source: trunk/sources/HeuristicLab.Algorithms.DataAnalysis/3.4/RandomForest/RandomForestClassification.cs @ 12504

Last change on this file since 12504 was 12504, checked in by mkommend, 5 years ago

#2025: Changed categories for all creatables.

File size: 7.9 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2015 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 HeuristicLab.Common;
23using HeuristicLab.Core;
24using HeuristicLab.Data;
25using HeuristicLab.Optimization;
26using HeuristicLab.Parameters;
27using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
28using HeuristicLab.Problems.DataAnalysis;
29
30namespace HeuristicLab.Algorithms.DataAnalysis {
31  /// <summary>
32  /// Random forest classification data analysis algorithm.
33  /// </summary>
34  [Item("Random Forest Classification", "Random forest classification data analysis algorithm (wrapper for ALGLIB).")]
35  [Creatable(CreatableAttribute.Categories.DataAnalysisClassification, Priority = 120)]
36  [StorableClass]
37  public sealed class RandomForestClassification : FixedDataAnalysisAlgorithm<IClassificationProblem> {
38    private const string RandomForestClassificationModelResultName = "Random forest classification solution";
39    private const string NumberOfTreesParameterName = "Number of trees";
40    private const string RParameterName = "R";
41    private const string MParameterName = "M";
42    private const string SeedParameterName = "Seed";
43    private const string SetSeedRandomlyParameterName = "SetSeedRandomly";
44
45    #region parameter properties
46    public IFixedValueParameter<IntValue> NumberOfTreesParameter {
47      get { return (IFixedValueParameter<IntValue>)Parameters[NumberOfTreesParameterName]; }
48    }
49    public IFixedValueParameter<DoubleValue> RParameter {
50      get { return (IFixedValueParameter<DoubleValue>)Parameters[RParameterName]; }
51    }
52    public IFixedValueParameter<DoubleValue> MParameter {
53      get { return (IFixedValueParameter<DoubleValue>)Parameters[MParameterName]; }
54    }
55    public IFixedValueParameter<IntValue> SeedParameter {
56      get { return (IFixedValueParameter<IntValue>)Parameters[SeedParameterName]; }
57    }
58    public IFixedValueParameter<BoolValue> SetSeedRandomlyParameter {
59      get { return (IFixedValueParameter<BoolValue>)Parameters[SetSeedRandomlyParameterName]; }
60    }
61    #endregion
62    #region properties
63    public int NumberOfTrees {
64      get { return NumberOfTreesParameter.Value.Value; }
65      set { NumberOfTreesParameter.Value.Value = value; }
66    }
67    public double R {
68      get { return RParameter.Value.Value; }
69      set { RParameter.Value.Value = value; }
70    }
71    public double M {
72      get { return MParameter.Value.Value; }
73      set { MParameter.Value.Value = value; }
74    }
75    public int Seed {
76      get { return SeedParameter.Value.Value; }
77      set { SeedParameter.Value.Value = value; }
78    }
79    public bool SetSeedRandomly {
80      get { return SetSeedRandomlyParameter.Value.Value; }
81      set { SetSeedRandomlyParameter.Value.Value = value; }
82    }
83    #endregion
84
85    [StorableConstructor]
86    private RandomForestClassification(bool deserializing) : base(deserializing) { }
87    private RandomForestClassification(RandomForestClassification original, Cloner cloner)
88      : base(original, cloner) {
89    }
90
91    public RandomForestClassification()
92      : base() {
93      Parameters.Add(new FixedValueParameter<IntValue>(NumberOfTreesParameterName, "The number of trees in the forest. Should be between 50 and 100", new IntValue(50)));
94      Parameters.Add(new FixedValueParameter<DoubleValue>(RParameterName, "The ratio of the training set that will be used in the construction of individual trees (0<r<=1). Should be adjusted depending on the noise level in the dataset in the range from 0.66 (low noise) to 0.05 (high noise). This parameter should be adjusted to achieve good generalization error.", new DoubleValue(0.3)));
95      Parameters.Add(new FixedValueParameter<DoubleValue>(MParameterName, "The ratio of features that will be used in the construction of individual trees (0<m<=1)", new DoubleValue(0.5)));
96      Parameters.Add(new FixedValueParameter<IntValue>(SeedParameterName, "The random seed used to initialize the new pseudo random number generator.", new IntValue(0)));
97      Parameters.Add(new FixedValueParameter<BoolValue>(SetSeedRandomlyParameterName, "True if the random seed should be set to a random value, otherwise false.", new BoolValue(true)));
98      Problem = new ClassificationProblem();
99    }
100
101    [StorableHook(HookType.AfterDeserialization)]
102    private void AfterDeserialization() {
103      if (!Parameters.ContainsKey(MParameterName))
104        Parameters.Add(new FixedValueParameter<DoubleValue>(MParameterName, "The ratio of features that will be used in the construction of individual trees (0<m<=1)", new DoubleValue(0.5)));
105      if (!Parameters.ContainsKey(SeedParameterName))
106        Parameters.Add(new FixedValueParameter<IntValue>(SeedParameterName, "The random seed used to initialize the new pseudo random number generator.", new IntValue(0)));
107      if (!Parameters.ContainsKey((SetSeedRandomlyParameterName)))
108        Parameters.Add(new FixedValueParameter<BoolValue>(SetSeedRandomlyParameterName, "True if the random seed should be set to a random value, otherwise false.", new BoolValue(true)));
109    }
110
111    public override IDeepCloneable Clone(Cloner cloner) {
112      return new RandomForestClassification(this, cloner);
113    }
114
115    #region random forest
116    protected override void Run() {
117      double rmsError, relClassificationError, outOfBagRmsError, outOfBagRelClassificationError;
118      if (SetSeedRandomly) Seed = new System.Random().Next();
119
120      var solution = CreateRandomForestClassificationSolution(Problem.ProblemData, NumberOfTrees, R, M, Seed, out rmsError, out relClassificationError, out outOfBagRmsError, out outOfBagRelClassificationError);
121      Results.Add(new Result(RandomForestClassificationModelResultName, "The random forest classification solution.", solution));
122      Results.Add(new Result("Root mean square error", "The root of the mean of squared errors of the random forest regression solution on the training set.", new DoubleValue(rmsError)));
123      Results.Add(new Result("Relative classification error", "Relative classification error of the random forest regression solution on the training set.", new PercentValue(relClassificationError)));
124      Results.Add(new Result("Root mean square error (out-of-bag)", "The out-of-bag root of the mean of squared errors of the random forest regression solution.", new DoubleValue(outOfBagRmsError)));
125      Results.Add(new Result("Relative classification error (out-of-bag)", "The out-of-bag relative classification error  of the random forest regression solution.", new PercentValue(outOfBagRelClassificationError)));
126    }
127
128    public static IClassificationSolution CreateRandomForestClassificationSolution(IClassificationProblemData problemData, int nTrees, double r, double m, int seed,
129      out double rmsError, out double relClassificationError, out double outOfBagRmsError, out double outOfBagRelClassificationError) {
130      var model = RandomForestModel.CreateClassificationModel(problemData, nTrees, r, m, seed, out rmsError, out relClassificationError, out outOfBagRmsError, out outOfBagRelClassificationError);
131      return new RandomForestClassificationSolution((IClassificationProblemData)problemData.Clone(), model);
132    }
133    #endregion
134  }
135}
Note: See TracBrowser for help on using the repository browser.