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 |
|
---|
22 | using System.Linq;
|
---|
23 | using HeuristicLab.Common;
|
---|
24 | using HeuristicLab.Core;
|
---|
25 | using HeuristicLab.Data;
|
---|
26 | using HeuristicLab.Operators;
|
---|
27 | using HeuristicLab.Optimization;
|
---|
28 | using HeuristicLab.Parameters;
|
---|
29 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
30 | using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
|
---|
31 | using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic;
|
---|
32 | using HeuristicLab.Problems.DataAnalysis.Symbolic;
|
---|
33 | using System.Collections.Generic;
|
---|
34 | using HeuristicLab.Problems.DataAnalysis.Symbolic.Symbols;
|
---|
35 | using HeuristicLab.Problems.DataAnalysis;
|
---|
36 | using HeuristicLab.Problems.DataAnalysis.Evaluators;
|
---|
37 |
|
---|
38 | namespace HeuristicLab.Problems.DataAnalysis.Regression.Symbolic.Analyzers {
|
---|
39 | [Item("BestSymbolicRegressionSolutionAnalyzer", "An operator for analyzing the best solution of symbolic regression problems given in symbolic expression tree encoding.")]
|
---|
40 | [StorableClass]
|
---|
41 | public sealed class BestSymbolicRegressionSolutionAnalyzer : SingleSuccessorOperator, ISymbolicRegressionAnalyzer {
|
---|
42 | private const string SymbolicExpressionTreeParameterName = "SymbolicExpressionTree";
|
---|
43 | private const string SymbolicExpressionTreeInterpreterParameterName = "SymbolicExpressionTreeInterpreter";
|
---|
44 | private const string ProblemDataParameterName = "ProblemData";
|
---|
45 | private const string QualityParameterName = "Quality";
|
---|
46 | private const string UpperEstimationLimitParameterName = "UpperEstimationLimit";
|
---|
47 | private const string LowerEstimationLimitParameterName = "LowerEstimationLimit";
|
---|
48 | private const string BestSolutionParameterName = "BestSolution";
|
---|
49 | private const string BestSolutionQualityParameterName = "BestSolutionQuality";
|
---|
50 | private const string ResultsParameterName = "Results";
|
---|
51 | private const string BestSolutionResultName = "Best solution (on validiation set)";
|
---|
52 | private const string BestSolutionInputvariableCountResultName = "Variables used by best solution";
|
---|
53 | private const string BestSolutionTrainingRSquared = "Best solution R² (training)";
|
---|
54 | private const string BestSolutionTestRSquared = "Best solution R² (test)";
|
---|
55 | private const string BestSolutionTrainingMse = "Best solution mean squared error (training)";
|
---|
56 | private const string BestSolutionTestMse = "Best solution mean squared error (test)";
|
---|
57 | private const string BestSolutionTrainingRelativeError = "Best solution average relative error (training)";
|
---|
58 | private const string BestSolutionTestRelativeError = "Best solution average relative error (test)";
|
---|
59 |
|
---|
60 | public ScopeTreeLookupParameter<SymbolicExpressionTree> SymbolicExpressionTreeParameter {
|
---|
61 | get { return (ScopeTreeLookupParameter<SymbolicExpressionTree>)Parameters[SymbolicExpressionTreeParameterName]; }
|
---|
62 | }
|
---|
63 | public IValueLookupParameter<ISymbolicExpressionTreeInterpreter> SymbolicExpressionTreeInterpreterParameter {
|
---|
64 | get { return (IValueLookupParameter<ISymbolicExpressionTreeInterpreter>)Parameters[SymbolicExpressionTreeInterpreterParameterName]; }
|
---|
65 | }
|
---|
66 | public IValueLookupParameter<DataAnalysisProblemData> ProblemDataParameter {
|
---|
67 | get { return (IValueLookupParameter<DataAnalysisProblemData>)Parameters[ProblemDataParameterName]; }
|
---|
68 | }
|
---|
69 | public ScopeTreeLookupParameter<DoubleValue> QualityParameter {
|
---|
70 | get { return (ScopeTreeLookupParameter<DoubleValue>)Parameters[QualityParameterName]; }
|
---|
71 | }
|
---|
72 | public IValueLookupParameter<DoubleValue> UpperEstimationLimitParameter {
|
---|
73 | get { return (IValueLookupParameter<DoubleValue>)Parameters[UpperEstimationLimitParameterName]; }
|
---|
74 | }
|
---|
75 | public IValueLookupParameter<DoubleValue> LowerEstimationLimitParameter {
|
---|
76 | get { return (IValueLookupParameter<DoubleValue>)Parameters[LowerEstimationLimitParameterName]; }
|
---|
77 | }
|
---|
