Beispiel #1
0
  public Graph createGraph(Clustering[] clusts) {
    Clustering c = clusts[0];
    // total number of items
    int n = c.instancesCount();

    Graph graph = new AdjListGraph();
    Object2LongOpenHashMap<Instance> mapping = new Object2LongOpenHashMap();

    Instance a, b;
    Node na, nb;
    // cluster membership
    int ca, cb;
    int x = 0;
    Edge edge;
    // accumulate evidence
    for (Clustering clust : clusts) {
      System.out.println("reducing " + (x++));
      for (int i = 1; i < n; i++) {
        a = clust.instance(i);
        na = fetchNode(graph, mapping, a);
        ca = clust.assignedCluster(a.getIndex());
        for (int j = 0; j < i; j++) {
          b = clust.instance(j);
          nb = fetchNode(graph, mapping, b);
          // for each pair of instances check if placed in the same cluster
          cb = clust.assignedCluster(b.getIndex());
          if (ca == cb) {
            edge = graph.getEdge(na, nb);
            // check if exists
            if (edge == null) {
              edge = graph.getFactory().newEdge(na, nb, 0, 0, false);
              graph.addEdge(edge);
            }
            // increase weight by 1
            edge.setWeight(edge.getWeight() + 1.0);
          }
        }
      }
    }
    return graph;
  }
Beispiel #2
0
  @Override
  public String runMetis(Graph graph, int k, Props params) {
    long current = System.currentTimeMillis();
    File file = new File("inputGraph-" + current);
    String filename = file.getName();
    Process p;
    try {
      graph.hMetisExport(file, false);
      if (metisFile == null) {
        // fetch the file just once
        metisFile = getBinary();
        System.out.println("metis path: " + metisFile.getAbsolutePath());
      }

      // run metis
      String space = " ";
      StringBuilder sb = new StringBuilder(metisFile.getAbsolutePath());
      sb.append(" ")
          .append(filename)
          .append(" ")
          .append(String.valueOf(k))
          .append(space)
          .append("-ufactor=")
          .append(String.valueOf(params.getDouble(UFACTOR, 5.0)))
          .append(space)
          .append("-nruns=")
          .append(String.valueOf(params.getInt(NRUNS, 10)))
          .append(space)
          .append("-ptype=")
          .append(String.valueOf(params.get(PTYPE, "rb")))
          .append(space)
          .append("-otype=")
          .append(String.valueOf(params.get(OTYPE, "cut")))
          .append(space);
      if (params.containsKey(CTYPE)) {
        sb.append("-ctype=").append(params.get(CTYPE)).append(space);
      }
      if (params.containsKey(RTYPE)) {
        sb.append("-rtype=").append(params.get(RTYPE)).append(space);
      }

      // sb.append(String.valueOf(vcycle)).append(space);
      p = Runtime.getRuntime().exec(sb.toString());
      p.waitFor();
      if (debug) {
        System.out.println("cmd: " + sb.toString());
        System.out.println("exit code: " + p.exitValue());
        if (p.exitValue() != 1) {
          // System.out.println(ExtBinHelper.readFile(file));
        }

        helper.readStdout(p);
        helper.readStderr(p);
      }
      file.delete();
    } catch (FileNotFoundException | UnsupportedEncodingException ex) {
      Exceptions.printStackTrace(ex);
    } catch (IOException | InterruptedException ex) {
      Exceptions.printStackTrace(ex);
    }
    return filename;
  }
Beispiel #3
0
 private Node createNode(Graph g, Object2LongOpenHashMap<Instance> mapping, Instance inst) {
   Node node = g.getFactory().newNode(inst);
   mapping.put(inst, node.getId());
   g.addNode(node);
   return node;
 }
Beispiel #4
0
 private Node fetchNode(Graph g, Object2LongOpenHashMap<Instance> mapping, Instance inst) {
   if (mapping.containsKey(inst)) {
     return g.getNode(mapping.getLong(inst));
   }
   return createNode(g, mapping, inst);
 }
Beispiel #5
0
  @Override
  public Clustering<E, C> reduce(
      Clustering[] clusts, Algorithm<E, C> alg, ColorGenerator cg, Props props) {
    Graph graph = createGraph(clusts);

    // degree of freedom
    double df;
    double w, attain;
    EdgeIterable neigh;
    PriorityQueue<DoubleElem> pq = new PriorityQueue<>(graph.getNodeCount());
    DoubleElem<Node> elem;
    // for each node compute attainment score
    for (Node node : graph.getNodes()) {
      neigh = graph.getEdges(node);
      df = neigh.size();
      w = 0.0;
      for (Edge ne : neigh) {
        w += ne.getWeight();
      }
      attain = w / df;
      elem = new DoubleElem<>(node, attain);
      pq.add(elem);
    }

    // number of clusters is just a hint
    int k = props.getInt(KMeans.K, 5);
    double relax = props.getDouble(RELAX, 0.5);
    Clustering<E, C> result = new ClusterList(k);
    Dataset<? extends Instance> dataset = clusts[0].getLookup().lookup(Dataset.class);
    result.lookupAdd(dataset);
    ObjectOpenHashSet<Node> blacklist = new ObjectOpenHashSet();
    Node node, other;
    Cluster curr;
    double maxW;
    while (!pq.isEmpty()) {
      elem = pq.poll();
      node = elem.getElem();
      if (!blacklist.contains(node)) {
        blacklist.add(node);
        curr = result.createCluster();
        if (cg != null) {
          curr.setColor(cg.next());
        }
        curr.add(node.getInstance());

        EdgeIterable iter = graph.getEdges(node);
        maxW = -1;
        for (Edge ne : iter) {
          if (ne.getWeight() > maxW) {
            maxW = ne.getWeight();
          }
        }
        // add immediate neighbours with max weight to same cluster
        if (maxW >= 0.0) {
          for (Edge ne : iter) {
            // when relax set to 0.0, only items with maximum weight
            // will be added to the same cluster
            w = ne.getWeight() + relax * ne.getWeight();
            if (w >= maxW) {
              if (!node.equals(ne.getSource())) {
                other = ne.getSource();
              } else {
                other = ne.getTarget();
              }
              if (!blacklist.contains(other)) {
                curr.add(other.getInstance());
                blacklist.add(other);
              }
            }
          }
        }
      }
    }
    // TODO merge some clusters

    return result;
  }