Shark machine learning library
About Shark
News!
Contribute
Credits and copyright
Downloads
Getting Started
Installation
Using the docs
Documentation
Tutorials
Quick references
Class list
Global functions
FAQ
Showroom
examples
Supervised
KernelRegression.cpp
Go to the documentation of this file.
1
//===========================================================================
2
/*!
3
*
4
*
5
* \brief Kernel-based regression methods example program.
6
*
7
*
8
*
9
* \author T. Glasmachers
10
* \date -
11
*
12
*
13
* \par Copyright 1995-2017 Shark Development Team
14
*
15
* <BR><HR>
16
* This file is part of Shark.
17
* <http://shark-ml.org/>
18
*
19
* Shark is free software: you can redistribute it and/or modify
20
* it under the terms of the GNU Lesser General Public License as published
21
* by the Free Software Foundation, either version 3 of the License, or
22
* (at your option) any later version.
23
*
24
* Shark is distributed in the hope that it will be useful,
25
* but WITHOUT ANY WARRANTY; without even the implied warranty of
26
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
27
* GNU Lesser General Public License for more details.
28
*
29
* You should have received a copy of the GNU Lesser General Public License
30
* along with Shark. If not, see <http://www.gnu.org/licenses/>.
31
*
32
*/
33
//===========================================================================
34
35
#include <
shark/LinAlg/Base.h
>
36
#include <
shark/Core/Random.h
>
37
#include <
shark/Models/Kernels/GaussianRbfKernel.h
>
38
#include <
shark/Algorithms/Trainers/EpsilonSvmTrainer.h
>
39
#include <
shark/Algorithms/Trainers/RegularizationNetworkTrainer.h
>
40
#include <
shark/ObjectiveFunctions/Loss/SquaredLoss.h
>
41
#include <
shark/Data/Dataset.h
>
42
#include <
shark/Data/DataDistribution.h
>
43
44
45
using namespace
shark
;
46
47
48
int
main
()
49
{
50
// experiment settings
51
unsigned
int
ell = 200;
52
unsigned
int
tests = 10000;
53
double
C = 10.0;
54
double
gamma = 1.0 / C;
55
double
epsilon = 0.03;
56
57
GaussianRbfKernel<>
kernel(0.1);
58
SquaredLoss<>
loss;
59
60
// generate dataset
61
Wave
problem;
62
RegressionDataset
training = problem.
generateDataset
(ell);
63
RegressionDataset
test = problem.
generateDataset
(tests);
64
65
// define the machines
66
KernelExpansion<RealVector>
svm[2] = {
67
KernelExpansion<RealVector>
(),
68
KernelExpansion<RealVector>
()
69
};
70
71
// define the corresponding trainers
72
AbstractTrainer<KernelExpansion<RealVector>
>* trainer[2];
73
trainer[0] =
new
EpsilonSvmTrainer<RealVector>
(&kernel, C, epsilon);
74
trainer[1] =
new
RegularizationNetworkTrainer<RealVector>
(&kernel, gamma);
75
76
for
(
unsigned
int
i=0; i<2; i++)
77
{
78
std::cout<<
"METHOD"
<<(i+1) <<
" "
<< trainer[i]->name().c_str()<<std::endl;
79
std::cout<<
"training ..."
<<std::flush;
80
trainer[i]->
train
(svm[i], training);
81
std::cout<<
"done"
<<std::endl;
82
83
Data<RealVector>
output = svm[i](training.
inputs
());
84
double
train_error = loss.
eval
(training.
labels
(), output);
85
std::cout<<
"training error: "
<<train_error<<std::endl;
86
output = svm[i](test.
inputs
());
87
double
test_error = loss.
eval
(test.
labels
(), output);
88
std::cout<<
" test error: "
<<test_error<<
"\n\n"
;
89
}
90
91
delete
trainer[0];
92
delete
trainer[1];
93
}