Package com.aamend.hadoop.clustering.cluster

Examples of com.aamend.hadoop.clustering.cluster.Cluster

42
43
44
45
46
47
48
49
50
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
79
    }

    @Override
    protected void map(Text key, CanopyWritable value, Context context) throws IOException, InterruptedException {

        Cluster cluster = value.get();
        int[] point = cluster.getCenter();
        boolean stronglyBound = false;
        for (Cluster canopy : canopies) {
            double dist = measure.distance(canopy.getCenter(), point);
            if (dist < t1) {
                KEY.set(Arrays.toString(canopy.getCenter()));
                Cluster newCluster;
                if (dist < t2) {
                    newCluster = new Canopy(canopy.getId(), point, value.get().getNum());
                    LOGGER.debug("Adding (T2) {} to Cluster {}", Arrays.toString(point),
                            Arrays.toString(canopy.getCenter()));
                } else {
                    newCluster = new Canopy(canopy.getId(), point, 0L);
                    LOGGER.debug("Adding (T1) {} to Cluster {}", Arrays.toString(point),
                            Arrays.toString(canopy.getCenter()));
                }

                context.write(KEY, new CanopyWritable(newCluster));
            }

            stronglyBound = stronglyBound || dist < t2;
        }
        if (!stronglyBound) {
            nextCanopyId++;
            Cluster canopy = new Canopy(nextCanopyId, point, value.get().getNum());
            canopies.add(canopy);
            LOGGER.debug("Creating a new Cluster {}", canopy.asFormattedString());
            KEY.set(Arrays.toString(canopy.getCenter()));
            context.write(KEY, value);
        }

    }
View Full Code Here
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
79
80
        boolean stronglyBound = false;
        for (Cluster canopy : canopies) {
            double dist = measure.distance(canopy.getCenter(), point);
            if (dist < t1) {
                KEY.set(Arrays.toString(canopy.getCenter()));
                Cluster newCluster;
                if (dist < t2) {
                    newCluster = new Canopy(canopy.getId(), point, 1L);
                    LOGGER.debug("Adding (T2) {} to Cluster center {}", Arrays.toString((int[]) value.get()),
                            Arrays.toString(canopy.getCenter()));
                } else {
                    newCluster = new Canopy(canopy.getId(), point, 0L);
                    LOGGER.debug("Adding (T1) {} to Cluster center {}", Arrays.toString((int[]) value.get()),
                            Arrays.toString(canopy.getCenter()));
                }
                context.write(KEY, new CanopyWritable(newCluster));
            }
            stronglyBound = stronglyBound || dist < t2;
        }
        if (!stronglyBound) {
            nextCanopyId++;
            Cluster canopy = new Canopy(nextCanopyId, point, 1L);
            LOGGER.debug("Creating a new Cluster {}", canopy.asFormattedString());
            canopies.add(canopy);
            KEY.set(Arrays.toString(canopy.getCenter()));
            context.write(KEY, new CanopyWritable(canopy));
        }

    }
View Full Code Here
46
47
48
49
50
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
79
80
        // Try to find a center that could minimize all data points
        List<int[]> points = new ArrayList<int[]>();

        long obs = 0L;
        Cluster clusterTemplate = null;
        for (CanopyWritable value : values) {
            if (clusterTemplate == null) {
                clusterTemplate = value.get();
            } else {
                obs += value.get().getNum();
                points.add(value.get().getCenter());
            }
        }

        // Increment number of observations for this cluster
        clusterTemplate.observe(obs);

        if (lastIteration) {
            if (clusterTemplate.getNum() < minObservations) {
                context.getCounter(COUNTER, COUNTER_REJECTED_CANOPY).increment(1L);
                return;
            }
        }

        LOGGER.info("Minimizing distance across {} data points in cluster center {}",
                points.size(), Arrays.toString(clusterTemplate.getCenter()));

        clusterTemplate.computeCenter(points, measure);
        nextCanopyId++;
        Cluster newCluster = new Canopy(nextCanopyId, clusterTemplate.getCenter(), clusterTemplate.getNum());
        context.getCounter(COUNTER, COUNTER_CANOPY).increment(1L);
        context.write(KEY, new CanopyWritable(newCluster));

    }
View Full Code Here
61
62
63
64
65
66
67
68
69
70
71
                                .newInstance(reader.getValueClass(), conf);

                int i = 0;
                while (reader.next(key, value)) {
                    i++;
                    Cluster cluster = value.get();
                    clusters.add(cluster);
                }

                IOUtils.closeStream(reader);
            }
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
            throws IOException, InterruptedException {

        // Get distance from that point to any cluster center
        double[] pdf = new double[clusters.size()];
        for (int i = 0; i < clusters.size(); i++) {
            Cluster cluster = clusters.get(i);
            pdf[i] = cluster.pdf((int[]) value.get(), measure);
        }

        // Get the cluster with smallest distance to that point
        double maxSimilarity = pdf[0];
        int maxSimilarityId = 0;
        for (int i = 1; i < pdf.length; i++) {
            if (pdf[i] > maxSimilarity) {
                maxSimilarity = pdf[i];
                maxSimilarityId = i;
            }
        }

        if (maxSimilarity < minSimilarity) {
            // Point could not be added to any cluster
            context.getCounter(COUNTER, COUNTER_NON_CLUSTERED).increment(1L);
            return;
        }

        // Point has been added to that cluster
        context.getCounter(COUNTER, COUNTER_CLUSTERED).increment(1L);
        Cluster cluster = clusters.get(maxSimilarityId);

        KEY.set(cluster.getId());
        context.write(KEY, new ObjectWritable(key));

    }
View Full Code Here
TOP

Related Classes of com.aamend.hadoop.clustering.cluster.Cluster

  • com.aamend.hadoop.clustering.mapreduce.CanopyCreateInitMapper
  • com.aamend.hadoop.clustering.mapreduce.CanopyCreateMapper
  • com.aamend.hadoop.clustering.mapreduce.CanopyCreateReducer
  • com.aamend.hadoop.clustering.mapreduce.ClusterDataMapper
    • 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.