Free cookie consent management tool by TermsFeed Policy Generator

source: branches/HeuristicLab.Problems.GaussianProcessTuning/HeuristicLab.Problems.Instances.DataAnalysis.GaussianProcessRegression/VariousInstanceProvider.cs @ 10662

Last change on this file since 10662 was 9338, checked in by gkronber, 12 years ago

#1967: minor adaptations necessary for the EuroCAST presentation

File size: 10.5 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2012 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
4 *
5 * This file is part of HeuristicLab.
6 *
7 * HeuristicLab is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
11 *
12 * HeuristicLab is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15 * GNU General Public License for more details.
16 *
17 * You should have received a copy of the GNU General Public License
18 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
19 */
20#endregion
21
22using System;
23using System.Collections.Generic;
24using System.Linq;
25using HeuristicLab.Algorithms.DataAnalysis;
26using HeuristicLab.Data;
27
28namespace HeuristicLab.Problems.Instances.DataAnalysis {
29  public class InstanceProvider : ArtificialRegressionInstanceProvider {
30    public override string Name {
31      get { return "GPR Benchmark Problems"; }
32    }
33    public override string Description {
34      get { return ""; }
35    }
36    public override Uri WebLink {
37      get { return new Uri("http://dev.heuristiclab.com/trac/hl/core/wiki/AdditionalMaterial"); }
38    }
39    public override string ReferencePublication {
40      get { return ""; }
41    }
42
43    public override IEnumerable<IDataDescriptor> GetDataDescriptors() {
44      List<IDataDescriptor> descriptorList = new List<IDataDescriptor>();
45      descriptorList.Add(new GaussianProcessSEIso());
46      descriptorList.Add(new GaussianProcessSEIso1());
47      descriptorList.Add(new GaussianProcessSEIso2());
48      descriptorList.Add(new GaussianProcessSEIso3());
49      descriptorList.Add(new GaussianProcessSEIso4());
50      descriptorList.Add(new GaussianProcessSEIso5());
51      descriptorList.Add(new GaussianProcessSEIso6());
52      descriptorList.Add(new GaussianProcessPolyTen());
53      descriptorList.Add(new GaussianProcessSEIsoDependentNoise());
54      descriptorList.Add(new GaussianProcess2dPeriodic());
55
56      {
57        var cov = new CovarianceSum();
58        cov.Terms.Add(new CovarianceSquaredExponentialIso());
59        cov.Terms.Add(new CovarianceSquaredExponentialIso());
60        cov.Terms.Add(new CovarianceNoise());
61        var hyp = new double[] { -2.8, -0.1, 0.5, 0.3, -1.5 };
62        descriptorList.Add(new GaussianProcessRegressionInstance("SE+SE", cov, hyp));
63      }
64      {
65        var cov = new CovarianceSum();
66        cov.Terms.Add(new CovarianceRationalQuadraticIso());
67        cov.Terms.Add(new CovarianceRationalQuadraticIso());
68        cov.Terms.Add(new CovarianceNoise());
69        var hyp = new double[] { -3, 0, 0, -1.5, 0, 2.5, -1.5 };
70        descriptorList.Add(new GaussianProcessRegressionInstance("RQ+RQ", cov, hyp));
71      }
72      {
73        var cov = new CovarianceSum();
74        cov.Terms.Add(new CovariancePeriodic());
75        cov.Terms.Add(new CovariancePeriodic());
76        cov.Terms.Add(new CovarianceNoise());
77        var hyp = new double[] { 0, -1.8, -1.5, 0, -0.5, -1, -2.1 };
78        descriptorList.Add(new GaussianProcessRegressionInstance("Periodic+Periodic", cov, hyp));
79      }
80      {
81        var cov = new CovarianceSum();
82        cov.Terms.Add(new CovarianceMaternIso());
83        cov.Terms.Add(new CovarianceMaternIso());
84        cov.Terms.Add(new CovarianceNoise());
