Free cookie consent management tool by TermsFeed Policy Generator

source: branches/HiveHiveEngine/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionSolutionBase.cs @ 7317

Last change on this file since 7317 was 7272, checked in by gkronber, 13 years ago

#1737 added online calculator for the mean error and results for the mean error of regression solutions on the training and test partitions.
File size: 13.5 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
21
22using System.Collections.Generic;
23using System.Linq;
24using HeuristicLab.Common;
25using HeuristicLab.Data;
26using HeuristicLab.Optimization;
27using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
28
29namespace HeuristicLab.Problems.DataAnalysis {
30  [StorableClass]
31  public abstract class RegressionSolutionBase : DataAnalysisSolution, IRegressionSolution {
32    private const string TrainingMeanSquaredErrorResultName = "Mean squared error (training)";
33    private const string TestMeanSquaredErrorResultName = "Mean squared error (test)";
34    private const string TrainingMeanAbsoluteErrorResultName = "Mean absolute error (training)";
35    private const string TestMeanAbsoluteErrorResultName = "Mean absolute error (test)";
36    private const string TrainingSquaredCorrelationResultName = "Pearson's R² (training)";
37    private const string TestSquaredCorrelationResultName = "Pearson's R² (test)";
38    private const string TrainingRelativeErrorResultName = "Average relative error (training)";
39    private const string TestRelativeErrorResultName = "Average relative error (test)";
40    private const string TrainingNormalizedMeanSquaredErrorResultName = "Normalized mean squared error (training)";
41    private const string TestNormalizedMeanSquaredErrorResultName = "Normalized mean squared error (test)";
42    private const string TrainingMeanErrorResultName = "Mean error (training)";
43    private const string TestMeanErrorResultName = "Mean error (test)";
44
45    public new IRegressionModel Model {
46      get { return (IRegressionModel)base.Model; }
47      protected set { base.Model = value; }
48    }
49
50    public new IRegressionProblemData ProblemData {
51      get { return (IRegressionProblemData)base.ProblemData; }
52      set { base.ProblemData = value; }
53    }
54
55    public abstract IEnumerable<double> EstimatedValues { get; }
56    public abstract IEnumerable<double> EstimatedTrainingValues { get; }
57    public abstract IEnumerable<double> EstimatedTestValues { get; }
58    public abstract IEnumerable<double> GetEstimatedValues(IEnumerable<int> rows);
59
60    #region Results
61    public double TrainingMeanSquaredError {
62      get { return ((DoubleValue)this[TrainingMeanSquaredErrorResultName].Value).Value; }
63      private set { ((DoubleValue)this[TrainingMeanSquaredErrorResultName].Value).Value = value; }
64    }
65    public double TestMeanSquaredError {
66      get { return ((DoubleValue)this[TestMeanSquaredErrorResultName].Value).Value; }
67      private set { ((DoubleValue)this[TestMeanSquaredErrorResultName].Value).Value = value; }
68    }
69    public double TrainingMeanAbsoluteError {
70      get { return ((DoubleValue)this[TrainingMeanAbsoluteErrorResultName].Value).Value; }
71      private set { ((DoubleValue)this[TrainingMeanAbsoluteErrorResultName].Value).Value = value; }
72    }
73    public double TestMeanAbsoluteError {
74      get { return ((DoubleValue)this[TestMeanAbsoluteErrorResultName].Value).Value; }
75      private set { ((DoubleValue)this[TestMeanAbsoluteErrorResultName].Value).Value = value; }
76    }
77    public double TrainingRSquared {
78      get { return ((DoubleValue)this[TrainingSquaredCorrelationResultName].Value).Value; }
79      private set { ((DoubleValue)this[TrainingSquaredCorrelationResultName].Value).Value = value; }
80    }
81    public double TestRSquared {
82      get { return ((DoubleValue)this[TestSquaredCorrelationResultName].Value).Value; }
83      private set { ((DoubleValue)this[TestSquaredCorrelationResultName].Value).Value = value; }
84    }
85    public double TrainingRelativeError {
86      get { return ((DoubleValue)this[TrainingRelativeErrorResultName].Value).Value; }
87      private set { ((DoubleValue)this[TrainingRelativeErrorResultName].Value).Value = value; }
