1 /**
2 * Copyright (c) 2011, The University of Southampton and the individual contributors.
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without modification,
6 * are permitted provided that the following conditions are met:
7 *
8 * * Redistributions of source code must retain the above copyright notice,
9 * this list of conditions and the following disclaimer.
10 *
11 * * Redistributions in binary form must reproduce the above copyright notice,
12 * this list of conditions and the following disclaimer in the documentation
13 * and/or other materials provided with the distribution.
14 *
15 * * Neither the name of the University of Southampton nor the names of its
16 * contributors may be used to endorse or promote products derived from this
17 * software without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
21 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
23 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
24 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
25 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
26 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
28 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
29 */
30 package org.openimaj.pgm.vb.lda.mle;
31
32 import org.openimaj.pgm.util.Corpus;
33
34 /**
35 * Holds the sufficient statistics for a maximum liklihood LDA
36 * as well as a single value for alpha (the parameter of the topic
37 * dirichlet prior)
38 * @author Sina Samangooei (ss@ecs.soton.ac.uk)
39 *
40 */
41 public class LDAModel{
42 /**
43 * the dirichelet perameter for every dimension of the topic multinomial prior
44 */
45 public double alpha;
46 /**
47 * The maximum likelihood sufficient statistics for estimation of Beta.
48 * This is the number of times a given word is in a given topic
49 */
50 public double[][] topicWord;
51 /**
52 * The maximum likelihood sufficient statistics for estimation of Beta.
53 * This is the number of words total in a given topic
54 */
55 public double[] topicTotal;
56 /**
57 * number of topics
58 */
59 public int ntopics;
60 int iteration;
61 double likelihood,oldLikelihood;
62
63 /**
64 * @param ntopics the number of topics in this LDA model
65 */
66 public LDAModel(int ntopics) {
67 this.ntopics = ntopics;
68 }
69
70 /**
71 * initialises the sufficient statistic holder based on ntopics and
72 * the {@link Corpus#vocabularySize()}. Alpha remains at 0
73 * @param corpus
74 */
75 public void prepare(Corpus corpus) {
76 this.topicWord = new double[ntopics][corpus.vocabularySize()];
77 this.topicTotal = new double[ntopics];
78 this.alpha = 0;
79 this.iteration = 0;
80 this.likelihood = 0;
81 this.oldLikelihood = Double.NEGATIVE_INFINITY;
82 }
83
84 /**
85 * initialises the sufficient statistic holder based on ntopics and
86 * the vocabularySize. Alpha remains at 0
87 * @param vocabularySize
88 */
89 public void prepare(int vocabularySize) {
90 this.topicWord = new double[ntopics][vocabularySize];
91 this.topicTotal = new double[ntopics];
92 this.alpha = 0;
93 }
94
95 /**
96 * Increment a topic and word index by d. The totals are left untouched
97 * @param topicIndex
98 * @param wordIndex
99 * @param d
100 */
101 public void incTopicWord(int topicIndex, int wordIndex, double d) {
102 this.topicWord[topicIndex][wordIndex] += d;
103 }
104
105 /**
106 * Increment a topic and word index by d. The totals are left untouched
107 * @param topicIndex
108 * @param d
109 */
110 public void incTopicTotal(int topicIndex, double d) {
111 this.topicTotal[topicIndex] += d;
112 }
113
114 /**
115 * @param initialAlpha the alpha parameter for the topic multinomial dirichelet prior
116 */
117 public void setAlpha(double initialAlpha) {
118 this.alpha = initialAlpha;
119 }
120
121 /**
122 * This method also swaps the likelihoods (i.e. oldLikelihood == likelihood, likelhood = 0)
123 * @return a blank copy with unset alpha matching the current model's configuration
124 */
125 public LDAModel newInstance() {
126 LDAModel ret = new LDAModel(ntopics);
127 ret.prepare(this.topicWord[0].length);
128 ret.iteration = this.iteration;
129 ret.likelihood = 0;
130 ret.oldLikelihood = this.likelihood;
131 return ret;
132 }
133 }