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
|
package de.lmu.ifi.dbs.elki.distance.distancefunction.subspace;
/*
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.BitSet;
import de.lmu.ifi.dbs.elki.data.NumberVector;
import de.lmu.ifi.dbs.elki.database.relation.Relation;
import de.lmu.ifi.dbs.elki.database.relation.RelationUtil;
import de.lmu.ifi.dbs.elki.distance.distancevalue.PreferenceVectorBasedCorrelationDistance;
import de.lmu.ifi.dbs.elki.index.preprocessed.preference.HiSCPreferenceVectorIndex;
import de.lmu.ifi.dbs.elki.logging.Logging;
import de.lmu.ifi.dbs.elki.utilities.ClassGenericsUtil;
import de.lmu.ifi.dbs.elki.utilities.FormatUtil;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.parameterization.Parameterization;
/**
* Distance function used in the HiSC algorithm.
*
* @author Elke Achtert
*
* @apiviz.has Instance
*
* @param <V> the type of NumberVector to compute the distances in between
*/
public class HiSCDistanceFunction<V extends NumberVector<?>> extends AbstractPreferenceVectorBasedCorrelationDistanceFunction<V, HiSCPreferenceVectorIndex<V>> {
/**
* Logger for debug.
*/
private static final Logging LOG = Logging.getLogger(HiSCDistanceFunction.class);
/**
* Constructor.
*
* @param indexFactory HiSC index factory
* @param epsilon Epsilon value
*/
public HiSCDistanceFunction(HiSCPreferenceVectorIndex.Factory<V> indexFactory, double epsilon) {
super(indexFactory, epsilon);
}
@Override
public <T extends V> Instance<T> instantiate(Relation<T> database) {
// We can't really avoid these warnings, due to a limitation in Java
// Generics (AFAICT)
@SuppressWarnings("unchecked")
HiSCPreferenceVectorIndex<T> indexinst = (HiSCPreferenceVectorIndex<T>) indexFactory.instantiate((Relation<V>) database);
return new Instance<T>(database, indexinst, getEpsilon(), this);
}
/**
* The actual instance bound to a particular database.
*
* @author Erich Schubert
*
* @param <V> the type of NumberVector to compute the distances in between
*/
public static class Instance<V extends NumberVector<?>> extends AbstractPreferenceVectorBasedCorrelationDistanceFunction.Instance<V, HiSCPreferenceVectorIndex<V>> {
/**
* Constructor.
*
* @param database Database
* @param index Preprocessed index
* @param epsilon Epsilon
* @param distanceFunction parent distance function
*/
public Instance(Relation<V> database, HiSCPreferenceVectorIndex<V> index, double epsilon, HiSCDistanceFunction<? super V> distanceFunction) {
super(database, index, epsilon, distanceFunction);
}
/**
* Computes the correlation distance between the two specified vectors
* according to the specified preference vectors.
*
* @param v1 first vector
* @param v2 second vector
* @param pv1 the first preference vector
* @param pv2 the second preference vector
* @return the correlation distance between the two specified vectors
*/
@Override
public PreferenceVectorBasedCorrelationDistance correlationDistance(V v1, V v2, BitSet pv1, BitSet pv2) {
BitSet commonPreferenceVector = (BitSet) pv1.clone();
commonPreferenceVector.and(pv2);
int dim = v1.getDimensionality();
// number of zero values in commonPreferenceVector
Integer subspaceDim = dim - commonPreferenceVector.cardinality();
// special case: v1 and v2 are in parallel subspaces
double dist1 = weightedDistance(v1, v2, pv1);
double dist2 = weightedDistance(v1, v2, pv2);
if(Math.max(dist1, dist2) > epsilon) {
subspaceDim++;
if(LOG.isDebugging()) {
//Representation<String> rep = rep.getObjectLabelQuery();
StringBuilder msg = new StringBuilder();
msg.append("\ndist1 ").append(dist1);
msg.append("\ndist2 ").append(dist2);
// msg.append("\nv1 ").append(rep.get(v1.getID()));
// msg.append("\nv2 ").append(rep.get(v2.getID()));
msg.append("\nsubspaceDim ").append(subspaceDim);
msg.append("\ncommon pv ").append(FormatUtil.format(dim, commonPreferenceVector));
LOG.debugFine(msg.toString());
}
}
// flip commonPreferenceVector for distance computation in common subspace
BitSet inverseCommonPreferenceVector = (BitSet) commonPreferenceVector.clone();
inverseCommonPreferenceVector.flip(0, dim);
return new PreferenceVectorBasedCorrelationDistance(RelationUtil.dimensionality(relation), subspaceDim, weightedDistance(v1, v2, inverseCommonPreferenceVector), commonPreferenceVector);
}
}
/**
* Parameterization class.
*
* @author Erich Schubert
*
* @apiviz.exclude
*/
public static class Parameterizer<V extends NumberVector<?>> extends AbstractPreferenceVectorBasedCorrelationDistanceFunction.Parameterizer<HiSCPreferenceVectorIndex.Factory<V>> {
@Override
protected void makeOptions(Parameterization config) {
super.makeOptions(config);
Class<HiSCPreferenceVectorIndex.Factory<V>> cls = ClassGenericsUtil.uglyCastIntoSubclass(HiSCPreferenceVectorIndex.Factory.class);
factory = config.tryInstantiate(cls);
}
@Override
protected HiSCDistanceFunction<V> makeInstance() {
return new HiSCDistanceFunction<V>(factory, epsilon);
}
}
}
|
