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

Example 16 with Clustering

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

the class ERiC method run.

/**
 * Performs the ERiC algorithm on the given database.
 *
 * @param relation Relation to process
 * @return Clustering result
 */
public Clustering<CorrelationModel> run(Database database, Relation<V> relation) {
    final int dimensionality = RelationUtil.dimensionality(relation);
    StepProgress stepprog = LOG.isVerbose() ? new StepProgress(3) : null;
    // Run Generalized DBSCAN
    LOG.beginStep(stepprog, 1, "Preprocessing local correlation dimensionalities and partitioning data");
    // FIXME: how to ensure we are running on the same relation?
    ERiCNeighborPredicate<V>.Instance npred = new ERiCNeighborPredicate<V>(settings).instantiate(database, relation);
    CorePredicate.Instance<DBIDs> cpred = new MinPtsCorePredicate(settings.minpts).instantiate(database);
    Clustering<Model> copacResult = new GeneralizedDBSCAN.Instance<>(npred, cpred, false).run();
    // extract correlation clusters
    LOG.beginStep(stepprog, 2, "Extract correlation clusters");
    List<List<Cluster<CorrelationModel>>> clusterMap = extractCorrelationClusters(copacResult, relation, dimensionality, npred);
    if (LOG.isDebugging()) {
        StringBuilder msg = new StringBuilder("Step 2: Extract correlation clusters...");
        for (int corrDim = 0; corrDim < clusterMap.size(); corrDim++) {
            List<Cluster<CorrelationModel>> correlationClusters = clusterMap.get(corrDim);
            msg.append("\n\ncorrDim ").append(corrDim);
            for (Cluster<CorrelationModel> cluster : correlationClusters) {
                msg.append("\n  cluster ").append(cluster).append(", ids: ").append(cluster.getIDs().size());
            // .append(", level: ").append(cluster.getLevel()).append(", index:
            // ").append(cluster.getLevelIndex());
            // msg.append("\n basis " +
            // cluster.getPCA().getWeakEigenvectors().toString(" ", NF) +
            // " ids " + cluster.getIDs().size());
            }
        }
        LOG.debugFine(msg.toString());
    }
    if (LOG.isVerbose()) {
        int clusters = 0;
        for (List<Cluster<CorrelationModel>> correlationClusters : clusterMap) {
            clusters += correlationClusters.size();
        }
        LOG.verbose(clusters + " clusters extracted.");
    }
    // build hierarchy
    LOG.beginStep(stepprog, 3, "Building hierarchy");
    Clustering<CorrelationModel> clustering = new Clustering<>("ERiC clustering", "eric-clustering");
    buildHierarchy(clustering, clusterMap, npred);
    if (LOG.isDebugging()) {
        StringBuilder msg = new StringBuilder("Step 3: Build hierarchy");
        for (int corrDim = 0; corrDim < clusterMap.size(); corrDim++) {
            List<Cluster<CorrelationModel>> correlationClusters = clusterMap.get(corrDim);
            for (Cluster<CorrelationModel> cluster : correlationClusters) {
                msg.append("\n  cluster ").append(cluster).append(", ids: ").append(cluster.getIDs().size());
                // ").append(cluster.getLevelIndex());
                for (It<Cluster<CorrelationModel>> iter = clustering.getClusterHierarchy().iterParents(cluster); iter.valid(); iter.advance()) {
                    msg.append("\n   parent ").append(iter.get());
                }
                for (It<Cluster<CorrelationModel>> iter = clustering.getClusterHierarchy().iterChildren(cluster); iter.valid(); iter.advance()) {
                    msg.append("\n   child ").append(iter.get());
                }
            }
        }
        LOG.debugFine(msg.toString());
    }
    LOG.setCompleted(stepprog);
    for (Cluster<CorrelationModel> rc : clusterMap.get(clusterMap.size() - 1)) {
        clustering.addToplevelCluster(rc);
    }
    return clustering;
}

Example 17 with Clustering

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

the class CanopyPreClustering method run.

