1 /*
2 * Licensed to the Apache Software Foundation (ASF) under one or more
3 * contributor license agreements. See the NOTICE file distributed with
4 * this work for additional information regarding copyright ownership.
5 * The ASF licenses this file to You under the Apache License, Version 2.0
6 * (the "License"); you may not use this file except in compliance with
7 * the License. You may obtain a copy of the License at
8 *
9 * https://www.apache.org/licenses/LICENSE-2.0
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 */
17
18 /*
19 * This is not the original file distributed by the Apache Software Foundation
20 * It has been modified by the Hipparchus project
21 */
22
23 package org.hipparchus.clustering.evaluation;
24
25 import java.util.List;
26
27 import org.hipparchus.clustering.CentroidCluster;
28 import org.hipparchus.clustering.Cluster;
29 import org.hipparchus.clustering.Clusterable;
30 import org.hipparchus.clustering.DoublePoint;
31 import org.hipparchus.clustering.distance.DistanceMeasure;
32 import org.hipparchus.clustering.distance.EuclideanDistance;
33
34 /**
35 * Base class for cluster evaluation methods.
36 *
37 * @param <T> type of the clustered points
38 */
39 public abstract class ClusterEvaluator<T extends Clusterable> {
40
41 /** The distance measure to use when evaluating the cluster. */
42 private final DistanceMeasure measure;
43
44 /**
45 * Creates a new cluster evaluator with an {@link EuclideanDistance}
46 * as distance measure.
47 */
48 public ClusterEvaluator() {
49 this(new EuclideanDistance());
50 }
51
52 /**
53 * Creates a new cluster evaluator with the given distance measure.
54 * @param measure the distance measure to use
55 */
56 public ClusterEvaluator(final DistanceMeasure measure) {
57 this.measure = measure;
58 }
59
60 /**
61 * Computes the evaluation score for the given list of clusters.
62 * @param clusters the clusters to evaluate
63 * @return the computed score
64 */
65 public abstract double score(List<? extends Cluster<T>> clusters);
66
67 /**
68 * Returns whether the first evaluation score is considered to be better
69 * than the second one by this evaluator.
70 * <p>
71 * Specific implementations shall override this method if the returned scores
72 * do not follow the same ordering, i.e. smaller score is better.
73 *
74 * @param score1 the first score
75 * @param score2 the second score
76 * @return {@code true} if the first score is considered to be better, {@code false} otherwise
77 */
78 public boolean isBetterScore(double score1, double score2) {
79 return score1 < score2;
80 }
81
82 /**
83 * Calculates the distance between two {@link Clusterable} instances
84 * with the configured {@link DistanceMeasure}.
85 *
86 * @param p1 the first clusterable
87 * @param p2 the second clusterable
88 * @return the distance between the two clusterables
89 */
90 protected double distance(final Clusterable p1, final Clusterable p2) {
91 return measure.compute(p1.getPoint(), p2.getPoint());
92 }
93
94 /**
95 * Computes the centroid for a cluster.
96 *
97 * @param cluster the cluster
98 * @return the computed centroid for the cluster,
99 * or {@code null} if the cluster does not contain any points
100 */
101 protected Clusterable centroidOf(final Cluster<T> cluster) {
102 final List<T> points = cluster.getPoints();
103 if (points.isEmpty()) {
104 return null;
105 }
106
107 // in case the cluster is of type CentroidCluster, no need to compute the centroid
108 if (cluster instanceof CentroidCluster) {
109 return ((CentroidCluster<T>) cluster).getCenter();
110 }
111
112 final int dimension = points.get(0).getPoint().length;
113 final double[] centroid = new double[dimension];
114 for (final T p : points) {
115 final double[] point = p.getPoint();
116 for (int i = 0; i < centroid.length; i++) {
117 centroid[i] += point[i];
118 }
119 }
120 for (int i = 0; i < centroid.length; i++) {
121 centroid[i] /= points.size();
122 }
123 return new DoublePoint(centroid);
124 }
125
126 }