-
-
Notifications
You must be signed in to change notification settings - Fork 1k
/
NewtonSVM.h
124 lines (103 loc) · 3.22 KB
/
NewtonSVM.h
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) 2012 Harshit Syal
* Copyright (C) 2012 Harshit Syal
*/
#ifndef _NEWTONSVM_H___
#define _NEWTONSVM_H___
#include <shogun/lib/common.h>
#include <shogun/machine/LinearMachine.h>
#include <shogun/features/DotFeatures.h>
#include <shogun/features/Labels.h>
namespace shogun
{
#ifdef HAVE_LAPACK
/** @brief NewtonSVM,
* In this Implementation linear SVM is trained in its primal form using Newton-like iterations.
* This Implementation is ported from the Olivier Chapelles fast newton based SVM solver, Which could be found here :http://mloss.org/software/view/30/
* For further information on this implementation of SVM refer to this paper: http://www.kyb.mpg.de/publications/attachments/neco_%5B0%5D.pdf
*/
class CNewtonSVM : public CLinearMachine
{
public:
/** default constructor */
CNewtonSVM();
/** constructor
* @param C constant C
* @param itr constant no of iterations
* @param traindat training features
* @param trainlab labels for features
*/
CNewtonSVM(float64_t C, CDotFeatures* traindat, CLabels* trainlab, int32_t itr=20);
virtual ~CNewtonSVM();
/** get classifier type
*
* @return classifier type NewtonSVM
*/
virtual inline EClassifierType get_classifier_type() { return CT_NEWTONSVM; }
/**
* set C
* @param C constant C
*/
inline void set_C(float64_t c) { C=c; }
/** get epsilon
* @return epsilon
*/
inline float64_t get_epsilon() { return epsilon; }
/**
* set epsilon
* @param epsilon constant epsilon
*/
inline void set_epsilon(float64_t e) { epsilon=e; }
/** get C
* @return C
*/
inline float64_t get_C() { return C; }
/** set if bias shall be enabled
* @param enable_bias if bias shall be enabled
*/
inline void set_bias_enabled(bool enable_bias) { use_bias=enable_bias; }
/** get if bias is enabled
* @return if bias is enabled
*/
inline bool get_bias_enabled() { return use_bias; }
/** set num_iter
* @return num_iter
*/
inline int32_t get_num_iter() {return num_iter;}
/** set iter
* @param num_iter number of iterations
*/
inline void set_num_iter(int32_t iter) { num_iter=iter; }
/** @return object name */
inline virtual const char* get_name() const { return "NewtonSVM"; }
protected:
/** train SVM classifier
*
* @param data training data (parameter can be avoided if distance or
* kernel-based classifiers are used and distance/kernels are
* initialized with train data)
*
* @return whether training was successful
*/
virtual bool train_machine(CFeatures* data=NULL);
private:
void obj_fun_linear(float64_t* weights, float64_t* out, float64_t* obj,
int32_t* sv, int32_t* numsv, float64_t* grad);
void line_search_linear(float64_t* weights, float64_t* d,
float64_t* out, float64_t* tx);
protected:
/** lambda=1/C */
float64_t lambda, C, epsilon;
float64_t prec;
int32_t x_n, x_d, num_iter;
/** if bias is used */
bool use_bias;
};
#endif //HAVE_LAPACK
}
#endif //_NEWTONSVM_H___