/**
 * Run the algorithm
 *
 * @param database Database
 * @param relation Relation to process
 */
public Clustering<PrototypeModel<O>> run(Database database, Relation<O> relation) {
    if (!(t1 >= t2)) {
        throw new AbortException("T1 must be at least as large as T2.");
    }
    DistanceQuery<O> dq = database.getDistanceQuery(relation, getDistanceFunction());
    ModifiableDBIDs ids = DBIDUtil.newHashSet(relation.getDBIDs());
    ArrayList<Cluster<PrototypeModel<O>>> clusters = new ArrayList<>();
    final int size = relation.size();
    FiniteProgress prog = LOG.isVerbose() ? new FiniteProgress("Canopy clustering", size, LOG) : null;
    DBIDVar first = DBIDUtil.newVar();
    while (!ids.isEmpty()) {
        // Remove first element:
        ids.pop(first);
        // Start a new cluster:
        ModifiableDBIDs cids = DBIDUtil.newArray();
        cids.add(first);
        // Compare to remaining objects:
        for (DBIDMIter iter = ids.iter(); iter.valid(); iter.advance()) {
            double dist = dq.distance(first, iter);
            // Inclusion threshold:
            if (dist > t1) {
                continue;
            }
            cids.add(iter);
            // Removal threshold:
            if (dist <= t2) {
                iter.remove();
            }
        }
        // TODO: remember the central object using a CanopyModel?
        // Construct cluster:
        clusters.add(new Cluster<>(cids, new SimplePrototypeModel<>(relation.get(first))));
        if (prog != null) {
            prog.setProcessed(size - ids.size(), LOG);
        }
    }
    LOG.ensureCompleted(prog);
    return new Clustering<>("Canopy clustering", "canopy-clustering", clusters);
}

Example 18 with Clustering

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

the class AffinityPropagationClusteringAlgorithm method run.