85        var hyp = new double[] { 0, 0, -1, 1, -4 };
86        descriptorList.Add(new GaussianProcessRegressionInstance("Matern1+Matern1", cov, hyp));
87      }
88      {
89        var cov = new CovarianceSum();
90        var m1 = new CovarianceMaternIso();
91        m1.DParameter.Value = m1.DParameter.ValidValues.First(v => v.Value == 3);
92        var m2 = new CovarianceMaternIso();
93        m2.DParameter.Value = m2.DParameter.ValidValues.First(v => v.Value == 3);
94        cov.Terms.Add(m1);
95        cov.Terms.Add(m2);
96        cov.Terms.Add(new CovarianceNoise());
97        var hyp = new double[] { -2.7, 0, -1, 1, -1.5 };
98        descriptorList.Add(new GaussianProcessRegressionInstance("Matern3+Matern3", cov, hyp));
99      }
100      {
101        var cov = new CovarianceSum();
102        cov.Terms.Add(new CovarianceRationalQuadraticIso());
103        cov.Terms.Add(new CovarianceSquaredExponentialIso());
104        cov.Terms.Add(new CovarianceNoise());
105        var hyp = new double[] { -1.5, -0.5, -3, -1, -1, -3 };
106        descriptorList.Add(new GaussianProcessRegressionInstance("RQ+SE", cov, hyp));
107      }
108      {
109        var cov = new CovarianceSum();
110        cov.Terms.Add(new CovarianceSquaredExponentialIso());
111        var prod = new CovarianceProduct();
112        prod.Factors.Add(new CovarianceLinear());
113        prod.Factors.Add(new CovarianceNoise());
114        cov.Terms.Add(prod);
115        cov.Terms.Add(new CovarianceNoise());
116        var hyp = new double[] { -3, 0, 0, -1.5 };
117        descriptorList.Add(new GaussianProcessRegressionInstance("SE+Linear*Noise", cov, hyp));
118      }
119      {
120        var cov = new CovarianceSum();
121        cov.Terms.Add(new CovarianceSquaredExponentialIso());
122        cov.Terms.Add(new CovariancePeriodic());
123        cov.Terms.Add(new CovarianceNoise());
124        var hyp = new double[] { -1, 0, 0, -1.5, 0, -2 };
125        descriptorList.Add(new GaussianProcessRegressionInstance("SE+Periodic", cov, hyp));
126      }
127
128      {
129        var cov = new CovarianceSum();
130        cov.Terms.Add(new CovarianceSquaredExponentialIso());
131        cov.Terms.Add(new CovarianceNoise());
132        var hyp = new double[] { -2.5, 0, -7 };
133        descriptorList.Add(new GaussianProcessRegressionInstance1D("1D: SE", cov, hyp));
134      }
135      {
136        var cov = new CovarianceSum();
137        cov.Terms.Add(new CovarianceSquaredExponentialIso());
138        cov.Terms.Add(new CovarianceNoise());
139        var hyp = new double[] { -2.5, 0, -7 };
140        descriptorList.Add(new GaussianProcessRegressionDemo("1D: SE Noise", cov, hyp));
141      }
142      {
143        var cov = new CovarianceSum();
144        cov.Terms.Add(new CovarianceRationalQuadraticIso());
145        cov.Terms.Add(new CovarianceNoise());
146        var hyp = new double[] { -2.5, 0, -1, -7 };
147        descriptorList.Add(new GaussianProcessRegressionDemo("1D: RQ Noise", cov, hyp));
148      }
149      {
150        var cov = new CovarianceSum();
151        var t = new CovarianceMaternIso();
152        t.DParameter.Value = t.DParameter.ValidValues.First(x => x.Value == 3);
153        cov.Terms.Add(t);
154        cov.Terms.Add(new CovarianceNoise());
155        var hyp = new double[] { -1.5, 0, -7 };
156        descriptorList.Add(new GaussianProcessRegressionDemo("1D: Matern3 Noise", cov, hyp));
157      }
158      {
159        var cov = new CovarianceSum();
160        var t = new CovariancePeriodic();
161        t.PeriodParameter.Value = new DoubleValue(0.3);
162        cov.Terms.Add(t);
163        cov.Terms.Add(new CovarianceNoise());
164        var hyp = new double[] { 0, 0, -7 };
