Free cookie consent management tool by TermsFeed Policy Generator

source: branches/LearningClassifierSystems/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionSolutionBase.cs @ 15509

Last change on this file since 15509 was 8798, checked in by mkommend, 12 years ago

#1081: Reintegrated time series modeling branch into trunk.

File size: 15.0 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.Collections.Generic;
23using HeuristicLab.Common;
24using HeuristicLab.Data;
25using HeuristicLab.Optimization;
26using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
27
28namespace HeuristicLab.Problems.DataAnalysis {
29  [StorableClass]
30  public abstract class RegressionSolutionBase : DataAnalysisSolution, IRegressionSolution {
31    protected const string TrainingMeanSquaredErrorResultName = "Mean squared error (training)";
32    protected const string TestMeanSquaredErrorResultName = "Mean squared error (test)";
33    protected const string TrainingMeanAbsoluteErrorResultName = "Mean absolute error (training)";
34    protected const string TestMeanAbsoluteErrorResultName = "Mean absolute error (test)";
35    protected const string TrainingSquaredCorrelationResultName = "Pearson's R² (training)";
36    protected const string TestSquaredCorrelationResultName = "Pearson's R² (test)";
37    protected const string TrainingRelativeErrorResultName = "Average relative error (training)";
38    protected const string TestRelativeErrorResultName = "Average relative error (test)";
39    protected const string TrainingNormalizedMeanSquaredErrorResultName = "Normalized mean squared error (training)";
40    protected const string TestNormalizedMeanSquaredErrorResultName = "Normalized mean squared error (test)";
41    protected const string TrainingMeanErrorResultName = "Mean error (training)";
42    protected const string TestMeanErrorResultName = "Mean error (test)";
43
44    protected const string TrainingMeanSquaredErrorResultDescription = "Mean of squared errors of the model on the training partition";
45    protected const string TestMeanSquaredErrorResultDescription = "Mean of squared errors of the model on the test partition";
46    protected const string TrainingMeanAbsoluteErrorResultDescription = "Mean of absolute errors of the model on the training partition";
47    protected const string TestMeanAbsoluteErrorResultDescription = "Mean of absolute errors of the model on the test partition";
48    protected const string TrainingSquaredCorrelationResultDescription = "Squared Pearson's correlation coefficient of the model output and the actual values on the training partition";
49    protected const string TestSquaredCorrelationResultDescription = "Squared Pearson's correlation coefficient of the model output and the actual values on the test partition";
50    protected const string TrainingRelativeErrorResultDescription = "Average of the relative errors of the model output and the actual values on the training partition";
51    protected const string TestRelativeErrorResultDescription = "Average of the relative errors of the model output and the actual values on the test partition";
52    protected const string TrainingNormalizedMeanSquaredErrorResultDescription = "Normalized mean of squared errors of the model on the training partition";
53    protected const string TestNormalizedMeanSquaredErrorResultDescription = "Normalized mean of squared errors of the model on the test partition";
54    protected const string TrainingMeanErrorResultDescription = "Mean of errors of the model on the training partition";
55    protected const string TestMeanErrorResultDescription = "Mean of errors of the model on the test partition";
56
57    public new IRegressionModel Model {
58      get { return (IRegressionModel)base.Model; }
59      protected set { base.Model = value; }
60    }
61
62    public new IRegressionProblemData ProblemData {
63      get { return (IRegressionProblemData)base.ProblemData; }
64      set { base.ProblemData = value; }
65    }
66
67    public abstract IEnumerable<double> EstimatedValues { get; }
68    public abstract IEnumerable<double> EstimatedTrainingValues { get; }
69    public abstract IEnumerable<double> EstimatedTestValues { get; }
70    public abstract IEnumerable<double> GetEstimatedValues(IEnumerable<int> rows);
71
72    #region Results
73    public double TrainingMeanSquaredError {
74      get { return ((DoubleValue)this[TrainingMeanSquaredErrorResultName].Value).Value; }
75      private set { ((DoubleValue)this[TrainingMeanSquaredErrorResultName].Value).Value = value; }
76    }
77    public double TestMeanSquaredError {
78      get { return ((DoubleValue)this[TestMeanSquaredErrorResultName].Value).Value; }