/**
 * Perform affinity propagation clustering.
 *
 * @param db Database
 * @param relation Relation
 * @return Clustering result
 */
public Clustering<MedoidModel> run(Database db, Relation<O> relation) {
    ArrayDBIDs ids = DBIDUtil.ensureArray(relation.getDBIDs());
    final int size = ids.size();
    int[] assignment = new int[size];
    double[][] s = initialization.getSimilarityMatrix(db, relation, ids);
    double[][] r = new double[size][size];
    double[][] a = new double[size][size];
    IndefiniteProgress prog = LOG.isVerbose() ? new IndefiniteProgress("Affinity Propagation Iteration", LOG) : null;
    MutableProgress aprog = LOG.isVerbose() ? new MutableProgress("Stable assignments", size + 1, LOG) : null;
    int inactive = 0;
    for (int iteration = 0; iteration < maxiter && inactive < convergence; iteration++) {
        // Update responsibility matrix:
        for (int i = 0; i < size; i++) {
            double[] ai = a[i], ri = r[i], si = s[i];
            // Find the two largest values (as initially maxk == i)
            double max1 = Double.NEGATIVE_INFINITY, max2 = Double.NEGATIVE_INFINITY;
            int maxk = -1;
            for (int k = 0; k < size; k++) {
                double val = ai[k] + si[k];
                if (val > max1) {
                    max2 = max1;
                    max1 = val;
                    maxk = k;
                } else if (val > max2) {
                    max2 = val;
                }
            }
            // With the maximum value known, update r:
            for (int k = 0; k < size; k++) {
                double val = si[k] - ((k != maxk) ? max1 : max2);
                ri[k] = ri[k] * lambda + val * (1. - lambda);
            }
        }
        // Update availability matrix
        for (int k = 0; k < size; k++) {
            // Compute sum of max(0, r_ik) for all i.
            // For r_kk, don't apply the max.
            double colposum = 0.;
            for (int i = 0; i < size; i++) {
                if (i == k || r[i][k] > 0.) {
                    colposum += r[i][k];
                }
            }
            for (int i = 0; i < size; i++) {
                double val = colposum;
                // Adjust column sum by the one extra term.
                if (i == k || r[i][k] > 0.) {
                    val -= r[i][k];
                }
                if (i != k && val > 0.) {
                    // min
                    val = 0.;
                }
                a[i][k] = a[i][k] * lambda + val * (1 - lambda);
            }
        }
        int changed = 0;
        for (int i = 0; i < size; i++) {
            double[] ai = a[i], ri = r[i];
            double max = Double.NEGATIVE_INFINITY;
            int maxj = -1;
            for (int j = 0; j < size; j++) {
                double v = ai[j] + ri[j];
                if (v > max || (i == j && v >= max)) {
                    max = v;
                    maxj = j;
                }
            }
            if (assignment[i] != maxj) {
                changed += 1;
                assignment[i] = maxj;
            }
        }
        inactive = (changed > 0) ? 0 : (inactive + 1);
        LOG.incrementProcessed(prog);
        if (aprog != null) {
            aprog.setProcessed(size - changed, LOG);
        }
    }
    if (aprog != null) {
        aprog.setProcessed(aprog.getTotal(), LOG);
    }
    LOG.setCompleted(prog);
    // Cluster map, by lead object
    Int2ObjectOpenHashMap<ModifiableDBIDs> map = new Int2ObjectOpenHashMap<>();
    DBIDArrayIter i1 = ids.iter();
    for (int i = 0; i1.valid(); i1.advance(), i++) {
        int c = assignment[i];
        // Add to cluster members:
        ModifiableDBIDs cids = map.get(c);
        if (cids == null) {
            cids = DBIDUtil.newArray();
            map.put(c, cids);
        }
        cids.add(i1);
    }
    // If we stopped early, the cluster lead might be in a different cluster.
    for (ObjectIterator<Int2ObjectOpenHashMap.Entry<ModifiableDBIDs>> iter = map.int2ObjectEntrySet().fastIterator(); iter.hasNext(); ) {
        Int2ObjectOpenHashMap.Entry<ModifiableDBIDs> entry = iter.next();
        final int key = entry.getIntKey();
        int targetkey = key;
        ModifiableDBIDs tids = null;
        // Chase arrows:
        while (ids == null && assignment[targetkey] != targetkey) {
            targetkey = assignment[targetkey];
            tids = map.get(targetkey);
        }
        if (tids != null && targetkey != key) {
            tids.addDBIDs(entry.getValue());
            iter.remove();
        }
    }
    Clustering<MedoidModel> clustering = new Clustering<>("Affinity Propagation Clustering", "ap-clustering");
    ModifiableDBIDs noise = DBIDUtil.newArray();
    for (ObjectIterator<Int2ObjectOpenHashMap.Entry<ModifiableDBIDs>> iter = map.int2ObjectEntrySet().fastIterator(); iter.hasNext(); ) {
        Int2ObjectOpenHashMap.Entry<ModifiableDBIDs> entry = iter.next();
        i1.seek(entry.getIntKey());
        if (entry.getValue().size() > 1) {
            MedoidModel mod = new MedoidModel(DBIDUtil.deref(i1));
            clustering.addToplevelCluster(new Cluster<>(entry.getValue(), mod));
        } else {
            noise.add(i1);
        }
    }
    if (noise.size() > 0) {
        MedoidModel mod = new MedoidModel(DBIDUtil.deref(noise.iter()));
        clustering.addToplevelCluster(new Cluster<>(noise, true, mod));
    }
    return clustering;
}

Example 19 with Clustering

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

the class NaiveAgglomerativeHierarchicalClustering3 method run.