88    }
89    public double TestRelativeError {
90      get { return ((DoubleValue)this[TestRelativeErrorResultName].Value).Value; }
91      private set { ((DoubleValue)this[TestRelativeErrorResultName].Value).Value = value; }
92    }
93    public double TrainingNormalizedMeanSquaredError {
94      get { return ((DoubleValue)this[TrainingNormalizedMeanSquaredErrorResultName].Value).Value; }
95      private set { ((DoubleValue)this[TrainingNormalizedMeanSquaredErrorResultName].Value).Value = value; }
96    }
97    public double TestNormalizedMeanSquaredError {
98      get { return ((DoubleValue)this[TestNormalizedMeanSquaredErrorResultName].Value).Value; }
99      private set { ((DoubleValue)this[TestNormalizedMeanSquaredErrorResultName].Value).Value = value; }
100    }
101    public double TrainingMeanError {
102      get { return ((DoubleValue)this[TrainingMeanErrorResultName].Value).Value; }
103      private set { ((DoubleValue)this[TrainingMeanErrorResultName].Value).Value = value; }
104    }
105    public double TestMeanError {
106      get { return ((DoubleValue)this[TestMeanErrorResultName].Value).Value; }
107      private set { ((DoubleValue)this[TestMeanErrorResultName].Value).Value = value; }
108    }
109    #endregion
110
111    [StorableConstructor]
112    protected RegressionSolutionBase(bool deserializing) : base(deserializing) { }
113    protected RegressionSolutionBase(RegressionSolutionBase original, Cloner cloner)
114      : base(original, cloner) {
115    }
116    protected RegressionSolutionBase(IRegressionModel model, IRegressionProblemData problemData)
117      : base(model, problemData) {
118      Add(new Result(TrainingMeanSquaredErrorResultName, "Mean of squared errors of the model on the training partition", new DoubleValue()));
119      Add(new Result(TestMeanSquaredErrorResultName, "Mean of squared errors of the model on the test partition", new DoubleValue()));
120      Add(new Result(TrainingMeanAbsoluteErrorResultName, "Mean of absolute errors of the model on the training partition", new DoubleValue()));
121      Add(new Result(TestMeanAbsoluteErrorResultName, "Mean of absolute errors of the model on the test partition", new DoubleValue()));
122      Add(new Result(TrainingSquaredCorrelationResultName, "Squared Pearson's correlation coefficient of the model output and the actual values on the training partition", new DoubleValue()));
123      Add(new Result(TestSquaredCorrelationResultName, "Squared Pearson's correlation coefficient of the model output and the actual values on the test partition", new DoubleValue()));
124      Add(new Result(TrainingRelativeErrorResultName, "Average of the relative errors of the model output and the actual values on the training partition", new PercentValue()));
125      Add(new Result(TestRelativeErrorResultName, "Average of the relative errors of the model output and the actual values on the test partition", new PercentValue()));
126      Add(new Result(TrainingNormalizedMeanSquaredErrorResultName, "Normalized mean of squared errors of the model on the training partition", new DoubleValue()));
127      Add(new Result(TestNormalizedMeanSquaredErrorResultName, "Normalized mean of squared errors of the model on the test partition", new DoubleValue()));
128      Add(new Result(TrainingMeanErrorResultName, "Mean of errors of the model on the training partition", new DoubleValue()));
129      Add(new Result(TestMeanErrorResultName, "Mean of errors of the model on the test partition", new DoubleValue()));
130    }
131
132    [StorableHook(HookType.AfterDeserialization)]
133    private void AfterDeserialization() {
134      // BackwardsCompatibility3.4
135
136      #region Backwards compatible code, remove with 3.5
137
138      if (!ContainsKey(TrainingMeanAbsoluteErrorResultName)) {
139        OnlineCalculatorError errorState;
140        Add(new Result(TrainingMeanAbsoluteErrorResultName, "Mean of absolute errors of the model on the training partition", new DoubleValue()));
141        double trainingMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(EstimatedTrainingValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndizes), out errorState);
142        TrainingMeanAbsoluteError = errorState == OnlineCalculatorError.None ? trainingMAE : double.NaN;
