1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 2002-2015 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
4 | * and the BEACON Center for the Study of Evolution in Action.
|
---|
5 | *
|
---|
6 | * This file is part of HeuristicLab.
|
---|
7 | *
|
---|
8 | * HeuristicLab is free software: you can redistribute it and/or modify
|
---|
9 | * it under the terms of the GNU General Public License as published by
|
---|
10 | * the Free Software Foundation, either version 3 of the License, or
|
---|
11 | * (at your option) any later version.
|
---|
12 | *
|
---|
13 | * HeuristicLab is distributed in the hope that it will be useful,
|
---|
14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
16 | * GNU General Public License for more details.
|
---|
17 | *
|
---|
18 | * You should have received a copy of the GNU General Public License
|
---|
19 | * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
|
---|
20 | */
|
---|
21 | #endregion
|
---|
22 |
|
---|
23 | using System.Collections.Generic;
|
---|
24 | using HeuristicLab.Common;
|
---|
25 | using HeuristicLab.Core;
|
---|
26 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
27 | using HeuristicLab.Problems.DataAnalysis;
|
---|
28 |
|
---|
29 | namespace HeuristicLab.Algorithms.DataAnalysis {
|
---|
30 | [StorableType("ECCC9F9F-E3B5-4D16-AB7F-B1221A5F7033")]
|
---|
31 | // this class is used as a surrogate for persistence of an actual GBT model
|
---|
32 | // since the actual GBT model would be very large when persisted we only store all necessary information to
|
---|
33 | // recalculate the actual GBT model on demand
|
---|
34 | [Item("Gradient boosted tree model", "")]
|
---|
35 | public sealed class GradientBoostedTreesModelSurrogate : NamedItem, IGradientBoostedTreesModel {
|
---|
36 | // don't store the actual model!
|
---|
37 | private IGradientBoostedTreesModel actualModel; // the actual model is only recalculated when necessary
|
---|
38 |
|
---|
39 | [Storable]
|
---|
40 | private readonly IRegressionProblemData trainingProblemData;
|
---|
41 | [Storable]
|
---|
42 | private readonly uint seed;
|
---|
43 | [Storable]
|
---|
44 | private ILossFunction lossFunction;
|
---|
45 | [Storable]
|
---|
46 | private double r;
|
---|
47 | [Storable]
|
---|
48 | private double m;
|
---|
49 | [Storable]
|
---|
50 | private double nu;
|
---|
51 | [Storable]
|
---|
52 | private int iterations;
|
---|
53 | [Storable]
|
---|
54 | private int maxSize;
|
---|
55 |
|
---|
56 |
|
---|
57 | [StorableConstructor]
|
---|
58 | private GradientBoostedTreesModelSurrogate(bool deserializing) : base(deserializing) { }
|
---|
59 |
|
---|
60 | private GradientBoostedTreesModelSurrogate(GradientBoostedTreesModelSurrogate original, Cloner cloner)
|
---|
61 | : base(original, cloner) {
|
---|
62 | if (original.actualModel != null) this.actualModel = cloner.Clone(original.actualModel);
|
---|
63 |
|
---|
64 | this.trainingProblemData = cloner.Clone(original.trainingProblemData);
|
---|
65 | this.lossFunction = cloner.Clone(original.lossFunction);
|
---|
66 | this.seed = original.seed;
|
---|
67 | this.iterations = original.iterations;
|
---|
68 | this.maxSize = original.maxSize;
|
---|
69 | this.r = original.r;
|
---|
70 | this.m = original.m;
|
---|
71 | this.nu = original.nu;
|
---|
72 | }
|
---|
73 |
|
---|
74 | // create only the surrogate model without an actual model
|
---|
75 | public GradientBoostedTreesModelSurrogate(IRegressionProblemData trainingProblemData, uint seed, ILossFunction lossFunction, int iterations, int maxSize, double r, double m, double nu)
|
---|
76 | : base("Gradient boosted tree model", string.Empty) {
|
---|
77 | this.trainingProblemData = trainingProblemData;
|
---|
78 | this.seed = seed;
|
---|
79 | this.lossFunction = lossFunction;
|
---|
80 | this.iterations = iterations;
|
---|
81 | this.maxSize = maxSize;
|
---|
82 | this.r = r;
|
---|
83 | this.m = m;
|
---|
84 | this.nu = nu;
|
---|
85 | }
|
---|
86 |
|
---|
87 | // wrap an actual model in a surrograte
|
---|
88 | public GradientBoostedTreesModelSurrogate(IRegressionProblemData trainingProblemData, uint seed, ILossFunction lossFunction, int iterations, int maxSize, double r, double m, double nu, IGradientBoostedTreesModel model)
|
---|
89 | : this(trainingProblemData, seed, lossFunction, iterations, maxSize, r, m, nu) {
|
---|
90 | this.actualModel = model;
|
---|
91 | }
|
---|
92 |
|
---|
93 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
94 | return new GradientBoostedTreesModelSurrogate(this, cloner);
|
---|
95 | }
|
---|
96 |
|
---|
97 | // forward message to actual model (recalculate model first if necessary)
|
---|
98 | public IEnumerable<double> GetEstimatedValues(IDataset dataset, IEnumerable<int> rows) {
|
---|
99 | if (actualModel == null) actualModel = RecalculateModel();
|
---|
100 | return actualModel.GetEstimatedValues(dataset, rows);
|
---|
101 | }
|
---|
102 |
|
---|
103 | public IRegressionSolution CreateRegressionSolution(IRegressionProblemData problemData) {
|
---|
104 | return new RegressionSolution(this, (IRegressionProblemData)problemData.Clone());
|
---|
105 | }
|
---|
106 |
|
---|
107 |
|
---|
108 | private IGradientBoostedTreesModel RecalculateModel() {
|
---|
109 | return GradientBoostedTreesAlgorithmStatic.TrainGbm(trainingProblemData, lossFunction, maxSize, nu, r, m, iterations, seed).Model;
|
---|
110 | }
|
---|
111 |
|
---|
112 | public IEnumerable<IRegressionModel> Models {
|
---|
113 | get {
|
---|
114 | if (actualModel == null) actualModel = RecalculateModel();
|
---|
115 | return actualModel.Models;
|
---|
116 | }
|
---|
117 | }
|
---|
118 |
|
---|
119 | public IEnumerable<double> Weights {
|
---|
120 | get {
|
---|
121 | if (actualModel == null) actualModel = RecalculateModel();
|
---|
122 | return actualModel.Weights;
|
---|
123 | }
|
---|
124 | }
|
---|
125 | }
|
---|
126 | }
|
---|