Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression.Views/3.4/InteractiveSymbolicRegressionSolutionSimplifierView.cs @ 8736

Last change on this file since 8736 was 8736, checked in by gkronber, 12 years ago

#1962: fixed bug in the view that caused a unit test fail.

File size: 7.4 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;
23using System.Collections.Generic;
24using System.Linq;
25using HeuristicLab.Common;
26using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
27using HeuristicLab.Problems.DataAnalysis.Symbolic.Views;
28
29namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Regression.Views {
30  public partial class InteractiveSymbolicRegressionSolutionSimplifierView : InteractiveSymbolicDataAnalysisSolutionSimplifierView {
31    private readonly ConstantTreeNode constantNode;
32    private readonly SymbolicExpressionTree tempTree;
33
34    public new SymbolicRegressionSolution Content {
35      get { return (SymbolicRegressionSolution)base.Content; }
36      set { base.Content = value; }
37    }
38
39    public InteractiveSymbolicRegressionSolutionSimplifierView()
40      : base() {
41      InitializeComponent();
42      this.Caption = "Interactive Regression Solution Simplifier";
43
44      constantNode = ((ConstantTreeNode)new Constant().CreateTreeNode());
45      ISymbolicExpressionTreeNode root = new ProgramRootSymbol().CreateTreeNode();
46      ISymbolicExpressionTreeNode start = new StartSymbol().CreateTreeNode();
47      root.AddSubtree(start);
48      tempTree = new SymbolicExpressionTree(root);
49    }
50
51    protected override void UpdateModel(ISymbolicExpressionTree tree) {
52      var model = new SymbolicRegressionModel(tree, Content.Model.Interpreter, Content.Model.LowerEstimationLimit, Content.Model.UpperEstimationLimit);
53      SymbolicRegressionModel.Scale(model, Content.ProblemData, Content.ProblemData.TargetVariable);
54      Content.Model = model;
55    }
56
57    protected override Dictionary<ISymbolicExpressionTreeNode, double> CalculateReplacementValues(ISymbolicExpressionTree tree) {
58      Dictionary<ISymbolicExpressionTreeNode, double> replacementValues = new Dictionary<ISymbolicExpressionTreeNode, double>();
59      foreach (ISymbolicExpressionTreeNode node in tree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPrefix()) {
60        replacementValues[node] = CalculateReplacementValue(node, tree);
61      }
62      return replacementValues;
63    }
64
65    protected override Dictionary<ISymbolicExpressionTreeNode, double> CalculateImpactValues(ISymbolicExpressionTree tree) {
66      var interpreter = Content.Model.Interpreter;
67      var dataset = Content.ProblemData.Dataset;
68      var rows = Content.ProblemData.TrainingIndices;
69      string targetVariable = Content.ProblemData.TargetVariable;
70      Dictionary<ISymbolicExpressionTreeNode, double> impactValues = new Dictionary<ISymbolicExpressionTreeNode, double>();
71      List<ISymbolicExpressionTreeNode> nodes = tree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPostfix().ToList();
72      var originalOutput = interpreter.GetSymbolicExpressionTreeValues(tree, dataset, rows).LimitToRange(Content.Model.LowerEstimationLimit, Content.Model.UpperEstimationLimit).ToArray();
73      var targetValues = dataset.GetDoubleValues(targetVariable, rows);
74      OnlineCalculatorError errorState;
75      double originalR2 = OnlinePearsonsRSquaredCalculator.Calculate(targetValues, originalOutput, out errorState);
76      if (errorState != OnlineCalculatorError.None) originalR2 = 0.0;
77
78      foreach (ISymbolicExpressionTreeNode node in nodes) {
79        var parent = node.Parent;
80        constantNode.Value = CalculateReplacementValue(node, tree);
81        ISymbolicExpressionTreeNode replacementNode = constantNode;
82        SwitchNode(parent, node, replacementNode);
83        var newOutput = interpreter.GetSymbolicExpressionTreeValues(tree, dataset, rows).LimitToRange(Content.Model.LowerEstimationLimit, Content.Model.UpperEstimationLimit);
84        double newR2 = OnlinePearsonsRSquaredCalculator.Calculate(targetValues, newOutput, out errorState);
85        if (errorState != OnlineCalculatorError.None) newR2 = 0.0;
86
87        // impact = 0 if no change
88        // impact < 0 if new solution is better
89        // impact > 0 if new solution is worse
90        impactValues[node] = originalR2 - newR2;
91        SwitchNode(parent, replacementNode, node);
92      }
93      return impactValues;
94    }
95
96    private double CalculateReplacementValue(ISymbolicExpressionTreeNode node, ISymbolicExpressionTree sourceTree) {
97      // remove old ADFs
98      while (tempTree.Root.SubtreeCount > 1) tempTree.Root.RemoveSubtree(1);
99      // clone ADFs of source tree
100      for (int i = 1; i < sourceTree.Root.SubtreeCount; i++) {
101        tempTree.Root.AddSubtree((ISymbolicExpressionTreeNode)sourceTree.Root.GetSubtree(i).Clone());
102      }
103      var start = tempTree.Root.GetSubtree(0);
104      while (start.SubtreeCount > 0) start.RemoveSubtree(0);
105      start.AddSubtree((ISymbolicExpressionTreeNode)node.Clone());
106      var interpreter = Content.Model.Interpreter;
107      var rows = Content.ProblemData.TrainingIndices;
108      return interpreter.GetSymbolicExpressionTreeValues(tempTree, Content.ProblemData.Dataset, rows)
109                         .LimitToRange(Content.Model.LowerEstimationLimit, Content.Model.UpperEstimationLimit)
110                         .Median();
111    }
112
113
114    private void SwitchNode(ISymbolicExpressionTreeNode root, ISymbolicExpressionTreeNode oldBranch, ISymbolicExpressionTreeNode newBranch) {
115      for (int i = 0; i < root.SubtreeCount; i++) {
116        if (root.GetSubtree(i) == oldBranch) {
117          root.RemoveSubtree(i);
118          root.InsertSubtree(i, newBranch);
119          return;
120        }
121      }
122    }
123
124    protected override void OnModelChanged() {
125      base.OnModelChanged();
126      if (Content != null)
127        btnOptimizeConstants.Enabled =
128          SymbolicRegressionConstantOptimizationEvaluator.CanOptimizeConstants(Content.Model.SymbolicExpressionTree);
129      else
130        btnOptimizeConstants.Enabled = false;
131    }
132    protected override void OnContentChanged() {
133      base.OnContentChanged();
134      base.OnModelChanged();
135      if (Content != null)
136        btnOptimizeConstants.Enabled =
137          SymbolicRegressionConstantOptimizationEvaluator.CanOptimizeConstants(Content.Model.SymbolicExpressionTree);
138      else
139        btnOptimizeConstants.Enabled = false;
140    }
141
142
143    protected override void btnOptimizeConstants_Click(object sender, EventArgs e) {
144      var model = Content.Model;
145      SymbolicRegressionConstantOptimizationEvaluator.OptimizeConstants(Content.Model.Interpreter, Content.Model.SymbolicExpressionTree, Content.ProblemData, Content.ProblemData.TrainingIndices,
146        applyLinearScaling: true, maxIterations: 50, upperEstimationLimit: model.UpperEstimationLimit, lowerEstimationLimit: model.LowerEstimationLimit);
147      UpdateModel(Content.Model.SymbolicExpressionTree);
148    }
149  }
150}
Note: See TracBrowser for help on using the repository browser.