78 | public ILookupParameter<SymbolicRegressionSolution> BestSolutionParameter {
|
---|
79 | get { return (ILookupParameter<SymbolicRegressionSolution>)Parameters[BestSolutionParameterName]; }
|
---|
80 | }
|
---|
81 | public ILookupParameter<DoubleValue> BestSolutionQualityParameter {
|
---|
82 | get { return (ILookupParameter<DoubleValue>)Parameters[BestSolutionQualityParameterName]; }
|
---|
83 | }
|
---|
84 | public ILookupParameter<ResultCollection> ResultsParameter {
|
---|
85 | get { return (ILookupParameter<ResultCollection>)Parameters[ResultsParameterName]; }
|
---|
86 | }
|
---|
87 |
|
---|
88 | public BestSymbolicRegressionSolutionAnalyzer()
|
---|
89 | : base() {
|
---|
90 | Parameters.Add(new ScopeTreeLookupParameter<SymbolicExpressionTree>(SymbolicExpressionTreeParameterName, "The symbolic expression trees to analyze."));
|
---|
91 | Parameters.Add(new ValueLookupParameter<ISymbolicExpressionTreeInterpreter>(SymbolicExpressionTreeInterpreterParameterName, "The interpreter that should be used for the analysis of symbolic expression trees."));
|
---|
92 | Parameters.Add(new ValueLookupParameter<DataAnalysisProblemData>(ProblemDataParameterName, "The problem data for which the symbolic expression tree is a solution."));
|
---|
93 | Parameters.Add(new ValueLookupParameter<DoubleValue>(UpperEstimationLimitParameterName, "The upper estimation limit that was set for the evaluation of the symbolic expression trees."));
|
---|
94 | Parameters.Add(new ValueLookupParameter<DoubleValue>(LowerEstimationLimitParameterName, "The lower estimation limit that was set for the evaluation of the symbolic expression trees."));
|
---|
95 | Parameters.Add(new ScopeTreeLookupParameter<DoubleValue>(QualityParameterName, "The qualities of the symbolic regression trees which should be analyzed."));
|
---|
96 | Parameters.Add(new LookupParameter<SymbolicRegressionSolution>(BestSolutionParameterName, "The best symbolic regression solution."));
|
---|
97 | Parameters.Add(new LookupParameter<DoubleValue>(BestSolutionQualityParameterName, "The quality of the best symbolic regression solution."));
|
---|
98 | Parameters.Add(new LookupParameter<ResultCollection>(ResultsParameterName, "The result collection where the best symbolic regression solution should be stored."));
|
---|
99 | }
|
---|
100 |
|
---|
101 | public override IOperation Apply() {
|
---|
102 | ItemArray<DoubleValue> qualities = QualityParameter.ActualValue;
|
---|
103 | ResultCollection results = ResultsParameter.ActualValue;
|
---|
104 | ISymbolicExpressionTreeInterpreter interpreter = SymbolicExpressionTreeInterpreterParameter.ActualValue;
|
---|
105 | ItemArray<SymbolicExpressionTree> expressions = SymbolicExpressionTreeParameter.ActualValue;
|
---|
106 | DataAnalysisProblemData problemData = ProblemDataParameter.ActualValue;
|
---|
107 | DoubleValue upperEstimationLimit = UpperEstimationLimitParameter.ActualValue;
|
---|
108 | DoubleValue lowerEstimationLimit = LowerEstimationLimitParameter.ActualValue;
|
---|
109 |
|
---|
110 | int i = qualities.Select((x, index) => new { index, x.Value }).OrderBy(x => x.Value).First().index;
|
---|
111 |
|
---|
112 | SymbolicRegressionSolution solution = BestSolutionParameter.ActualValue;
|
---|
113 | if (solution == null) {
|
---|
114 | var model = new SymbolicRegressionModel((ISymbolicExpressionTreeInterpreter)interpreter.Clone(), expressions[i], GetInputVariables(expressions[i]));
|
---|
115 | solution = new SymbolicRegressionSolution(problemData, model, lowerEstimationLimit.Value, upperEstimationLimit.Value);
|
---|
116 | BestSolutionParameter.ActualValue = solution;
|
---|
117 | BestSolutionQualityParameter.ActualValue = qualities[i];
|
---|
118 | results.Add(new Result(BestSolutionResultName, solution));
|
---|
119 | results.Add(new Result(BestSolutionInputvariableCountResultName, new IntValue(model.InputVariables.Count())));
|
---|
120 | #region calculate R2,MSE,Rel Error
|
---|
121 | double[] trainingValues = problemData.Dataset.GetVariableValues(
|
---|
122 | problemData.TargetVariable.Value,
|
---|
123 | problemData.TrainingSamplesStart.Value,
|
---|
124 | problemData.TrainingSamplesEnd.Value);
|
---|
125 | double[] testValues = problemData.Dataset.GetVariableValues(
|
---|
126 | problemData.TargetVariable.Value,
