Package org.apache.mahout.math.neighborhood

Examples of org.apache.mahout.math.neighborhood.UpdatableSearcher

47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
    this.conf = conf;
  }

  @Override
  public Iterable<Centroid> call() {
    UpdatableSearcher searcher = StreamingKMeansUtilsMR.searcherFromConfiguration(conf);
    int numClusters = conf.getInt(StreamingKMeansDriver.ESTIMATED_NUM_MAP_CLUSTERS, 1);
    double estimateDistanceCutoff = conf.getFloat(StreamingKMeansDriver.ESTIMATED_DISTANCE_CUTOFF,
        StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF);

    Iterator<Centroid> datapointsIterator = datapoints.iterator();
    if (estimateDistanceCutoff == StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF) {
      List<Centroid> estimatePoints = Lists.newArrayListWithExpectedSize(NUM_ESTIMATE_POINTS);
      while (datapointsIterator.hasNext() && estimatePoints.size() < NUM_ESTIMATE_POINTS) {
        estimatePoints.add(datapointsIterator.next());
      }
      estimateDistanceCutoff = ClusteringUtils.estimateDistanceCutoff(estimatePoints, searcher.getDistanceMeasure());
    }

    StreamingKMeans clusterer = new StreamingKMeans(searcher, numClusters, estimateDistanceCutoff);
    while (datapointsIterator.hasNext()) {
      clusterer.cluster(datapointsIterator.next());
View Full Code Here
51
52
53
54
55
56
57
58
59
60
61
  @Override
  public void setup(Context context) {
    // At this point the configuration received from the Driver is assumed to be valid.
    // No other checks are made.
    Configuration conf = context.getConfiguration();
    UpdatableSearcher searcher = StreamingKMeansUtilsMR.searcherFromConfiguration(conf);
    int numClusters = conf.getInt(StreamingKMeansDriver.ESTIMATED_NUM_MAP_CLUSTERS, 1);
    double estimatedDistanceCutoff = conf.getFloat(StreamingKMeansDriver.ESTIMATED_DISTANCE_CUTOFF,
        StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF);
    if (estimatedDistanceCutoff == StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF) {
      estimateDistanceCutoff = true;
View Full Code Here
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
   * index is i.
   */
  public static List<OnlineSummarizer> summarizeClusterDistances(Iterable<? extends Vector> datapoints,
                                                                 Iterable<? extends Vector> centroids,
                                                                 DistanceMeasure distanceMeasure) {
    UpdatableSearcher searcher = new ProjectionSearch(distanceMeasure, 3, 1);
    searcher.addAll(centroids);
    List<OnlineSummarizer> summarizers = Lists.newArrayList();
    if (searcher.size() == 0) {
      return summarizers;
    }
    for (int i = 0; i < searcher.size(); ++i) {
      summarizers.add(new OnlineSummarizer());
    }
    for (Vector v : datapoints) {
      Centroid closest = (Centroid)searcher.search(v,  1).get(0).getValue();
      OnlineSummarizer summarizer = summarizers.get(closest.getIndex());
      summarizer.add(distanceMeasure.distance(v, closest));
    }
    return summarizers;
  }
View Full Code Here
56
57
58
59
60
61
62
63
64
   * @param centroids iterable of Centroids.
   * @return the total cost described above.
   */
  public static double totalClusterCost(Iterable<? extends Vector> datapoints, Iterable<? extends Vector> centroids) {
    DistanceMeasure distanceMeasure = new EuclideanDistanceMeasure();
    UpdatableSearcher searcher = new ProjectionSearch(distanceMeasure, 3, 1);
    searcher.addAll(centroids);
    return totalClusterCost(datapoints, searcher);
  }
View Full Code Here
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
    }
  }

