source: trunk/yat/classifier/NBC.h @ 959

Last change on this file since 959 was 959, checked in by Peter, 14 years ago

Fixed so NBC and SVM are throwing when unexpected DataLookup2D is
apssed to make_classifier or predict.

Speeding up NBC::predict by separating weighted code from
unweighted. Also fixed some bugs in NBC.

  • Property svn:eol-style set to native
  • Property svn:keywords set to Id
File size: 3.1 KB
Line 
1#ifndef _theplu_yat_classifier_nbc_
2#define _theplu_yat_classifier_nbc_
3
4// $Id: NBC.h 959 2007-10-10 16:49:39Z peter $
5
6/*
7  Copyright (C) 2006 Jari Häkkinen, Markus Ringnér, Peter Johansson
8  Copyright (C) 2007 Peter Johansson
9
10  This file is part of the yat library, http://trac.thep.lu.se/trac/yat
11
12  The yat library is free software; you can redistribute it and/or
13  modify it under the terms of the GNU General Public License as
14  published by the Free Software Foundation; either version 2 of the
15  License, or (at your option) any later version.
16
17  The yat library is distributed in the hope that it will be useful,
18  but WITHOUT ANY WARRANTY; without even the implied warranty of
19  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
20  General Public License for more details.
21
22  You should have received a copy of the GNU General Public License
23  along with this program; if not, write to the Free Software
24  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
25  02111-1307, USA.
26*/
27
28#include "SupervisedClassifier.h"
29#include "yat/utility/matrix.h"
30
31namespace theplu {
32namespace yat {
33namespace classifier { 
34
35  class DataLookup1D;
36  class DataLookup2D;
37  class MatrixLookup;
38  class MatrixLookupWeighted;
39  class Target;
40
41  /**
42     @brief Naive Bayesian Classification.
43 
44     Each class is modelled as a multinormal distribution with
45     features being independent: \f$ p(x|c) = \prod
46     \frac{1}{\sqrt{2\pi\sigma_i^2}} \exp \left(
47     \frac{(x_i-m_i)^2}{2\sigma_i^2)} \right)\f$
48  */
49  class NBC : public SupervisedClassifier
50  {
51 
52  public:
53    ///
54    /// Constructor taking the training data, the target vector, and
55    /// the distance measure as input.
56    ///
57    NBC(const MatrixLookup&, const Target&);
58   
59    ///
60    /// Constructor taking the training data with weights, the target
61    /// vector, the distance measure, and a weight matrix for the
62    /// training data as input.
63    ///
64    NBC(const MatrixLookupWeighted&, const Target&);
65
66    virtual ~NBC();
67
68    const DataLookup2D& data(void) const;
69
70
71    SupervisedClassifier* make_classifier(const DataLookup2D&, 
72                                          const Target&) const;
73   
74    ///
75    /// Train the classifier using the training data.
76    ///
77    /// For each class mean and variance are estimated for each
78    /// feature (see Averager and AveragerWeighted for details).
79    ///
80    /// @return true if training succedeed.
81    ///
82    bool train();
83
84   
85    /**
86       Each sample (column) in \a data is predicted and predictions
87       are returned in the corresponding column in passed \a res. Each
88       row in \a res corresponds to a class. The prediction is the
89       estimated probability that sample belong to class \f$ j \f$
90
91       \f$ P_j = \frac{1}{Z}\prod_i{\frac{1}{\sigma_i}}
92       \exp(\frac{w_i(x_i-\mu_i)^2}{\sigma_i^2})\f$, where \f$ \mu_i
93       \f$ and \f$ \sigma_i^2 \f$ are the estimated mean and variance,
94       respectively. If \a data is a MatrixLookup is equivalent to
95       using all weight equal to unity.
96    */
97    void predict(const DataLookup2D& data, utility::matrix& res) const;
98
99
100  private:
101    utility::matrix centroids_;
102    utility::matrix sigma2_;
103    const DataLookup2D& data_;
104
105    double sum_logsigma(size_t i) const;
106
107
108  };
109 
110}}} // of namespace classifier, yat, and theplu
111
112#endif
Note: See TracBrowser for help on using the repository browser.