|
---|
127 | problemData.TestSamplesStart.Value,
|
---|
128 | problemData.TestSamplesEnd.Value);
|
---|
129 | double trainingR2 = SimpleRSquaredEvaluator.Calculate(trainingValues, solution.EstimatedTrainingValues);
|
---|
130 | double testR2 = SimpleRSquaredEvaluator.Calculate(testValues, solution.EstimatedTestValues);
|
---|
131 | double trainingMse = SimpleMSEEvaluator.Calculate(trainingValues, solution.EstimatedTrainingValues);
|
---|
132 | double testMse = SimpleMSEEvaluator.Calculate(testValues, solution.EstimatedTestValues);
|
---|
133 | double trainingRelError = SimpleMeanAbsolutePercentageErrorEvaluator.Calculate(trainingValues, solution.EstimatedTrainingValues);
|
---|
134 | double testRelError = SimpleMeanAbsolutePercentageErrorEvaluator.Calculate(testValues, solution.EstimatedTestValues);
|
---|
135 | results.Add(new Result(BestSolutionTrainingRSquared, new DoubleValue(trainingR2)));
|
---|
136 | results.Add(new Result(BestSolutionTestRSquared, new DoubleValue(testR2)));
|
---|
137 | results.Add(new Result(BestSolutionTrainingMse, new DoubleValue(trainingMse)));
|
---|
138 | results.Add(new Result(BestSolutionTestMse, new DoubleValue(testMse)));
|
---|
139 | results.Add(new Result(BestSolutionTrainingRelativeError, new DoubleValue(trainingRelError)));
|
---|
140 | results.Add(new Result(BestSolutionTestRelativeError, new DoubleValue(testRelError)));
|
---|
141 | #endregion
|
---|
142 | } else {
|
---|
143 | if (BestSolutionQualityParameter.ActualValue.Value > qualities[i].Value) {
|
---|
144 | var model = new SymbolicRegressionModel((ISymbolicExpressionTreeInterpreter)interpreter.Clone(), expressions[i], GetInputVariables(expressions[i]));
|
---|
145 | solution = new SymbolicRegressionSolution(problemData, model, lowerEstimationLimit.Value, upperEstimationLimit.Value);
|
---|
146 | BestSolutionParameter.ActualValue = solution;
|
---|
147 | BestSolutionQualityParameter.ActualValue = qualities[i];
|
---|
148 | results[BestSolutionResultName].Value = solution;
|
---|
149 | results[BestSolutionInputvariableCountResultName].Value = new IntValue(model.InputVariables.Count());
|
---|
150 | #region update R2,MSE, Rel Error
|
---|
151 | double[] trainingValues = problemData.Dataset.GetVariableValues(
|
---|
152 | problemData.TargetVariable.Value,
|
---|
153 | problemData.TrainingSamplesStart.Value,
|
---|
154 | problemData.TrainingSamplesEnd.Value);
|
---|
155 | double[] testValues = problemData.Dataset.GetVariableValues(
|
---|
156 | problemData.TargetVariable.Value,
|
---|
157 | problemData.TestSamplesStart.Value,
|
---|
158 | problemData.TestSamplesEnd.Value);
|
---|
159 | double trainingR2 = SimpleRSquaredEvaluator.Calculate(trainingValues, solution.EstimatedTrainingValues);
|
---|
160 | double testR2 = SimpleRSquaredEvaluator.Calculate(testValues, solution.EstimatedTestValues);
|
---|
161 | double trainingMse = SimpleMSEEvaluator.Calculate(trainingValues, solution.EstimatedTrainingValues);
|
---|
162 | double testMse = SimpleMSEEvaluator.Calculate(testValues, solution.EstimatedTestValues);
|
---|
163 | double trainingRelError = SimpleMeanAbsolutePercentageErrorEvaluator.Calculate(trainingValues, solution.EstimatedTrainingValues);
|
---|
164 | double testRelError = SimpleMeanAbsolutePercentageErrorEvaluator.Calculate(testValues, solution.EstimatedTestValues);
|
---|
165 | results[BestSolutionTrainingRSquared].Value = new DoubleValue(trainingR2);
|
---|
166 | results[BestSolutionTestRSquared].Value = new DoubleValue(testR2);
|
---|
167 | results[BestSolutionTrainingMse].Value = new DoubleValue(trainingMse);
|
---|
168 | results[BestSolutionTestMse].Value = new DoubleValue(testMse);
|
---|
169 | results[BestSolutionTrainingRelativeError].Value = new DoubleValue(trainingRelError);
|
---|
170 | results[BestSolutionTestRelativeError].Value = new DoubleValue(testRelError);
|
---|
171 | #endregion
|
---|
172 | }
|
---|
173 | }
|
---|
174 |
|
---|
175 | return base.Apply();
|
---|
176 | }
|
---|
177 |
|
---|
178 | private IEnumerable<string> GetInputVariables(SymbolicExpressionTree tree) {
|
---|
179 | return (from varNode in tree.IterateNodesPrefix().OfType<VariableTreeNode>()
|
---|
180 | select varNode.VariableName).Distinct();
|
---|
181 | }
|
---|
182 | }
|
---|
183 | }
|
---|