source: branches/2971_named_intervals/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression/3.4/SymbolicRegressionSolution.cs @ 16589

Last change on this file since 16589 was 16589, checked in by chaider, 5 months ago

#2971

  • Added new Evaluator to evaulate Pearson RSquared with repsect to given constraints
  • changes in SymbolicRegressionSolution
File size: 11.5 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2018 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.Core;
26using HeuristicLab.Data;
27using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
28using HeuristicLab.Optimization;
29using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
30using HeuristicLab.Problems.DataAnalysis.Implementation;
31
32namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Regression {
33  /// <summary>
34  /// Represents a symbolic regression solution (model + data) and attributes of the solution like accuracy and complexity
35  /// </summary>
36  [StorableClass]
37  [Item(Name = "SymbolicRegressionSolution", Description = "Represents a symbolic regression solution (model + data) and attributes of the solution like accuracy and complexity.")]
38  public sealed class SymbolicRegressionSolution : RegressionSolution, ISymbolicRegressionSolution {
39    private const string ModelLengthResultName = "Model Length";
40    private const string ModelDepthResultName = "Model Depth";
41
42    private const string EstimationLimitsResultsResultName = "Estimation Limits Results";
43    private const string EstimationLimitsResultName = "Estimation Limits";
44    private const string TrainingUpperEstimationLimitHitsResultName = "Training Upper Estimation Limit Hits";
45    private const string TestLowerEstimationLimitHitsResultName = "Test Lower Estimation Limit Hits";
46    private const string TrainingLowerEstimationLimitHitsResultName = "Training Lower Estimation Limit Hits";
47    private const string TestUpperEstimationLimitHitsResultName = "Test Upper Estimation Limit Hits";
48    private const string TrainingNaNEvaluationsResultName = "Training NaN Evaluations";
49    private const string TestNaNEvaluationsResultName = "Test NaN Evaluations";
50
51    private const string IntervalEvaluationResultName = "Interval Evaluation";
52    private const string EstimatedDerivationInterval = "Interval";
53
54    public new ISymbolicRegressionModel Model {
55      get { return (ISymbolicRegressionModel)base.Model; }
56      set { base.Model = value; }
57    }
58    ISymbolicDataAnalysisModel ISymbolicDataAnalysisSolution.Model {
59      get { return (ISymbolicDataAnalysisModel)base.Model; }
60    }
61    public int ModelLength {
62      get { return ((IntValue)this[ModelLengthResultName].Value).Value; }
63      private set { ((IntValue)this[ModelLengthResultName].Value).Value = value; }
64    }
65
66    public int ModelDepth {
67      get { return ((IntValue)this[ModelDepthResultName].Value).Value; }
68      private set { ((IntValue)this[ModelDepthResultName].Value).Value = value; }
69    }
70
71    private ResultCollection EstimationLimitsResultCollection {
72      get { return (ResultCollection)this[EstimationLimitsResultsResultName].Value; }
73    }
74    public DoubleLimit EstimationLimits {
75      get { return (DoubleLimit)EstimationLimitsResultCollection[EstimationLimitsResultName].Value; }
76    }
77
78    public int TrainingUpperEstimationLimitHits {
79      get { return ((IntValue)EstimationLimitsResultCollection[TrainingUpperEstimationLimitHitsResultName].Value).Value; }
80      private set { ((IntValue)EstimationLimitsResultCollection[TrainingUpperEstimationLimitHitsResultName].Value).Value = value; }
81    }
82    public int TestUpperEstimationLimitHits {
83      get { return ((IntValue)EstimationLimitsResultCollection[TestUpperEstimationLimitHitsResultName].Value).Value; }