79      private set { ((DoubleValue)this[TestMeanSquaredErrorResultName].Value).Value = value; }
80    }
81    public double TrainingMeanAbsoluteError {
82      get { return ((DoubleValue)this[TrainingMeanAbsoluteErrorResultName].Value).Value; }
83      private set { ((DoubleValue)this[TrainingMeanAbsoluteErrorResultName].Value).Value = value; }
84    }
85    public double TestMeanAbsoluteError {
86      get { return ((DoubleValue)this[TestMeanAbsoluteErrorResultName].Value).Value; }
87      private set { ((DoubleValue)this[TestMeanAbsoluteErrorResultName].Value).Value = value; }
88    }
89    public double TrainingRSquared {
90      get { return ((DoubleValue)this[TrainingSquaredCorrelationResultName].Value).Value; }
91      private set { ((DoubleValue)this[TrainingSquaredCorrelationResultName].Value).Value = value; }
92    }
93    public double TestRSquared {
94      get { return ((DoubleValue)this[TestSquaredCorrelationResultName].Value).Value; }
95      private set { ((DoubleValue)this[TestSquaredCorrelationResultName].Value).Value = value; }
96    }
97    public double TrainingRelativeError {
98      get { return ((DoubleValue)this[TrainingRelativeErrorResultName].Value).Value; }
99      private set { ((DoubleValue)this[TrainingRelativeErrorResultName].Value).Value = value; }
100    }
101    public double TestRelativeError {
102      get { return ((DoubleValue)this[TestRelativeErrorResultName].Value).Value; }
103      private set { ((DoubleValue)this[TestRelativeErrorResultName].Value).Value = value; }
104    }
105    public double TrainingNormalizedMeanSquaredError {
106      get { return ((DoubleValue)this[TrainingNormalizedMeanSquaredErrorResultName].Value).Value; }
107      private set { ((DoubleValue)this[TrainingNormalizedMeanSquaredErrorResultName].Value).Value = value; }
108    }
109    public double TestNormalizedMeanSquaredError {
110      get { return ((DoubleValue)this[TestNormalizedMeanSquaredErrorResultName].Value).Value; }
111      private set { ((DoubleValue)this[TestNormalizedMeanSquaredErrorResultName].Value).Value = value; }
112    }
113    public double TrainingMeanError {
114      get { return ((DoubleValue)this[TrainingMeanErrorResultName].Value).Value; }
115      private set { ((DoubleValue)this[TrainingMeanErrorResultName].Value).Value = value; }
116    }
117    public double TestMeanError {
118      get { return ((DoubleValue)this[TestMeanErrorResultName].Value).Value; }
119      private set { ((DoubleValue)this[TestMeanErrorResultName].Value).Value = value; }
120    }
121    #endregion
122
123    [StorableConstructor]
124    protected RegressionSolutionBase(bool deserializing) : base(deserializing) { }
125    protected RegressionSolutionBase(RegressionSolutionBase original, Cloner cloner)
126      : base(original, cloner) {
127    }
128    protected RegressionSolutionBase(IRegressionModel model, IRegressionProblemData problemData)
129      : base(model, problemData) {
130      Add(new Result(TrainingMeanSquaredErrorResultName, TrainingMeanSquaredErrorResultDescription, new DoubleValue()));
131      Add(new Result(TestMeanSquaredErrorResultName, TestMeanSquaredErrorResultDescription, new DoubleValue()));
132      Add(new Result(TrainingMeanAbsoluteErrorResultName, TrainingMeanAbsoluteErrorResultDescription, new DoubleValue()));
133      Add(new Result(TestMeanAbsoluteErrorResultName, TestMeanAbsoluteErrorResultDescription, new DoubleValue()));
134      Add(new Result(TrainingSquaredCorrelationResultName, TrainingSquaredCorrelationResultDescription, new DoubleValue()));
135      Add(new Result(TestSquaredCorrelationResultName, TestSquaredCorrelationResultDescription, new DoubleValue()));
136      Add(new Result(TrainingRelativeErrorResultName, TrainingRelativeErrorResultDescription, new PercentValue()));
137      Add(new Result(TestRelativeErrorResultName, TestRelativeErrorResultDescription, new PercentValue()));
138      Add(new Result(TrainingNormalizedMeanSquaredErrorResultName, TrainingNormalizedMeanSquaredErrorResultDescription, new DoubleValue()));
139      Add(new Result(TestNormalizedMeanSquaredErrorResultName, TestNormalizedMeanSquaredErrorResultDescription, new DoubleValue()));
140      Add(new Result(TrainingMeanErrorResultName, TrainingMeanErrorResultDescription, new DoubleValue()));
141      Add(new Result(TestMeanErrorResultName, TestMeanErrorResultDescription, new DoubleValue()));
142    }
143
144    [StorableHook(HookType.AfterDeserialization)]
145    private void AfterDeserialization() {
146      // BackwardsCompatibility3.4
147
148      #region Backwards compatible code, remove with 3.5
149
150      if (!ContainsKey(TrainingMeanAbsoluteErrorResultName)) {
151        OnlineCalculatorError errorState;