  @Test
  public void testClustering() {
    UpdatableSearcher searcher = new BruteSearch(new SquaredEuclideanDistanceMeasure());
    BallKMeans clusterer = new BallKMeans(searcher, 1 << NUM_DIMENSIONS, NUM_ITERATIONS);

    long startTime = System.currentTimeMillis();
    clusterer.cluster(syntheticData.getFirst());
    long endTime = System.currentTimeMillis();

    assertEquals("Total weight not preserved", totalWeight(syntheticData.getFirst()), totalWeight(clusterer), 1.0e-9);

    // Verify that each corner of the cube has a centroid very nearby.
    // This is probably FALSE for large-dimensional spaces!
    OnlineSummarizer summarizer = new OnlineSummarizer();
    for (Vector mean : syntheticData.getSecond()) {
      WeightedThing<Vector> v = searcher.search(mean, 1).get(0);
      summarizer.add(v.getWeight());
    }
    assertTrue(String.format("Median weight [%f] too large [>%f]", summarizer.getMedian(),
        DISTRIBUTION_RADIUS), summarizer.getMedian() < DISTRIBUTION_RADIUS);

    double clusterTime = (endTime - startTime) / 1000.0;
    System.out.printf("%s\n%.2f for clustering\n%.1f us per row\n\n",
        searcher.getClass().getName(), clusterTime,
        clusterTime / syntheticData.getFirst().size() * 1.0e6);

    // Verify that the total weight of the centroids near each corner is correct.
    double[] cornerWeights = new double[1 << NUM_DIMENSIONS];
    Searcher trueFinder = new BruteSearch(new EuclideanDistanceMeasure());
View Full Code Here
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
    this.conf = conf;
  }

  @Override
  public Iterable<Centroid> call() {
    UpdatableSearcher searcher = StreamingKMeansUtilsMR.searcherFromConfiguration(conf);
    int numClusters = conf.getInt(StreamingKMeansDriver.ESTIMATED_NUM_MAP_CLUSTERS, 1);
    double estimateDistanceCutoff = conf.getFloat(StreamingKMeansDriver.ESTIMATED_DISTANCE_CUTOFF,
        StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF);

    Iterator<Centroid> dataPointsIterator = dataPoints.iterator();

    if (estimateDistanceCutoff == StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF) {
      List<Centroid> estimatePoints = Lists.newArrayListWithExpectedSize(NUM_ESTIMATE_POINTS);
      while (dataPointsIterator.hasNext() && estimatePoints.size() < NUM_ESTIMATE_POINTS) {
        Centroid centroid = dataPointsIterator.next();
        estimatePoints.add(centroid);
      }

      if (log.isInfoEnabled()) {
        log.info("Estimated Points: {}", estimatePoints.size());
      }
      estimateDistanceCutoff = ClusteringUtils.estimateDistanceCutoff(estimatePoints, searcher.getDistanceMeasure());
    }

    StreamingKMeans streamingKMeans = new StreamingKMeans(searcher, numClusters, estimateDistanceCutoff);