84      private set { ((IntValue)EstimationLimitsResultCollection[TestUpperEstimationLimitHitsResultName].Value).Value = value; }
85    }
86    public int TrainingLowerEstimationLimitHits {
87      get { return ((IntValue)EstimationLimitsResultCollection[TrainingLowerEstimationLimitHitsResultName].Value).Value; }
88      private set { ((IntValue)EstimationLimitsResultCollection[TrainingLowerEstimationLimitHitsResultName].Value).Value = value; }
89    }
90    public int TestLowerEstimationLimitHits {
91      get { return ((IntValue)EstimationLimitsResultCollection[TestLowerEstimationLimitHitsResultName].Value).Value; }
92      private set { ((IntValue)EstimationLimitsResultCollection[TestLowerEstimationLimitHitsResultName].Value).Value = value; }
93    }
94    public int TrainingNaNEvaluations {
95      get { return ((IntValue)EstimationLimitsResultCollection[TrainingNaNEvaluationsResultName].Value).Value; }
96      private set { ((IntValue)EstimationLimitsResultCollection[TrainingNaNEvaluationsResultName].Value).Value = value; }
97    }
98    public int TestNaNEvaluations {
99      get { return ((IntValue)EstimationLimitsResultCollection[TestNaNEvaluationsResultName].Value).Value; }
100      private set { ((IntValue)EstimationLimitsResultCollection[TestNaNEvaluationsResultName].Value).Value = value; }
101    }
102
103    private NamedIntervals IntervalEvaluaitonCollection => 
104      (NamedIntervals) this[IntervalEvaluationResultName].Value;
105
106
107
108    [StorableConstructor]
109    private SymbolicRegressionSolution(bool deserializing) : base(deserializing) { }
110    private SymbolicRegressionSolution(SymbolicRegressionSolution original, Cloner cloner)
111      : base(original, cloner) {
112    }
113    public SymbolicRegressionSolution(ISymbolicRegressionModel model, IRegressionProblemData problemData)
114      : base(model, problemData) {
115      foreach (var node in model.SymbolicExpressionTree.Root.IterateNodesPrefix().OfType<SymbolicExpressionTreeTopLevelNode>())
116        node.SetGrammar(null);
117
118      Add(new Result(ModelLengthResultName, "Length of the symbolic regression model.", new IntValue()));
119      Add(new Result(ModelDepthResultName, "Depth of the symbolic regression model.", new IntValue()));
120
121      ResultCollection estimationLimitResults = new ResultCollection();
122      estimationLimitResults.Add(new Result(EstimationLimitsResultName, "", new DoubleLimit()));
123      estimationLimitResults.Add(new Result(TrainingUpperEstimationLimitHitsResultName, "", new IntValue()));
124      estimationLimitResults.Add(new Result(TestUpperEstimationLimitHitsResultName, "", new IntValue()));
125      estimationLimitResults.Add(new Result(TrainingLowerEstimationLimitHitsResultName, "", new IntValue()));
126      estimationLimitResults.Add(new Result(TestLowerEstimationLimitHitsResultName, "", new IntValue()));
127      estimationLimitResults.Add(new Result(TrainingNaNEvaluationsResultName, "", new IntValue()));
128      estimationLimitResults.Add(new Result(TestNaNEvaluationsResultName, "", new IntValue()));
129      Add(new Result(EstimationLimitsResultsResultName, "Results concerning the estimation limits of symbolic regression solution", estimationLimitResults));
130      RecalculateResults();
131
132      var namedIntervalCollection = new NamedIntervals();
133      Add(new Result(IntervalEvaluationResultName, "Results concerning the derivation of symbolic regression solution", namedIntervalCollection));
134      GetIntervalEvaulations(namedIntervalCollection);
135    }
136
137    public override IDeepCloneable Clone(Cloner cloner) {
138      return new SymbolicRegressionSolution(this, cloner);
139    }
140
141    [StorableHook(HookType.AfterDeserialization)]
142    private void AfterDeserialization() {
143      if (!ContainsKey(EstimationLimitsResultsResultName)) {
