1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
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
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
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
|
package de.lmu.ifi.dbs.elki.algorithm.statistics;
/*
This file is part of ELKI:
Environment for Developing KDD-Applications Supported by Index-Structures
Copyright (C) 2012
Ludwig-Maximilians-Universität München
Lehr- und Forschungseinheit für Datenbanksysteme
ELKI Development Team
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
import java.util.ArrayList;
import java.util.Collection;
import de.lmu.ifi.dbs.elki.algorithm.AbstractDistanceBasedAlgorithm;
import de.lmu.ifi.dbs.elki.algorithm.clustering.trivial.ByLabelOrAllInOneClustering;
import de.lmu.ifi.dbs.elki.data.Cluster;
import de.lmu.ifi.dbs.elki.data.DoubleVector;
import de.lmu.ifi.dbs.elki.data.model.Model;
import de.lmu.ifi.dbs.elki.data.type.TypeInformation;
import de.lmu.ifi.dbs.elki.data.type.TypeUtil;
import de.lmu.ifi.dbs.elki.database.Database;
import de.lmu.ifi.dbs.elki.database.ids.DBIDIter;
import de.lmu.ifi.dbs.elki.database.query.distance.DistanceQuery;
import de.lmu.ifi.dbs.elki.database.query.knn.KNNQuery;
import de.lmu.ifi.dbs.elki.database.relation.Relation;
import de.lmu.ifi.dbs.elki.distance.distancefunction.DistanceFunction;
import de.lmu.ifi.dbs.elki.distance.distanceresultlist.KNNResult;
import de.lmu.ifi.dbs.elki.distance.distancevalue.NumberDistance;
import de.lmu.ifi.dbs.elki.evaluation.roc.ROC;
import de.lmu.ifi.dbs.elki.logging.Logging;
import de.lmu.ifi.dbs.elki.logging.progress.FiniteProgress;
import de.lmu.ifi.dbs.elki.math.MeanVariance;
import de.lmu.ifi.dbs.elki.result.CollectionResult;
import de.lmu.ifi.dbs.elki.result.HistogramResult;
import de.lmu.ifi.dbs.elki.utilities.datastructures.histogram.DoubleStaticHistogram;
import de.lmu.ifi.dbs.elki.utilities.documentation.Description;
import de.lmu.ifi.dbs.elki.utilities.documentation.Title;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.OptionID;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.constraints.GreaterEqualConstraint;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.parameterization.Parameterization;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.parameters.IntParameter;
/**
* Evaluate a distance function with respect to kNN queries. For each point, the
* neighbors are sorted by distance, then the ROC AUC is computed. A score of 1
* means that the distance function provides a perfect ordering of relevant
* neighbors first, then irrelevant neighbors. A value of 0.5 can be obtained by
* random sorting. A value of 0 means the distance function is inverted, i.e. a
* similarity.
*
* TODO: Add sampling
*
* @author Erich Schubert
* @param <O> Object type
* @param <D> Distance type
*/
@Title("Ranking Quality Histogram")
@Description("Evaluates the effectiveness of a distance function via the obtained rankings.")
public class RankingQualityHistogram<O, D extends NumberDistance<D, ?>> extends AbstractDistanceBasedAlgorithm<O, D, CollectionResult<DoubleVector>> {
/**
* The logger for this class.
*/
private static final Logging LOG = Logging.getLogger(RankingQualityHistogram.class);
/**
* Option to configure the number of bins to use.
*/
public static final OptionID HISTOGRAM_BINS_ID = new OptionID("rankqual.bins", "Number of bins to use in the histogram");
/**
* Number of bins to use.
*/
int numbins = 100;
/**
* Constructor.
*
* @param distanceFunction Distance function to evaluate
* @param numbins Number of bins
*/
public RankingQualityHistogram(DistanceFunction<? super O, D> distanceFunction, int numbins) {
super(distanceFunction);
this.numbins = numbins;
}
/**
* Process a database
*
* @param database Database to process
* @param relation Relation to process
* @return Histogram of ranking qualities
*/
public HistogramResult<DoubleVector> run(Database database, Relation<O> relation) {
final DistanceQuery<O, D> distanceQuery = database.getDistanceQuery(relation, getDistanceFunction());
final KNNQuery<O, D> knnQuery = database.getKNNQuery(distanceQuery, relation.size());
if (LOG.isVerbose()) {
LOG.verbose("Preprocessing clusters...");
}
// Cluster by labels
Collection<Cluster<Model>> split = (new ByLabelOrAllInOneClustering()).run(database).getAllClusters();
DoubleStaticHistogram hist = new DoubleStaticHistogram(numbins, 0.0, 1.0);
if (LOG.isVerbose()) {
LOG.verbose("Processing points...");
}
FiniteProgress progress = LOG.isVerbose() ? new FiniteProgress("Computing ROC AUC values", relation.size(), LOG) : null;
MeanVariance mv = new MeanVariance();
// sort neighbors
for (Cluster<?> clus : split) {
for (DBIDIter iter = clus.getIDs().iter(); iter.valid(); iter.advance()) {
KNNResult<D> knn = knnQuery.getKNNForDBID(iter, relation.size());
double result = ROC.computeROCAUCDistanceResult(relation.size(), clus, knn);
mv.put(result);
hist.increment(result, 1. / relation.size());
if (progress != null) {
progress.incrementProcessed(LOG);
}
}
}
if (progress != null) {
progress.ensureCompleted(LOG);
}
// Transform Histogram into a Double Vector array.
Collection<DoubleVector> res = new ArrayList<DoubleVector>(relation.size());
for (DoubleStaticHistogram.Iter iter = hist.iter(); iter.valid(); iter.advance()) {
DoubleVector row = new DoubleVector(new double[] { iter.getCenter(), iter.getValue() });
res.add(row);
}
HistogramResult<DoubleVector> result = new HistogramResult<DoubleVector>("Ranking Quality Histogram", "ranking-histogram", res);
result.addHeader("Mean: " + mv.getMean() + " Variance: " + mv.getSampleVariance());
return result;
}
@Override
public TypeInformation[] getInputTypeRestriction() {
return TypeUtil.array(getDistanceFunction().getInputTypeRestriction());
}
@Override
protected Logging getLogger() {
return LOG;
}
/**
* Parameterization class.
*
* @author Erich Schubert
*
* @apiviz.exclude
*/
public static class Parameterizer<O, D extends NumberDistance<D, ?>> extends AbstractDistanceBasedAlgorithm.Parameterizer<O, D> {
/**
* Number of bins.
*/
protected int numbins = 20;
@Override
protected void makeOptions(Parameterization config) {
super.makeOptions(config);
final IntParameter param = new IntParameter(HISTOGRAM_BINS_ID, 100);
param.addConstraint(new GreaterEqualConstraint(2));
if (config.grab(param)) {
numbins = param.getValue();
}
}
@Override
protected RankingQualityHistogram<O, D> makeInstance() {
return new RankingQualityHistogram<O, D>(distanceFunction, numbins);
}
}
}
|