152        Add(new Result(TrainingMeanAbsoluteErrorResultName, "Mean of absolute errors of the model on the training partition", new DoubleValue()));
153        double trainingMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(EstimatedTrainingValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices), out errorState);
154        TrainingMeanAbsoluteError = errorState == OnlineCalculatorError.None ? trainingMAE : double.NaN;
155      }
156
157      if (!ContainsKey(TestMeanAbsoluteErrorResultName)) {
158        OnlineCalculatorError errorState;
159        Add(new Result(TestMeanAbsoluteErrorResultName, "Mean of absolute errors of the model on the test partition", new DoubleValue()));
160        double testMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(EstimatedTestValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices), out errorState);
161        TestMeanAbsoluteError = errorState == OnlineCalculatorError.None ? testMAE : double.NaN;
162      }
163
164      if (!ContainsKey(TrainingMeanErrorResultName)) {
165        OnlineCalculatorError errorState;
166        Add(new Result(TrainingMeanErrorResultName, "Mean of errors of the model on the training partition", new DoubleValue()));
167        double trainingME = OnlineMeanErrorCalculator.Calculate(EstimatedTrainingValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices), out errorState);
168        TrainingMeanError = errorState == OnlineCalculatorError.None ? trainingME : double.NaN;
169      }
170      if (!ContainsKey(TestMeanErrorResultName)) {
171        OnlineCalculatorError errorState;
172        Add(new Result(TestMeanErrorResultName, "Mean of errors of the model on the test partition", new DoubleValue()));
173        double testME = OnlineMeanErrorCalculator.Calculate(EstimatedTestValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices), out errorState);
174        TestMeanError = errorState == OnlineCalculatorError.None ? testME : double.NaN;
175      }
176      #endregion
177    }
178
179    protected override void RecalculateResults() {
180      CalculateRegressionResults();
181    }
182
183    protected void CalculateRegressionResults() {
184      IEnumerable<double> estimatedTrainingValues = EstimatedTrainingValues; // cache values
185      IEnumerable<double> originalTrainingValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices);
186      IEnumerable<double> estimatedTestValues = EstimatedTestValues; // cache values
187      IEnumerable<double> originalTestValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices);
188
189      OnlineCalculatorError errorState;
190      double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
191      TrainingMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingMSE : double.NaN;
192      double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
193      TestMeanSquaredError = errorState == OnlineCalculatorError.None ? testMSE : double.NaN;
194
195      double trainingMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
196      TrainingMeanAbsoluteError = errorState == OnlineCalculatorError.None ? trainingMAE : double.NaN;
197      double testMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
198      TestMeanAbsoluteError = errorState == OnlineCalculatorError.None ? testMAE : double.NaN;
199
200      double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
201      TrainingRSquared = errorState == OnlineCalculatorError.None ? trainingR2 : double.NaN;
202      double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
203      TestRSquared = errorState == OnlineCalculatorError.None ? testR2 : double.NaN;
204
205      double trainingRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
206      TrainingRelativeError = errorState == OnlineCalculatorError.None ? trainingRelError : double.NaN;
207      double testRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
208      TestRelativeError = errorState == OnlineCalculatorError.None ? testRelError : double.NaN;
209
210      double trainingNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
211      TrainingNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingNMSE : double.NaN;
212      double testNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
213      TestNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? testNMSE : double.NaN;
214
215      double trainingME = OnlineMeanErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
216      TrainingMeanError = errorState == OnlineCalculatorError.None ? trainingME : double.NaN;
217      double testME = OnlineMeanErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
218      TestMeanError = errorState == OnlineCalculatorError.None ? testME : double.NaN;
219    }
220  }
221}
Note: See TracBrowser for help on using the repository browser.