters(clusters, dimensions, relation);
        if (objectiveFunction < bestObjective) {
            // restart counting loops
            loops = 0;
            bestObjective = objectiveFunction;
            m_best = m_current;
            m_bad = computeBadMedoids(m_current, clusters, (int) (relation.size() * 0.1 / k));
        }
        m_current = computeM_current(medoids, m_best, m_bad, random);
        loops++;
        if (cprogress != null) {
            cprogress.setProcessed(clusters.size(), LOG);
        }
    }
    LOG.setCompleted(cprogress);
    // refinement phase
    if (LOG.isVerbose()) {
        LOG.verbose("3. Refinement phase...");
    }
    List<Pair<double[], long[]>> dimensions = findDimensions(clusters, relation);
    List<PROCLUSCluster> finalClusters = finalAssignment(dimensions, relation);
    // build result
    int numClusters = 1;
    Clustering<SubspaceModel> result = new Clustering<>("ProClus clustering", "proclus-clustering");
    for (PROCLUSCluster c : finalClusters) {
        Cluster<SubspaceModel> cluster = new Cluster<>(c.objectIDs);
        cluster.setModel(new SubspaceModel(new Subspace(c.getDimensions()), c.centroid));
        cluster.setName("cluster_" + numClusters++);
        result.addToplevelCluster(cluster);
    }
    return result;
}

Example 18 with Cluster

use of de.lmu.ifi.dbs.elki.data.Cluster in project elki by elki-project.

the class SUBCLU method bestSubspace.

/**
 * Determines the {@code d}-dimensional subspace of the {@code (d+1)}
 * -dimensional candidate with minimal number of objects in the cluster.
 *
 * @param subspaces the list of {@code d}-dimensional subspaces containing
 *        clusters
 * @param candidate the {@code (d+1)}-dimensional candidate subspace
 * @param clusterMap the mapping of subspaces to clusters
 * @return the {@code d}-dimensional subspace of the {@code (d+1)}
 *         -dimensional candidate with minimal number of objects in the
 *         cluster
 */
private Subspace bestSubspace(List<Subspace> subspaces, Subspace candidate, TreeMap<Subspace, List<Cluster<Model>>> clusterMap) {
    Subspace bestSubspace = null;
    for (Subspace subspace : subspaces) {
        int min = Integer.MAX_VALUE;
        if (subspace.isSubspace(candidate)) {
            List<Cluster<Model>> clusters = clusterMap.get(subspace);
            for (Cluster<Model> cluster : clusters) {
                int clusterSize = cluster.size();
                if (clusterSize < min) {
                    min = clusterSize;
                    bestSubspace = subspace;
                }
            }
        }
    }
    return bestSubspace;
}

Example 19 with Cluster

use of de.lmu.ifi.dbs.elki.data.Cluster in project elki by elki-project.

the class CLIQUE method run.

/**
 * Performs the CLIQUE algorithm on the given database.
 *
 * @param relation Data relation to process
 * @return Clustering result
 */
public Clustering<SubspaceModel> run(Relation<V> relation) {
    final int dimensionality = RelationUtil.dimensionality(relation);
    StepProgress step = new StepProgress(2);
    // 1. Identification of subspaces that contain clusters
    step.beginStep(1, "Identification of subspaces that contain clusters", LOG);
    ArrayList<List<CLIQUESubspace<V>>> dimensionToDenseSubspaces = new ArrayList<>(dimensionality);
    List<CLIQUESubspace<V>> denseSubspaces = findOneDimensionalDenseSubspaces(relation);
    dimensionToDenseSubspaces.add(denseSubspaces);
    if (LOG.isVerbose()) {
        LOG.verbose("1-dimensional dense subspaces: " + denseSubspaces.size());
    }
    if (LOG.isDebugging()) {
        for (CLIQUESubspace<V> s : denseSubspaces) {
            LOG.debug(s.toString(" "));
        }
    }
    for (int k = 2; k <= dimensionality && !denseSubspaces.isEmpty(); k++) {
        denseSubspaces = findDenseSubspaces(relation, denseSubspaces);
        assert (dimensionToDenseSubspaces.size() == k - 1);
        dimensionToDenseSubspaces.add(denseSubspaces);
        if (LOG.isVerbose()) {
            LOG.verbose(k + "-dimensional dense subspaces: " + denseSubspaces.size());
        }
        if (LOG.isDebugging()) {
            for (CLIQUESubspace<V> s : denseSubspaces) {
                LOG.debug(s.toString(" "));
            }
        }
    }
    // 2. Identification of clusters
    step.beginStep(2, "Identification of clusters", LOG);
    // build result
    Clustering<SubspaceModel> result = new Clustering<>("CLIQUE clustering", "clique-clustering");
    for (int dim = 0; dim < dimensionToDenseSubspaces.size(); dim++) {
        List<CLIQUESubspace<V>> subspaces = dimensionToDenseSubspaces.get(dim);
        List<Pair<Subspace, ModifiableDBIDs>> modelsAndClusters = determineClusters(subspaces);
        if (LOG.isVerbose()) {
            LOG.verbose((dim + 1) + "-dimensional clusters: " + modelsAndClusters.size());
        }
        for (Pair<Subspace, ModifiableDBIDs> modelAndCluster : modelsAndClusters) {
            Cluster<SubspaceModel> newCluster = new Cluster<>(modelAndCluster.second);
            newCluster.setModel(new SubspaceModel(modelAndCluster.first, Centroid.make(relation, modelAndCluster.second).getArrayRef()));
            result.addToplevelCluster(newCluster);
        }
    }
    return result;
}

Example 20 with Cluster

use of de.lmu.ifi.dbs.elki.data.Cluster in project elki by elki-project.

the class ByLabelOrAllInOneClustering method run.

@Override
public Clustering<Model> run(Database database) {
    // Prefer a true class label
    try {
        Relation<ClassLabel> relation = database.getRelation(TypeUtil.CLASSLABEL);
        return run(relation);
    } catch (NoSupportedDataTypeException e) {
    // Ignore.
    }
    try {
        Relation<ClassLabel> relation = database.getRelation(TypeUtil.GUESSED_LABEL);
        return run(relation);
    } catch (NoSupportedDataTypeException e) {
    // Ignore.
    }
    final DBIDs ids = database.getRelation(TypeUtil.ANY).getDBIDs();
    Clustering<Model> result = new Clustering<>("All-in-one trivial Clustering", "allinone-clustering");
    Cluster<Model> c = new Cluster<Model>(ids, ClusterModel.CLUSTER);
    result.addToplevelCluster(c);
    return result;
}