Examples of de.lmu.ifi.dbs.elki.database.ids.ModifiableDBIDs

Package de.lmu.ifi.dbs.elki.database.ids

Examples of de.lmu.ifi.dbs.elki.database.ids.ModifiableDBIDs

de.lmu.ifi.dbs.elki.database.ids.ModifiableDBIDs
Interface for a generic modifiable DBID collection. @author Erich Schubert

          e = exp / dist;
        }
        E.set(j, i, e);
      }
      // Convert kNN Heap into DBID array
      ModifiableDBIDs nids = DBIDUtil.newArray(heap.size());
      while(!heap.isEmpty()) {
        nids.add(heap.poll().getDBID());
      }
      neighbors.put(id, nids);
    }
    // normalize the adjacent Matrix
    // Sum based normalization - don't use E.normalizeColumns()

View Full Code Here

   * @param p1 First Point of the selected rectangle
   * @param p2 Second Point of the selected rectangle
   */
  private void updateSelection(Projection proj, SVGPoint p1, SVGPoint p2) {
    DBIDSelection selContext = context.getSelection();
    ModifiableDBIDs selection;
    if(selContext != null) {
      selection = DBIDUtil.newHashSet(selContext.getSelectedIds());
    }
    else {
      selection = DBIDUtil.newHashSet();
    }
    DoubleDoublePair[] ranges;


    if(p1 == null || p2 == null) {
      logger.warning("no rect selected: p1: " + p1 + " p2: " + p2);
    }
    else {
      double x1 = Math.min(p1.getX(), p2.getX());
      double x2 = Math.max(p1.getX(), p2.getX());
      double y1 = Math.max(p1.getY(), p2.getY());
      double y2 = Math.min(p1.getY(), p2.getY());


      if(selContext instanceof RangeSelection) {
        ranges = ((RangeSelection) selContext).getRanges();
      }
      else {
        ranges = new DoubleDoublePair[dim];
      }
      updateSelectionRectKoordinates(x1, x2, y1, y2, ranges);


      selection.clear();
      boolean idIn = true;
      for(DBID id : rel.iterDBIDs()) {
        NV dbTupel = rel.get(id);
        idIn = true;
        for(int i = 0; i < dim; i++) {
          if(ranges != null && ranges[i] != null) {
            if(dbTupel.doubleValue(i + 1) < ranges[i].first || dbTupel.doubleValue(i + 1) > ranges[i].second) {
              idIn = false;
              break;
            }
          }
        }
        if(idIn == true) {
          selection.add(id);
        }
      }
      context.setSelection(new RangeSelection(selection, ranges));
    }
  }

View Full Code Here

  protected void handleDelete() {
    if (!(database instanceof UpdatableDatabase)) {
      throw new UnsupportedOperationException("Database not updatable.");
    }
    UpdatableDatabase upd = (UpdatableDatabase) database;
    ModifiableDBIDs todel = DBIDUtil.newHashSet();
    ModifiableDBIDs remain = DBIDUtil.newHashSet(dbids);
    for(int row : table.getSelectedRows()) {
      final DBID id = dbids.get(row);
      todel.add(id);
      remain.remove(id);
    }
    // Unselect first ...
    context.setSelection(new DBIDSelection(remain));
    // Now delete them.
    for(DBID id : todel) {

View Full Code Here

    if(!initialized) {
      initialize(entries.get(0));
    }


    Map<DBID, KNNHeap<D>> knnHeaps = new HashMap<DBID, KNNHeap<D>>(entries.size());
    ModifiableDBIDs ids = DBIDUtil.newArray(entries.size());


    // insert
    for(MkCoPEntry<D> entry : entries) {
      DBID id = entry.getRoutingObjectID();
      // create knnList for the object
      knnHeaps.put(id, new KNNHeap<D>(k_max + 1, getDistanceQuery().infiniteDistance()));


      ids.add(id);
      // insert the object
      super.insert(entry, false);
    }


    // do batch nn

View Full Code Here

    if(k > this.k_max) {
      throw new IllegalArgumentException("Parameter k has to be less or equal than " + "parameter kmax of the MCop-Tree!");
    }


    List<DistanceResultPair<D>> result = new ArrayList<DistanceResultPair<D>>();
    ModifiableDBIDs candidates = DBIDUtil.newArray();
    doReverseKNNQuery(k, id, result, candidates);


    // refinement of candidates
    Map<DBID, KNNHeap<D>> knnLists = new HashMap<DBID, KNNHeap<D>>();
    for(DBID cid : candidates) {
      knnLists.put(cid, new KNNHeap<D>(k, getDistanceQuery().infiniteDistance()));
    }
    batchNN(getRoot(), candidates, knnLists);


