Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/sources/HeuristicLab.Algorithms.DataAnalysis/3.4/GaussianProcess/MeanFunctions/MeanProduct.cs @ 8929

Last change on this file since 8929 was 8929, checked in by gkronber, 11 years ago

#1902: moved covariance and mean functions to folders

File size: 3.5 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
21using System.Linq;
22using HeuristicLab.Common;
23using HeuristicLab.Core;
24using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
25
26namespace HeuristicLab.Algorithms.DataAnalysis {
27  [StorableClass]
28  [Item(Name = "MeanProduct", Description = "Product of mean functions for Gaussian processes.")]
29  public sealed class MeanProduct : Item, IMeanFunction {
30    [Storable]
31    private ItemList<IMeanFunction> factors;
32
33    [Storable]
34    private int numberOfVariables;
35
36    public ItemList<IMeanFunction> Factors {
37      get { return factors; }
38    }
39
40    [StorableConstructor]
41    private MeanProduct(bool deserializing)
42      : base(deserializing) {
43    }
44
45    private MeanProduct(MeanProduct original, Cloner cloner)
46      : base(original, cloner) {
47      this.factors = cloner.Clone(original.factors);
48      this.numberOfVariables = original.numberOfVariables;
49    }
50
51    public MeanProduct() {
52      this.factors = new ItemList<IMeanFunction>();
53    }
54    public override IDeepCloneable Clone(Cloner cloner) {
55      return new MeanProduct(this, cloner);
56    }
57
58    public int GetNumberOfParameters(int numberOfVariables) {
59      this.numberOfVariables = numberOfVariables;
60      return factors.Select(t => t.GetNumberOfParameters(numberOfVariables)).Sum();
61    }
62
63    public void SetParameter(double[] hyp) {
64      int offset = 0;
65      foreach (var t in factors) {
66        var numberOfParameters = t.GetNumberOfParameters(numberOfVariables);
67        t.SetParameter(hyp.Skip(offset).Take(numberOfParameters).ToArray());
68        offset += numberOfParameters;
69      }
70    }
71
72    public double[] GetMean(double[,] x) {
73      var res = factors.First().GetMean(x);
74      foreach (var t in factors.Skip(1)) {
75        var a = t.GetMean(x);
76        for (int i = 0; i < res.Length; i++) res[i] *= a[i];
77      }
78      return res;
79    }
80
81    public double[] GetGradients(int k, double[,] x) {
82      double[] res = Enumerable.Repeat(1.0, x.GetLength(0)).ToArray();
83      // find index of factor for the given k
84      int j = 0;
85      while (k >= factors[j].GetNumberOfParameters(numberOfVariables)) {
86        k -= factors[j].GetNumberOfParameters(numberOfVariables);
87        j++;
88      }
89      for (int i = 0; i < factors.Count; i++) {
90        var f = factors[i];
91        if (i == j) {
92          // multiply gradient
93          var g = f.GetGradients(k, x);
94          for (int ii = 0; ii < res.Length; ii++) res[ii] *= g[ii];
95        } else {
96          // multiply mean
97          var m = f.GetMean(x);
98          for (int ii = 0; ii < res.Length; ii++) res[ii] *= m[ii];
99        }
100      }
101      return res;
102    }
103  }
104}
Note: See TracBrowser for help on using the repository browser.