Free cookie consent management tool by TermsFeed Policy Generator

source: branches/2971_named_intervals/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/DiscriminantFunctionClassificationSolutionBase.cs @ 16644

Last change on this file since 16644 was 16644, checked in by gkronber, 5 years ago

#2971: removed duplicate usings of HEAL.Attic and unnecessary usings

File size: 7.5 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2019 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;
23using System.Collections.Generic;
24using System.Linq;
25using HeuristicLab.Common;
26using HeuristicLab.Core;
27using HeuristicLab.Data;
28using HeuristicLab.Optimization;
29using HEAL.Attic;
30
31namespace HeuristicLab.Problems.DataAnalysis {
32  /// <summary>
33  /// Represents a classification solution that uses a discriminant function and classification thresholds.
34  /// </summary>
35  [StorableType("3668EBE0-128C-4BC4-902C-161670F98FAD")]
36  [Item("DiscriminantFunctionClassificationSolution", "Represents a classification solution that uses a discriminant function and classification thresholds.")]
37  public abstract class DiscriminantFunctionClassificationSolutionBase : ClassificationSolutionBase, IDiscriminantFunctionClassificationSolution {
38    private const string TrainingMeanSquaredErrorResultName = "Mean squared error (training)";
39    private const string TestMeanSquaredErrorResultName = "Mean squared error (test)";
40    private const string TrainingRSquaredResultName = "Pearson's R² (training)";
41    private const string TestRSquaredResultName = "Pearson's R² (test)";
42
43    public new IDiscriminantFunctionClassificationModel Model {
44      get { return (IDiscriminantFunctionClassificationModel)base.Model; }
45      protected set {
46        if (value != null && value != Model) {
47          if (Model != null) {
48            Model.ThresholdsChanged -= new EventHandler(Model_ThresholdsChanged);
49          }
50          value.ThresholdsChanged += new EventHandler(Model_ThresholdsChanged);
51          base.Model = value;
52        }
53      }
54    }
55
56    #region Results
57    public double TrainingMeanSquaredError {
58      get { return ((DoubleValue)this[TrainingMeanSquaredErrorResultName].Value).Value; }
59      private set { ((DoubleValue)this[TrainingMeanSquaredErrorResultName].Value).Value = value; }
60    }
61    public double TestMeanSquaredError {
62      get { return ((DoubleValue)this[TestMeanSquaredErrorResultName].Value).Value; }
63      private set { ((DoubleValue)this[TestMeanSquaredErrorResultName].Value).Value = value; }
64    }
65    public double TrainingRSquared {
66      get { return ((DoubleValue)this[TrainingRSquaredResultName].Value).Value; }
67      private set { ((DoubleValue)this[TrainingRSquaredResultName].Value).Value = value; }
68    }
69    public double TestRSquared {
70      get { return ((DoubleValue)this[TestRSquaredResultName].Value).Value; }
71      private set { ((DoubleValue)this[TestRSquaredResultName].Value).Value = value; }
72    }
73    #endregion
74
75    [StorableConstructor]
76    protected DiscriminantFunctionClassificationSolutionBase(StorableConstructorFlag _) : base(_) { }
77    protected DiscriminantFunctionClassificationSolutionBase(DiscriminantFunctionClassificationSolutionBase original, Cloner cloner)
78      : base(original, cloner) {
79      RegisterEventHandler();
80    }
81    protected DiscriminantFunctionClassificationSolutionBase(IDiscriminantFunctionClassificationModel model, IClassificationProblemData problemData)
82      : base(model, problemData) {
83      Add(new Result(TrainingMeanSquaredErrorResultName, "Mean of squared errors of the model on the training partition", new DoubleValue()));
84      Add(new Result(TestMeanSquaredErrorResultName, "Mean of squared errors of the model on the test partition", new DoubleValue()));
85      Add(new Result(TrainingRSquaredResultName, "Squared Pearson's correlation coefficient of the model output and the actual values on the training partition", new DoubleValue()));
86      Add(new Result(TestRSquaredResultName, "Squared Pearson's correlation coefficient of the model output and the actual values on the test partition", new DoubleValue()));
87      RegisterEventHandler();
88    }
89
90    [StorableHook(HookType.AfterDeserialization)]
91    private void AfterDeserialization() {
92      RegisterEventHandler();
93    }
94
95    protected void CalculateRegressionResults() {
96      double[] estimatedTrainingValues = EstimatedTrainingValues.ToArray(); // cache values
97      double[] originalTrainingValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices).ToArray();
98      double[] estimatedTestValues = EstimatedTestValues.ToArray(); // cache values
99      double[] originalTestValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices).ToArray();
100
101      OnlineCalculatorError errorState;
102      double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
103      TrainingMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingMSE : double.NaN;
104      double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
105      TestMeanSquaredError = errorState == OnlineCalculatorError.None ? testMSE : double.NaN;
106
107      double trainingR = OnlinePearsonsRCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
108      TrainingRSquared = errorState == OnlineCalculatorError.None ? trainingR*trainingR : double.NaN;
109      double testR = OnlinePearsonsRCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
110      TestRSquared = errorState == OnlineCalculatorError.None ? testR*testR : double.NaN;
111
112      double trainingNormalizedGini = NormalizedGiniCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
113      if (errorState != OnlineCalculatorError.None) trainingNormalizedGini = double.NaN;
114      double testNormalizedGini = NormalizedGiniCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
115      if (errorState != OnlineCalculatorError.None) testNormalizedGini = double.NaN;
116
117      TrainingNormalizedGiniCoefficient = trainingNormalizedGini;
118      TestNormalizedGiniCoefficient = testNormalizedGini;
119    }
120
121    private void RegisterEventHandler() {
122      Model.ThresholdsChanged += new EventHandler(Model_ThresholdsChanged);
123    }
124    private void DeregisterEventHandler() {
125      Model.ThresholdsChanged -= new EventHandler(Model_ThresholdsChanged);
126    }
127    private void Model_ThresholdsChanged(object sender, EventArgs e) {
128      OnModelThresholdsChanged(e);
129    }
130
131    protected virtual void OnModelThresholdsChanged(EventArgs e) {
132      OnModelChanged();
133    }
134
135    public abstract IEnumerable<double> EstimatedValues { get; }
136    public abstract IEnumerable<double> EstimatedTrainingValues { get; }
137    public abstract IEnumerable<double> EstimatedTestValues { get; }
138
139    public abstract IEnumerable<double> GetEstimatedValues(IEnumerable<int> rows);
140
141    protected override void RecalculateResults() {
142      base.RecalculateResults();
143      CalculateRegressionResults();
144    }
145  }
146}
Note: See TracBrowser for help on using the repository browser.