    Collections.sort(result);
    // Collections.sort(candidates);


    rkNNStatistics.addCandidates(candidates.size());
    rkNNStatistics.addTrueHits(result.size());


    for(DBID cid : candidates) {
      for(DistanceResultPair<D> qr : knnLists.get(id)) {
        if(qr.getDBID().equals(id)) {

View Full Code Here

    // iterative phase
    if(logger.isVerbose()) {
      logger.verbose("2. Iterative phase...");
    }
    double bestObjective = Double.POSITIVE_INFINITY;
    ModifiableDBIDs m_best = null;
    ModifiableDBIDs m_bad = null;
    ModifiableDBIDs m_current = initialSet(medoids, k, random);


    if(logger.isDebugging()) {
      StringBuffer msg = new StringBuffer();
      msg.append("\n");
      msg.append("m_c ").append(m_current).append("\n");

View Full Code Here

   * @param random random number generator
   * @return a piercing set of m medoids from the specified sample set
   */
  private ModifiableDBIDs greedy(DistanceQuery<V, DoubleDistance> distFunc, DBIDs sampleSet, int m, Random random) {
    ArrayModifiableDBIDs s = DBIDUtil.newArray(sampleSet);
    ModifiableDBIDs medoids = DBIDUtil.newHashSet();


    // m_1 is random point of S
    DBID m_i = s.remove(random.nextInt(s.size()));
    medoids.add(m_i);
    if(logger.isDebugging()) {
      logger.debugFiner("medoids " + medoids);
    }


    // compute distances between each point in S and m_i
    Map<DBID, DistanceResultPair<DoubleDistance>> distances = new HashMap<DBID, DistanceResultPair<DoubleDistance>>();
    for(DBID id : s) {
      DoubleDistance dist = distFunc.distance(id, m_i);
      distances.put(id, new GenericDistanceResultPair<DoubleDistance>(dist, id));
    }


    for(int i = 1; i < m; i++) {
      // choose medoid m_i to be far from prevois medoids
      List<DistanceResultPair<DoubleDistance>> d = new ArrayList<DistanceResultPair<DoubleDistance>>(distances.values());
      Collections.sort(d);


      m_i = d.get(d.size() - 1).getDBID();
      medoids.add(m_i);
      s.remove(m_i);
      distances.remove(m_i);


      // compute distances of each point to closest medoid
      for(DBID id : s) {

View Full Code Here

   * @param random random number generator
   * @return a set of k elements from the specified sample set
   */
  private ModifiableDBIDs initialSet(DBIDs sampleSet, int k, Random random) {
    ArrayModifiableDBIDs s = DBIDUtil.newArray(sampleSet);
    ModifiableDBIDs initialSet = DBIDUtil.newHashSet();
    while(initialSet.size() < k) {
      DBID next = s.remove(random.nextInt(s.size()));
      initialSet.add(next);
    }
    return initialSet;
  }

View Full Code Here

    ArrayModifiableDBIDs m_list = DBIDUtil.newArray(m);
    for(DBID m_i : m_best) {
      m_list.remove(m_i);
    }


    ModifiableDBIDs m_current = DBIDUtil.newHashSet();
    for(DBID m_i : m_best) {
      if(m_bad.contains(m_i)) {
        int currentSize = m_current.size();
        while(m_current.size() == currentSize) {
          DBID next = m_list.remove(random.nextInt(m_list.size()));
          m_current.add(next);
        }
      }
      else {
        m_current.add(m_i);
      }
    }


    return m_current;
  }

View Full Code Here

          minDist = currentDist;
        }
      }
      // add p to cluster with mindist
      assert minDist != null;
      ModifiableDBIDs ids = clusterIDs.get(minDist.getDBID());
      ids.add(p_id);
    }


    Map<DBID, PROCLUSCluster> clusters = new HashMap<DBID, PROCLUSCluster>();
    for(DBID m_i : dimensions.keySet()) {
      ModifiableDBIDs objectIDs = clusterIDs.get(m_i);
      if(!objectIDs.isEmpty()) {
        Set<Integer> clusterDimensions = dimensions.get(m_i);
        V centroid = DatabaseUtil.centroid(database, objectIDs);
        clusters.put(m_i, new PROCLUSCluster(objectIDs, clusterDimensions, centroid));
      }
    }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of de.lmu.ifi.dbs.elki.database.ids.ModifiableDBIDs

de.lmu.ifi.dbs.elki.algorithm.clustering.AbstractProjectedDBSCAN

de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.CASH

de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.cash.CASHInterval

de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.cash.CASHIntervalSplit

de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.COPAC

de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.ERiC

de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.LMCLUS

de.lmu.ifi.dbs.elki.algorithm.clustering.DBSCAN

de.lmu.ifi.dbs.elki.algorithm.clustering.KMeans

de.lmu.ifi.dbs.elki.algorithm.clustering.kmeans.AbstractKMeans

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.