/**
 * Run the algorithm
 *
 * @param db Database
 * @param relation Relation
 * @return Clustering hierarchy
 */
public Result run(Database db, Relation<O> relation) {
    DistanceQuery<O> dq = db.getDistanceQuery(relation, getDistanceFunction());
    ArrayDBIDs ids = DBIDUtil.ensureArray(relation.getDBIDs());
    final int size = ids.size();
    if (size > 0x10000) {
        throw new AbortException("This implementation does not scale to data sets larger than " + 0x10000 + " instances (~17 GB RAM), which results in an integer overflow.");
    }
    if (Linkage.SINGLE.equals(linkage)) {
        LOG.verbose("Notice: SLINK is a much faster algorithm for single-linkage clustering!");
    }
    // Compute the initial (lower triangular) distance matrix.
    double[] scratch = new double[triangleSize(size)];
    DBIDArrayIter ix = ids.iter(), iy = ids.iter();
    // Position counter - must agree with computeOffset!
    int pos = 0;
    boolean square = Linkage.WARD.equals(linkage) && !getDistanceFunction().isSquared();
    for (int x = 0; ix.valid(); x++, ix.advance()) {
        iy.seek(0);
        for (int y = 0; y < x; y++, iy.advance()) {
            scratch[pos] = dq.distance(ix, iy);
            // Ward uses variances -- i.e. squared values
            if (square) {
                scratch[pos] *= scratch[pos];
            }
            pos++;
        }
    }
    // Initialize space for result:
    double[] height = new double[size];
    Arrays.fill(height, Double.POSITIVE_INFINITY);
    // Parent node, to track merges
    // have every object point to itself initially
    ArrayModifiableDBIDs parent = DBIDUtil.newArray(ids);
    // Active clusters, when not trivial.
    Int2ReferenceMap<ModifiableDBIDs> clusters = new Int2ReferenceOpenHashMap<>();
    // Repeat until everything merged, except the desired number of clusters:
    final int stop = size - numclusters;
    FiniteProgress prog = LOG.isVerbose() ? new FiniteProgress("Agglomerative clustering", stop, LOG) : null;
    for (int i = 0; i < stop; i++) {
        double min = Double.POSITIVE_INFINITY;
        int minx = -1, miny = -1;
        for (int x = 0; x < size; x++) {
            if (height[x] < Double.POSITIVE_INFINITY) {
                continue;
            }
            final int xbase = triangleSize(x);
            for (int y = 0; y < x; y++) {
                if (height[y] < Double.POSITIVE_INFINITY) {
                    continue;
                }
                final int idx = xbase + y;
                if (scratch[idx] < min) {
                    min = scratch[idx];
                    minx = x;
                    miny = y;
                }
            }
        }
        assert (minx >= 0 && miny >= 0);
        // Avoid allocating memory, by reusing existing iterators:
        ix.seek(minx);
        iy.seek(miny);
        // Perform merge in data structure: x -> y
        // Since y < x, prefer keeping y, dropping x.
        height[minx] = min;
        parent.set(minx, iy);
        // Merge into cluster
        ModifiableDBIDs cx = clusters.get(minx);
        ModifiableDBIDs cy = clusters.get(miny);
        // cluster sizes, for averaging
        int sizex = 1, sizey = 1;
        if (cy == null) {
            cy = DBIDUtil.newHashSet();
            cy.add(iy);
        } else {
            sizey = cy.size();
        }
        if (cx == null) {
            cy.add(ix);
        } else {
            sizex = cx.size();
            cy.addDBIDs(cx);
            clusters.remove(minx);
        }
        clusters.put(miny, cy);
        // Update distance matrix. Note: miny < minx
        // Implementation note: most will not need sizej, and could save the
        // hashmap lookup.
        final int xbase = triangleSize(minx), ybase = triangleSize(miny);
        // Write to (y, j), with j < y
        for (int j = 0; j < miny; j++) {
            if (height[j] < Double.POSITIVE_INFINITY) {
                continue;
            }
            final DBIDs idsj = clusters.get(j);
            final int sizej = (idsj == null) ? 1 : idsj.size();
            scratch[ybase + j] = linkage.combine(sizex, scratch[xbase + j], sizey, scratch[ybase + j], sizej, min);
        }
        // Write to (j, y), with y < j < x
        for (int j = miny + 1; j < minx; j++) {
            if (height[j] < Double.POSITIVE_INFINITY) {
                continue;
            }
            final int jbase = triangleSize(j);
            final DBIDs idsj = clusters.get(j);
            final int sizej = (idsj == null) ? 1 : idsj.size();
            scratch[jbase + miny] = linkage.combine(sizex, scratch[xbase + j], sizey, scratch[jbase + miny], sizej, min);
        }
        // Write to (j, y), with y < x < j
        for (int j = minx + 1; j < size; j++) {
            if (height[j] < Double.POSITIVE_INFINITY) {
                continue;
            }
            final DBIDs idsj = clusters.get(j);
            final int sizej = (idsj == null) ? 1 : idsj.size();
            final int jbase = triangleSize(j);
            scratch[jbase + miny] = linkage.combine(sizex, scratch[jbase + minx], sizey, scratch[jbase + miny], sizej, min);
        }
        LOG.incrementProcessed(prog);
    }
    LOG.ensureCompleted(prog);
    // Build the clustering result
    final Clustering<Model> dendrogram = new Clustering<>("Hierarchical-Clustering", "hierarchical-clustering");
    for (int x = 0; x < size; x++) {
        if (height[x] < Double.POSITIVE_INFINITY) {
            DBIDs cids = clusters.get(x);
            if (cids == null) {
                ix.seek(x);
                cids = DBIDUtil.deref(ix);
            }
            Cluster<Model> cluster = new Cluster<>("Cluster", cids);
            dendrogram.addToplevelCluster(cluster);
        }
    }
    return dendrogram;
}