    // datapointsIterator could be empty if no estimate distance was initially provided
View Full Code Here
51
52
53
54
55
56
57
58
59
60
61
  @Override
  public void setup(Context context) {
    // At this point the configuration received from the Driver is assumed to be valid.
    // No other checks are made.
    Configuration conf = context.getConfiguration();
    UpdatableSearcher searcher = StreamingKMeansUtilsMR.searcherFromConfiguration(conf);
    int numClusters = conf.getInt(StreamingKMeansDriver.ESTIMATED_NUM_MAP_CLUSTERS, 1);
    double estimatedDistanceCutoff = conf.getFloat(StreamingKMeansDriver.ESTIMATED_DISTANCE_CUTOFF,
        StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF);
    if (estimatedDistanceCutoff == StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF) {
      estimateDistanceCutoff = true;
View Full Code Here
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
   * index is i.
   */
  public static List<OnlineSummarizer> summarizeClusterDistances(Iterable<? extends Vector> datapoints,
                                                                 Iterable<? extends Vector> centroids,
                                                                 DistanceMeasure distanceMeasure) {
    UpdatableSearcher searcher = new ProjectionSearch(distanceMeasure, 3, 1);
    searcher.addAll(centroids);
    List<OnlineSummarizer> summarizers = Lists.newArrayList();
    if (searcher.size() == 0) {
      return summarizers;
    }
    for (int i = 0; i < searcher.size(); ++i) {
      summarizers.add(new OnlineSummarizer());
    }
    for (Vector v : datapoints) {
      Centroid closest = (Centroid)searcher.search(v,  1).get(0).getValue();
      OnlineSummarizer summarizer = summarizers.get(closest.getIndex());
      summarizer.add(distanceMeasure.distance(v, closest));
    }
    return summarizers;
  }
View Full Code Here
73
74
75
76
77
78
79
80
81
   * @param centroids iterable of Centroids.
   * @return the total cost described above.
   */
  public static double totalClusterCost(Iterable<? extends Vector> datapoints, Iterable<? extends Vector> centroids) {
    DistanceMeasure distanceMeasure = new EuclideanDistanceMeasure();
    UpdatableSearcher searcher = new ProjectionSearch(distanceMeasure, 3, 1);
    searcher.addAll(centroids);
    return totalClusterCost(datapoints, searcher);
  }
View Full Code Here
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
    }
  }

  @Test
  public void testClustering() {
    UpdatableSearcher searcher = new BruteSearch(new SquaredEuclideanDistanceMeasure());
    BallKMeans clusterer = new BallKMeans(searcher, 1 << NUM_DIMENSIONS, NUM_ITERATIONS);

    long startTime = System.currentTimeMillis();
    Pair<List<Centroid>, List<Centroid>> data = syntheticData;
    clusterer.cluster(data.getFirst());
    long endTime = System.currentTimeMillis();

    long hash = 0;
    for (Centroid centroid : data.getFirst()) {
      for (Vector.Element element : centroid.all()) {
        hash = 31 * hash + 17 * element.index() + Double.toHexString(element.get()).hashCode();
      }
    }
    System.out.printf("Hash = %08x\n", hash);

    assertEquals("Total weight not preserved", totalWeight(syntheticData.getFirst()), totalWeight(clusterer), 1.0e-9);

    // Verify that each corner of the cube has a centroid very nearby.
    // This is probably FALSE for large-dimensional spaces!
    OnlineSummarizer summarizer = new OnlineSummarizer();
    for (Vector mean : syntheticData.getSecond()) {
      WeightedThing<Vector> v = searcher.search(mean, 1).get(0);
      summarizer.add(v.getWeight());
    }
    assertTrue(String.format("Median weight [%f] too large [>%f]", summarizer.getMedian(),
        DISTRIBUTION_RADIUS), summarizer.getMedian() < DISTRIBUTION_RADIUS);

    double clusterTime = (endTime - startTime) / 1000.0;
    System.out.printf("%s\n%.2f for clustering\n%.1f us per row\n\n",
        searcher.getClass().getName(), clusterTime,
        clusterTime / syntheticData.getFirst().size() * 1.0e6);

    // Verify that the total weight of the centroids near each corner is correct.
    double[] cornerWeights = new double[1 << NUM_DIMENSIONS];
    Searcher trueFinder = new BruteSearch(new EuclideanDistanceMeasure());
View Full Code Here
TOP

Related Classes of org.apache.mahout.math.neighborhood.UpdatableSearcher

  • org.apache.mahout.clustering.ClusteringUtils
  • org.apache.mahout.clustering.streaming.cluster.BallKMeansTest
  • org.apache.mahout.clustering.streaming.mapreduce.StreamingKMeansMapper
  • org.apache.mahout.clustering.streaming.mapreduce.StreamingKMeansThread
    • Copyright © 2018 www.massapicom. All rights reserved.
    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.