144        ResultCollection estimationLimitResults = new ResultCollection();
145        estimationLimitResults.Add(new Result(EstimationLimitsResultName, "", new DoubleLimit()));
146        estimationLimitResults.Add(new Result(TrainingUpperEstimationLimitHitsResultName, "", new IntValue()));
147        estimationLimitResults.Add(new Result(TestUpperEstimationLimitHitsResultName, "", new IntValue()));
148        estimationLimitResults.Add(new Result(TrainingLowerEstimationLimitHitsResultName, "", new IntValue()));
149        estimationLimitResults.Add(new Result(TestLowerEstimationLimitHitsResultName, "", new IntValue()));
150        estimationLimitResults.Add(new Result(TrainingNaNEvaluationsResultName, "", new IntValue()));
151        estimationLimitResults.Add(new Result(TestNaNEvaluationsResultName, "", new IntValue()));
152        Add(new Result(EstimationLimitsResultsResultName, "Results concerning the estimation limits of symbolic regression solution", estimationLimitResults));
153        CalculateResults();
154
155        var namedIntervalCollection = new NamedIntervals();
156        Add(new Result(IntervalEvaluationResultName, "Results concerning the derivation of symbolic regression solution", namedIntervalCollection));
157        GetIntervalEvaulations(namedIntervalCollection);
158      }
159    }
160
161    protected override void RecalculateResults() {
162      base.RecalculateResults();
163      CalculateResults();
164    }
165
166    private void GetIntervalEvaulations(NamedIntervals intervalEvaluation) {
167      var interpreter = new IntervalInterpreter();
168      var variableRanges = (ProblemData as RegressionProblemData)?.VariableRangesParameter.Value.VariableIntervals;
169
170      if (variableRanges != null) {
171        intervalEvaluation.Add($"Target {ProblemData.TargetVariable}", new Interval(variableRanges[ProblemData.TargetVariable].LowerBound, variableRanges[ProblemData.TargetVariable].UpperBound));
172        intervalEvaluation.Add("Modell Interval", interpreter.GetSymbolicExressionTreeInterval(Model.SymbolicExpressionTree, variableRanges));
173
174        foreach (var derivate in variableRanges) {
175          if (derivate.Key != ProblemData.TargetVariable) {
176            var derived = DerivativeCalculator.Derive(Model.SymbolicExpressionTree, derivate.Key);
177            var derivedResultInterval = interpreter.GetSymbolicExressionTreeInterval(derived, variableRanges);
178            intervalEvaluation.Add(" ∂f/∂" + derivate.Key,
179              new Interval(derivedResultInterval.LowerBound, derivedResultInterval.UpperBound));
180          }
181        }
182      }
183    }
184   
185    private void CalculateResults() {
186      ModelLength = Model.SymbolicExpressionTree.Length;
187      ModelDepth = Model.SymbolicExpressionTree.Depth;
188
189      EstimationLimits.Lower = Model.LowerEstimationLimit;
190      EstimationLimits.Upper = Model.UpperEstimationLimit;
191
192      TrainingUpperEstimationLimitHits = EstimatedTrainingValues.Count(x => x.IsAlmost(Model.UpperEstimationLimit));
193      TestUpperEstimationLimitHits = EstimatedTestValues.Count(x => x.IsAlmost(Model.UpperEstimationLimit));
194      TrainingLowerEstimationLimitHits = EstimatedTrainingValues.Count(x => x.IsAlmost(Model.LowerEstimationLimit));
195      TestLowerEstimationLimitHits = EstimatedTestValues.Count(x => x.IsAlmost(Model.LowerEstimationLimit));
196      TrainingNaNEvaluations = Model.Interpreter.GetSymbolicExpressionTreeValues(Model.SymbolicExpressionTree, ProblemData.Dataset, ProblemData.TrainingIndices).Count(double.IsNaN);
197      TestNaNEvaluations = Model.Interpreter.GetSymbolicExpressionTreeValues(Model.SymbolicExpressionTree, ProblemData.Dataset, ProblemData.TestIndices).Count(double.IsNaN);
198    }
199  }
200}
Note: See TracBrowser for help on using the repository browser.