Example 20 with Clustering

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

the class UKMeans method run.

/**
 * Run the clustering.
 *
 * @param database the Database
 * @param relation the Relation
 * @return Clustering result
 */
public Clustering<?> run(final Database database, final Relation<DiscreteUncertainObject> relation) {
    if (relation.size() <= 0) {
        return new Clustering<>("Uk-Means Clustering", "ukmeans-clustering");
    }
    // Choose initial means randomly
    DBIDs sampleids = DBIDUtil.randomSample(relation.getDBIDs(), k, rnd);
    List<double[]> means = new ArrayList<>(k);
    for (DBIDIter iter = sampleids.iter(); iter.valid(); iter.advance()) {
        means.add(ArrayLikeUtil.toPrimitiveDoubleArray(relation.get(iter).getCenterOfMass()));
    }
    // Setup cluster assignment store
    List<ModifiableDBIDs> clusters = new ArrayList<>();
    for (int i = 0; i < k; i++) {
        clusters.add(DBIDUtil.newHashSet((int) (relation.size() * 2. / k)));
    }
    WritableIntegerDataStore assignment = DataStoreUtil.makeIntegerStorage(relation.getDBIDs(), DataStoreFactory.HINT_TEMP | DataStoreFactory.HINT_HOT, -1);
    double[] varsum = new double[k];
    IndefiniteProgress prog = LOG.isVerbose() ? new IndefiniteProgress("UK-Means iteration", LOG) : null;
    DoubleStatistic varstat = LOG.isStatistics() ? new DoubleStatistic(this.getClass().getName() + ".variance-sum") : null;
    int iteration = 0;
    for (; maxiter <= 0 || iteration < maxiter; iteration++) {
        LOG.incrementProcessed(prog);
        boolean changed = assignToNearestCluster(relation, means, clusters, assignment, varsum);
        logVarstat(varstat, varsum);
        // Stop if no cluster assignment changed.
        if (!changed) {
            break;
        }
        // Recompute means.
        means = means(clusters, means, relation);
    }
    LOG.setCompleted(prog);
    if (LOG.isStatistics()) {
        LOG.statistics(new LongStatistic(KEY + ".iterations", iteration));
    }
    // Wrap result
    Clustering<KMeansModel> result = new Clustering<>("Uk-Means Clustering", "ukmeans-clustering");
    for (int i = 0; i < clusters.size(); i++) {
        DBIDs ids = clusters.get(i);
        if (ids.isEmpty()) {
            continue;
        }
        result.addToplevelCluster(new Cluster<>(ids, new KMeansModel(means.get(i), varsum[i])));
    }
    return result;
}