165        descriptorList.Add(new GaussianProcessRegressionDemo("1D: Periodic Noise", cov, hyp));
166      }
167      {
168        var cov = new CovarianceSum();
169        cov.Terms.Add(new CovarianceRationalQuadraticIso());
170        cov.Terms.Add(new CovarianceNoise());
171        var hyp = new double[] { -2.5, 0, -2, -7 };
172        descriptorList.Add(new GaussianProcessRegressionInstance1D("1D: RQ", cov, hyp));
173      }
174      {
175        var cov = new CovarianceSum();
176        cov.Terms.Add(new CovarianceMaternIso());
177        cov.Terms.Add(new CovarianceNoise());
178        var hyp = new double[] { -2, 0, -7 };
179        descriptorList.Add(new GaussianProcessRegressionInstance1D("1D: Matern1", cov, hyp));
180      }
181      {
182        var cov = new CovarianceSum();
183        cov.Terms.Add(new CovariancePeriodic());
184        cov.Terms.Add(new CovarianceNoise());
185        var hyp = new double[] { 0, -1.3, 0, -7 };
186        descriptorList.Add(new GaussianProcessRegressionInstance1D("1D: Periodic", cov, hyp));
187      }
188      {
189        var cov = new CovarianceSum();
190        cov.Terms.Add(new CovarianceSquaredExponentialIso());
191        cov.Terms.Add(new CovarianceNoise());
192        var hyp = new double[] { -2.5, 0, -7 };
193        descriptorList.Add(new GaussianProcessRegressionInstance1D("1D: SE", cov, hyp));
194      }
195      {
196        var cov = new CovarianceSum();
197        var m1 = new CovarianceMask();
198        m1.SelectedDimensionsParameter.Value = new IntArray(new int[] { 0 });
199        m1.CovarianceFunctionParameter.Value = new CovarianceSquaredExponentialIso();
200
201        var m2 = new CovarianceMask();
202        m2.SelectedDimensionsParameter.Value = new IntArray(new int[] { 1 });
203        m2.CovarianceFunctionParameter.Value = new CovarianceRationalQuadraticIso();
204
205        cov.Terms.Add(m1);
206        cov.Terms.Add(m2);
207        cov.Terms.Add(new CovarianceNoise());
208        var hyp = new double[] { -2.5, 0, -2.0, 0, -2, -7 };
209        descriptorList.Add(new GaussianProcessRegressionInstance2D("2D: SE+RQ", cov, hyp));
210      }
211      {
212        var cov = new CovarianceSum();
213        var m1 = new CovarianceMask();
214        m1.SelectedDimensionsParameter.Value = new IntArray(new int[] { 0 });
215        m1.CovarianceFunctionParameter.Value = new CovarianceSquaredExponentialIso();
216
217        var m2 = new CovarianceMask();
218        m2.SelectedDimensionsParameter.Value = new IntArray(new int[] { 1 });
219        m2.CovarianceFunctionParameter.Value = new CovarianceMaternIso();
220
221        cov.Terms.Add(m1);
222        cov.Terms.Add(m2);
223        cov.Terms.Add(new CovarianceNoise());
224        var hyp = new double[] { -2.5, 0, 2, 0, -7 };
225        descriptorList.Add(new GaussianProcessRegressionInstance2D("2D: SE+Matern1", cov, hyp));
226      }
227      {
228        var cov = new CovarianceSum();
229        var m1 = new CovarianceMask();
230        m1.SelectedDimensionsParameter.Value = new IntArray(new int[] { 0 });
231        m1.CovarianceFunctionParameter.Value = new CovarianceSquaredExponentialIso();
232
233        var m2 = new CovarianceMask();
234        m2.SelectedDimensionsParameter.Value = new IntArray(new int[] { 1 });
235        m2.CovarianceFunctionParameter.Value = new CovariancePeriodic();
236
237        cov.Terms.Add(m1);
238        cov.Terms.Add(m2);
239        cov.Terms.Add(new CovarianceNoise());
240        var hyp = new double[] { -2.5, 0, 0, -1.3, 0, -7 };
241        descriptorList.Add(new GaussianProcessRegressionInstance2D("2D: SE+Periodic", cov, hyp));
242      }
243
244
245
246      return descriptorList;
247    }
248  }
249}
250
Note: See TracBrowser for help on using the repository browser.