Free cookie consent management tool by TermsFeed Policy Generator

source: branches/DataAnalysis/HeuristicLab.Problems.DataAnalysis.Regression/3.3/Symbolic/Analyzers/SymbolicRegressionOverfittingAnalyzer.cs @ 6627

Last change on this file since 6627 was 5275, checked in by gkronber, 14 years ago

Merged changes from trunk to data analysis exploration branch and added fractional distance metric evaluator. #1142

File size: 7.6 KB
RevLine 
[5275]1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2010 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.Analysis;
25using HeuristicLab.Common;
26using HeuristicLab.Core;
27using HeuristicLab.Data;
28using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
29using HeuristicLab.Operators;
30using HeuristicLab.Optimization;
31using HeuristicLab.Parameters;
32using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
33using HeuristicLab.Problems.DataAnalysis.Evaluators;
34using HeuristicLab.Problems.DataAnalysis.Symbolic;
35using System;
36
37namespace HeuristicLab.Problems.DataAnalysis.Regression.Symbolic.Analyzers {
38  [Item("SymbolicRegressionOverfittingAnalyzer", "Calculates and tracks correlation of training and validation fitness of symbolic regression models.")]
39  [StorableClass]
40  public sealed class SymbolicRegressionOverfittingAnalyzer : SymbolicRegressionValidationAnalyzer, ISymbolicRegressionAnalyzer {
41    private const string MaximizationParameterName = "Maximization";
42    private const string QualityParameterName = "Quality";
43    private const string TrainingValidationCorrelationParameterName = "TrainingValidationCorrelation";
44    private const string TrainingValidationCorrelationTableParameterName = "TrainingValidationCorrelationTable";
45    private const string LowerCorrelationThresholdParameterName = "LowerCorrelationThreshold";
46    private const string UpperCorrelationThresholdParameterName = "UpperCorrelationThreshold";
47    private const string OverfittingParameterName = "IsOverfitting";
48    private const string ResultsParameterName = "Results";
49
50    #region parameter properties
51    public ScopeTreeLookupParameter<DoubleValue> QualityParameter {
52      get { return (ScopeTreeLookupParameter<DoubleValue>)Parameters[QualityParameterName]; }
53    }
54    public ILookupParameter<BoolValue> MaximizationParameter {
55      get { return (ILookupParameter<BoolValue>)Parameters[MaximizationParameterName]; }
56    }
57    public ILookupParameter<DoubleValue> TrainingValidationQualityCorrelationParameter {
58      get { return (ILookupParameter<DoubleValue>)Parameters[TrainingValidationCorrelationParameterName]; }
59    }
60    public ILookupParameter<DataTable> TrainingValidationQualityCorrelationTableParameter {
61      get { return (ILookupParameter<DataTable>)Parameters[TrainingValidationCorrelationTableParameterName]; }
62    }
63    public IValueLookupParameter<DoubleValue> LowerCorrelationThresholdParameter {
64      get { return (IValueLookupParameter<DoubleValue>)Parameters[LowerCorrelationThresholdParameterName]; }
65    }
66    public IValueLookupParameter<DoubleValue> UpperCorrelationThresholdParameter {
67      get { return (IValueLookupParameter<DoubleValue>)Parameters[UpperCorrelationThresholdParameterName]; }
68    }
69    public ILookupParameter<BoolValue> OverfittingParameter {
70      get { return (ILookupParameter<BoolValue>)Parameters[OverfittingParameterName]; }
71    }
72    public ILookupParameter<ResultCollection> ResultsParameter {
73      get { return (ILookupParameter<ResultCollection>)Parameters[ResultsParameterName]; }
74    }
75    #endregion
76    #region properties
77    public BoolValue Maximization {
78      get { return MaximizationParameter.ActualValue; }
79    }
80    #endregion
81
82    [StorableConstructor]
83    private SymbolicRegressionOverfittingAnalyzer(bool deserializing) : base(deserializing) { }
84    private SymbolicRegressionOverfittingAnalyzer(SymbolicRegressionOverfittingAnalyzer original, Cloner cloner) : base(original, cloner) { }
85    public SymbolicRegressionOverfittingAnalyzer()
86      : base() {
87      Parameters.Add(new ScopeTreeLookupParameter<DoubleValue>(QualityParameterName, "Training fitness"));
88      Parameters.Add(new LookupParameter<BoolValue>(MaximizationParameterName, "The direction of optimization."));
89      Parameters.Add(new LookupParameter<DoubleValue>(TrainingValidationCorrelationParameterName, "Correlation of training and validation fitnesses"));
90      Parameters.Add(new LookupParameter<DataTable>(TrainingValidationCorrelationTableParameterName, "Data table of training and validation fitness correlation values over the whole run."));
91      Parameters.Add(new ValueLookupParameter<DoubleValue>(LowerCorrelationThresholdParameterName, "Lower threshold for correlation value that marks the boundary from non-overfitting to overfitting.", new DoubleValue(0.65)));
92      Parameters.Add(new ValueLookupParameter<DoubleValue>(UpperCorrelationThresholdParameterName, "Upper threshold for correlation value that marks the boundary from overfitting to non-overfitting.", new DoubleValue(0.75)));
93      Parameters.Add(new LookupParameter<BoolValue>(OverfittingParameterName, "Boolean indicator for overfitting."));
94      Parameters.Add(new LookupParameter<ResultCollection>(ResultsParameterName, "The results collection."));
95    }
96
97    [StorableHook(HookType.AfterDeserialization)]
98    private void AfterDeserialization() {
99    }
100
101    public override IDeepCloneable Clone(Cloner cloner) {
102      return new SymbolicRegressionOverfittingAnalyzer(this, cloner);
103    }
104
105    protected override void Analyze(SymbolicExpressionTree[] trees, double[] validationQuality) {
106      double[] trainingQuality = QualityParameter.ActualValue.Select(x => x.Value).ToArray();
107
108      double r = alglib.spearmancorr2(trainingQuality, validationQuality);
109
110      TrainingValidationQualityCorrelationParameter.ActualValue = new DoubleValue(r);
111
112      if (TrainingValidationQualityCorrelationTableParameter.ActualValue == null) {
113        var dataTable = new DataTable("Training and validation fitness correlation table", "Data table of training and validation fitness correlation values over the whole run.");
114        dataTable.Rows.Add(new DataRow("Training and validation fitness correlation", "Training and validation fitness correlation values"));
115        TrainingValidationQualityCorrelationTableParameter.ActualValue = dataTable;
116        ResultsParameter.ActualValue.Add(new Result(TrainingValidationCorrelationTableParameterName, dataTable));
117      }
118
119      TrainingValidationQualityCorrelationTableParameter.ActualValue.Rows["Training and validation fitness correlation"].Values.Add(r);
120
121      double correlationThreshold;
122      if (OverfittingParameter.ActualValue != null && OverfittingParameter.ActualValue.Value) {
123        // if is already overfitting => have to reach the upper threshold to switch back to non-overfitting state
124        correlationThreshold = UpperCorrelationThresholdParameter.ActualValue.Value;
125      } else {
126        // if currently in non-overfitting state => have to reach to lower threshold to switch to overfitting state
127        correlationThreshold = LowerCorrelationThresholdParameter.ActualValue.Value;
128      }
129      bool overfitting = r < correlationThreshold;
130
131      OverfittingParameter.ActualValue = new BoolValue(overfitting);
132    }
133  }
134}
Note: See TracBrowser for help on using the repository browser.