143      }
144
145      if (!ContainsKey(TestMeanAbsoluteErrorResultName)) {
146        OnlineCalculatorError errorState;
147        Add(new Result(TestMeanAbsoluteErrorResultName, "Mean of absolute errors of the model on the test partition", new DoubleValue()));
148        double testMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(EstimatedTestValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndizes), out errorState);
149        TestMeanAbsoluteError = errorState == OnlineCalculatorError.None ? testMAE : double.NaN;
150      }
151
152      if (!ContainsKey(TrainingMeanErrorResultName)) {
153        OnlineCalculatorError errorState;
154        Add(new Result(TrainingMeanErrorResultName, "Mean of errors of the model on the training partition", new DoubleValue()));
155        double trainingME = OnlineMeanErrorCalculator.Calculate(EstimatedTrainingValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndizes), out errorState);
156        TrainingMeanError = errorState == OnlineCalculatorError.None ? trainingME : double.NaN;
157      }
158      if (!ContainsKey(TestMeanErrorResultName)) {
159        OnlineCalculatorError errorState;
160        Add(new Result(TestMeanErrorResultName, "Mean of errors of the model on the test partition", new DoubleValue()));
161        double testME = OnlineMeanErrorCalculator.Calculate(EstimatedTestValues, ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndizes), out errorState);
162        TestMeanError = errorState == OnlineCalculatorError.None ? testME : double.NaN;
163      }
164      #endregion
165    }
166
167    protected void CalculateResults() {
168      double[] estimatedTrainingValues = EstimatedTrainingValues.ToArray(); // cache values
169      double[] originalTrainingValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndizes).ToArray();
170      double[] estimatedTestValues = EstimatedTestValues.ToArray(); // cache values
171      double[] originalTestValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndizes).ToArray();
172
173      OnlineCalculatorError errorState;
174      double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
175      TrainingMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingMSE : double.NaN;
176      double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
177      TestMeanSquaredError = errorState == OnlineCalculatorError.None ? testMSE : double.NaN;
178
179      double trainingMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
180      TrainingMeanAbsoluteError = errorState == OnlineCalculatorError.None ? trainingMAE : double.NaN;
181      double testMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
182      TestMeanAbsoluteError = errorState == OnlineCalculatorError.None ? testMAE : double.NaN;
183
184      double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
185      TrainingRSquared = errorState == OnlineCalculatorError.None ? trainingR2 : double.NaN;
186      double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
187      TestRSquared = errorState == OnlineCalculatorError.None ? testR2 : double.NaN;
188
189      double trainingRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
190      TrainingRelativeError = errorState == OnlineCalculatorError.None ? trainingRelError : double.NaN;
191      double testRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
192      TestRelativeError = errorState == OnlineCalculatorError.None ? testRelError : double.NaN;
193
194      double trainingNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
195      TrainingNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingNMSE : double.NaN;
196      double testNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
197      TestNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? testNMSE : double.NaN;
198
199      double trainingME = OnlineMeanErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
200      TrainingMeanError = errorState == OnlineCalculatorError.None ? trainingME : double.NaN;
201      double testME = OnlineMeanErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
202      TestMeanError = errorState == OnlineCalculatorError.None ? testME : double.NaN;
203    }
204  }
205}
Note: See TracBrowser for help on using the repository browser.