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 }