/usr/include/shogun/regression/svr/LibSVR.h is in libshogun-dev 3.2.0-7.3build4.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 | /*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* Written (W) 1999-2009 Soeren Sonnenburg
* Written (W) 2013 Heiko Strathmann
* Copyright (C) 1999-2009 Fraunhofer Institute FIRST and Max-Planck-Society
*/
#ifndef _LIBSVR_H___
#define _LIBSVR_H___
#include <stdio.h>
#include <shogun/lib/common.h>
#include <shogun/classifier/svm/SVM.h>
#include <shogun/lib/external/shogun_libsvm.h>
#include <shogun/regression/Regression.h>
namespace shogun
{
/** @brief Class LibSVR, performs support vector regression using LibSVM.
*
* The SVR solution can be expressed as
* \f[
* f({\bf x})=\sum_{i=1}^{N} \alpha_i k({\bf x}, {\bf x_i})+b
* \f]
*
* where \f$\alpha\f$ and \f$b\f$ are determined in training, i.e. using a
* pre-specified kernel, a given tube-epsilon for the epsilon insensitive
* loss, the follwoing quadratic problem is minimized (using sequential
* minimal decomposition (SMO))
*
* \f{eqnarray*}
* \max_{{\bf \alpha},{\bf \alpha}^*} &-\frac{1}{2}\sum_{i,j=1}^N(\alpha_i-\alpha_i^*)(\alpha_j-\alpha_j^*){\bf x}_i^T {\bf x}_j -\sum_{i=1}^N(\alpha_i+\alpha_i^*)\epsilon - \sum_{i=1}^N(\alpha_i-\alpha_i^*)y_i\\
* \mbox{wrt}:& {\bf \alpha},{\bf \alpha}^*\in{\bf R}^N\\
* \mbox{s.t.}:& 0\leq \alpha_i,\alpha_i^*\leq C,\, \forall i=1\dots N\\
* &\sum_{i=1}^N(\alpha_i-\alpha_i^*)y_i=0
* \f}
*
*
* Note that the SV regression problem is reduced to the standard SV
* classification problem by introducing artificial labels \f$-y_i\f$ which
* leads to the epsilon insensitive loss constraints * \f{eqnarray*}
* {\bf w}^T{\bf x}_i+b-c_i-\xi_i\leq 0,&\, \forall i=1\dots N\\
* -{\bf w}^T{\bf x}_i-b-c_i^*-\xi_i^*\leq 0,&\, \forall i=1\dots N
* \f}
* with \f$c_i=y_i+ \epsilon\f$ and \f$c_i^*=-y_i+ \epsilon\f$
*
* This class also support the \f$\nu\f$-SVR regression version of the problem,
* where \f$\nu\f$ replaces the \f$\epsilon\f$ parameter and represents an
* upper bound on the fraction of margin errors and a lower bound on the fraction
* of support vectors. While it is easier to interpret, the resulting
* optimization problem usually takes longer to solve. Note that these different
* parameters do not result in different predictive power. For a given problem,
* the best SVR for each parametrization will lead to the same results.
* See the letter "Training \f$\nu\f$-Support Vector Regression: Theory and Algorithms" by
* Chih-Chung Chang and Chih-Jen Lin for the relation of \f$\epsilon\f$-SVR and
* \f$\nu\f$-SVR.
*/
#ifndef DOXYGEN_SHOULD_SKIP_THIS
enum LIBSVR_SOLVER_TYPE
{
LIBSVR_EPSILON_SVR = 1,
LIBSVR_NU_SVR = 2
};
#endif
class CLibSVR : public CSVM
{
public:
/** problem type */
MACHINE_PROBLEM_TYPE(PT_REGRESSION);
/** default constructor, creates a EPISOLON-SVR */
CLibSVR();
/** constructor
*
* @param C constant C
* @param svr_param tube epsilon or SVR-NU depending on solver type
* @param k kernel
* @param lab labels
* @param st solver type to use, EPSILON-SVR or NU-SVR
*/
CLibSVR(float64_t C, float64_t svr_param, CKernel* k, CLabels* lab,
LIBSVR_SOLVER_TYPE st=LIBSVR_EPSILON_SVR);
virtual ~CLibSVR();
/** get classifier type
*
* @return classifie type LIBSVR
*/
virtual EMachineType get_classifier_type();
/** @return object name */
virtual const char* get_name() const { return "LibSVR"; }
protected:
/** train regression
*
* @param data training data (parameter can be avoided if distance or
* kernel-based regressor are used and distance/kernels are
* initialized with train data)
*
* @return whether training was successful
*/
virtual bool train_machine(CFeatures* data=NULL);
protected:
/** SVM problem */
svm_problem problem;
/** SVM parameter */
svm_parameter param;
/** SVM model */
struct svm_model* model;
/** solver type */
LIBSVR_SOLVER_TYPE solver_type;
};
}
#endif
|