source: trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression/3.4/SymbolicRegressionSolutionImpactValuesCalculator.cs @ 12720

Last change on this file since 12720 was 12720, checked in by bburlacu, 6 years ago

#2359: Changed the impact calculators so that the quality value necessary for impacts calculation is calculated with a separate method. Refactored the CalculateImpactAndReplacementValues method to return the new quality in an out-parameter (adjusted method signature in interface accordingly). Added Evaluate method to the regression and classification pruning operators that re-evaluates the tree using the problem evaluator after pruning was performed.

File size: 5.3 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 System;
23using System.Collections.Generic;
24using HeuristicLab.Common;
25using HeuristicLab.Core;
26using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
27using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
28
29namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Regression {
30  [StorableClass]
31  [Item("SymbolicRegressionSolutionImpactValuesCalculator", "Calculate symbolic expression tree node impact values for regression problems.")]
32  public class SymbolicRegressionSolutionImpactValuesCalculator : SymbolicDataAnalysisSolutionImpactValuesCalculator {
33    public SymbolicRegressionSolutionImpactValuesCalculator() { }
34
35    protected SymbolicRegressionSolutionImpactValuesCalculator(SymbolicRegressionSolutionImpactValuesCalculator original, Cloner cloner)
36      : base(original, cloner) { }
37    public override IDeepCloneable Clone(Cloner cloner) {
38      return new SymbolicRegressionSolutionImpactValuesCalculator(this, cloner);
39    }
40
41    [StorableConstructor]
42    protected SymbolicRegressionSolutionImpactValuesCalculator(bool deserializing) : base(deserializing) { }
43    public override double CalculateReplacementValue(ISymbolicDataAnalysisModel model, ISymbolicExpressionTreeNode node, IDataAnalysisProblemData problemData, IEnumerable<int> rows) {
44      var regressionModel = (ISymbolicRegressionModel)model;
45      var regressionProblemData = (IRegressionProblemData)problemData;
46
47      return CalculateReplacementValue(node, regressionModel.SymbolicExpressionTree, regressionModel.Interpreter, regressionProblemData.Dataset, rows);
48    }
49
50    public override double CalculateImpactValue(ISymbolicDataAnalysisModel model, ISymbolicExpressionTreeNode node, IDataAnalysisProblemData problemData, IEnumerable<int> rows, double qualityForImpactsCalculation = double.NaN) {
51      double impactValue, replacementValue, newQualityForImpactsCalculation;
52      CalculateImpactAndReplacementValues(model, node, problemData, rows, out impactValue, out replacementValue, out newQualityForImpactsCalculation, qualityForImpactsCalculation);
53      return impactValue;
54    }
55
56    public override void CalculateImpactAndReplacementValues(ISymbolicDataAnalysisModel model, ISymbolicExpressionTreeNode node,
57      IDataAnalysisProblemData problemData, IEnumerable<int> rows, out double impactValue, out double replacementValue, out double newQualityForImpactsCalculation,
58      double qualityForImpactsCalculation = Double.NaN) {
59      var regressionModel = (ISymbolicRegressionModel)model;
60      var regressionProblemData = (IRegressionProblemData)problemData;
61
62      var dataset = regressionProblemData.Dataset;
63      var targetValues = dataset.GetDoubleValues(regressionProblemData.TargetVariable, rows);
64
65      OnlineCalculatorError errorState;
66      if (double.IsNaN(qualityForImpactsCalculation))
67        qualityForImpactsCalculation = CalculateQualityForImpacts(regressionModel, regressionProblemData, rows);
68
69      replacementValue = CalculateReplacementValue(regressionModel, node, regressionProblemData, rows);
70      var constantNode = new ConstantTreeNode(new Constant()) { Value = replacementValue };
71
72      var cloner = new Cloner();
73      var tempModel = cloner.Clone(regressionModel);
74      var tempModelNode = (ISymbolicExpressionTreeNode)cloner.GetClone(node);
75
76      var tempModelParentNode = tempModelNode.Parent;
77      int i = tempModelParentNode.IndexOfSubtree(tempModelNode);
78      tempModelParentNode.RemoveSubtree(i);
79      tempModelParentNode.InsertSubtree(i, constantNode);
80
81      var estimatedValues = tempModel.GetEstimatedValues(dataset, rows);
82      double r = OnlinePearsonsRCalculator.Calculate(targetValues, estimatedValues, out errorState);
83      if (errorState != OnlineCalculatorError.None) r = 0.0;
84      newQualityForImpactsCalculation = r * r;
85
86      impactValue = qualityForImpactsCalculation - newQualityForImpactsCalculation;
87    }
88
89    public static double CalculateQualityForImpacts(ISymbolicRegressionModel model, IRegressionProblemData problemData, IEnumerable<int> rows) {
90      var estimatedValues = model.GetEstimatedValues(problemData.Dataset, rows); // also bounds the values
91      var targetValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, rows);
92      OnlineCalculatorError errorState;
93      var r = OnlinePearsonsRCalculator.Calculate(targetValues, estimatedValues, out errorState);
94      var quality = r * r;
95      if (errorState != OnlineCalculatorError.None) return double.NaN;
96      return quality;
97    }
98  }
99}
Note: See TracBrowser for help on using the repository browser.