Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/sources/HeuristicLab.Tests/HeuristicLab.Algorithms.DataAnalysis-3.4/GaussianProcessModelTest.cs @ 8463

Last change on this file since 8463 was 8463, checked in by gkronber, 12 years ago

#1902 improved GPR implementation

File size: 4.2 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.Linq;
23using HeuristicLab.Algorithms.DataAnalysis;
24using HeuristicLab.Problems.Instances.DataAnalysis;
25using Microsoft.VisualStudio.TestTools.UnitTesting;
26
27namespace HeuristicLab.Algorithms.DataAnalysis_34.Tests {
28  [TestClass]
29
30  // reference values calculated with Rasmussen's GPML MATLAB package
31  public class GaussianProcessModelTest {
32    [TestMethod]
33    [DeploymentItem(@"HeuristicLab.Algorithms.DataAnalysis-3.4/co2.txt")]
34    public void GaussianProcessModelOutputTest() {
35      var provider = new RegressionCSVInstanceProvider();
36      var problemData = provider.ImportData("co2.txt");
37
38      var targetVariable = "interpolated";
39      var allowedInputVariables = new string[] { "decimal date" };
40      var rows = Enumerable.Range(0, 401);
41
42      var meanFunction = new MeanConst();
43      var covarianceFunction = new CovarianceSum();
44      covarianceFunction.Terms.Add(new CovarianceSEiso());
45      var prod = new CovarianceProd();
46      prod.Factors.Add(new CovarianceSEiso());
47      prod.Factors.Add(new CovariancePeriodic());
48      covarianceFunction.Terms.Add(prod);
49
50      {
51        var hyp = new double[] { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
52        var model = new GaussianProcessModel(problemData.Dataset, targetVariable, allowedInputVariables, rows, hyp,
53                                             meanFunction,
54                                             covarianceFunction);
55        Assert.AreEqual(4.3170e+004, model.NegativeLogLikelihood, 1);
56
57        var dHyp = model.GetHyperparameterGradients();
58        Assert.AreEqual(-2.0171e+003, dHyp[0], 1);
59        Assert.AreEqual(-248.7932, dHyp[1], 1E-2);
60        var dHypCovExpected = new double[] { -0.5550e4, -5.5533e4, -0.2511e4, -2.7625e4, -1.3033e4, 0.0289e4, -2.7625e4 };
61        AssertEqual(dHypCovExpected, dHyp.Skip(2).ToArray(), 1);
62
63        var predTrain = model.GetEstimatedValues(problemData.Dataset, new int[] { 0, 400 }).ToArray();
64        Assert.AreEqual(310.5930, predTrain[0], 1e-3);
65        Assert.AreEqual(347.9993, predTrain[1], 1e-3);
66      }
67
68      {
69        var hyp = new double[] { 0.716427415979145, 0.029973094285941, 0.455535210579926, 3.438647883940457, 1.464114485889487, 3.001788584487478, 3.815289323309630, 4.374914122810222, 3.001788584487478 };
70        var model = new GaussianProcessModel(problemData.Dataset, targetVariable, allowedInputVariables, rows, hyp,
71                                             meanFunction,
72                                             covarianceFunction);
73        Assert.AreEqual(872.8448, model.NegativeLogLikelihood, 1e-3);
74
75        var dHyp = model.GetHyperparameterGradients();
76        Assert.AreEqual(0.8621, dHyp[0], 1e-3);
77        Assert.AreEqual(-0.0046, dHyp[1], 1e-3);
78        var dHypCovExpected = new double[] { 0.2652, -0.2386, 0.1706, -0.1744, 0.0000, 0.0000, -0.1744 };
79        AssertEqual(dHypCovExpected, dHyp.Skip(2).ToArray(), 1e-3);
80
81        var predTrain = model.GetEstimatedValues(problemData.Dataset, new int[] { 0, 400 }).ToArray();
82        Assert.AreEqual(315.3692, predTrain[0], 1e-3);
83        Assert.AreEqual(356.6076, predTrain[1], 1e-3);
84      }
85    }
86
87
88    private void AssertEqual(double[] expected, double[] actual, double delta = 1E-3) {
89      Assert.AreEqual(expected.Length, actual.Length);
90      for (int i = 0; i < expected.Length; i++)
91        Assert.AreEqual(expected[i], actual[i], delta);
92    }
93  }
94}
Note: See TracBrowser for help on using the repository browser.