| 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.Evaluators;
|
|---|
| 32 | using HeuristicLab.Problems.DataAnalysis.Symbolic;
|
|---|
| 33 |
|
|---|
| 34 | namespace HeuristicLab.Problems.DataAnalysis.Regression.Symbolic {
|
|---|
| 35 | /// <summary>
|
|---|
| 36 | /// An operator for visualizing the best symbolic regression solution based on the validation set.
|
|---|
| 37 | /// </summary>
|
|---|
| 38 | [Item("BestSymbolicExpressionTreeVisualizer", "An operator for visualizing the best symbolic regression solution based on the validation set.")]
|
|---|
| 39 | [StorableClass]
|
|---|
| 40 | public sealed class BestValidationSymbolicRegressionSolutionVisualizer : SingleSuccessorOperator, ISingleObjectiveSolutionsVisualizer, ISolutionsVisualizer {
|
|---|
| 41 | private const string SymbolicExpressionTreeInterpreterParameterName = "SymbolicExpressionTreeInterpreter";
|
|---|
| 42 | private const string UpperEstimationLimitParameterName = "UpperEstimationLimit";
|
|---|
| 43 | private const string LowerEstimationLimitParameterName = "LowerEstimationLimit";
|
|---|
| 44 | private const string SymbolicRegressionModelParameterName = "SymbolicRegressionModel";
|
|---|
| 45 | private const string DataAnalysisProblemDataParameterName = "DataAnalysisProblemData";
|
|---|
| 46 | private const string BestValidationSolutionParameterName = "BestValidationSolution";
|
|---|
| 47 | private const string ValidationSamplesStartParameterName = "ValidationSamplesStart";
|
|---|
| 48 | private const string ValidationSamplesEndParameterName = "ValidationSamplesEnd";
|
|---|
| 49 | private const string QualityParameterName = "Quality";
|
|---|
| 50 | private const string ResultsParameterName = "Results";
|
|---|
| 51 |
|
|---|
| 52 | #region parameter properties
|
|---|
| 53 | public ILookupParameter<ISymbolicExpressionTreeInterpreter> SymbolicExpressionTreeInterpreterParameter {
|
|---|
| 54 | get { return (ILookupParameter<ISymbolicExpressionTreeInterpreter>)Parameters[SymbolicExpressionTreeInterpreterParameterName]; }
|
|---|
| 55 | }
|
|---|
| 56 | public IValueLookupParameter<DoubleValue> UpperEstimationLimitParameter {
|
|---|
| 57 | get { return (IValueLookupParameter<DoubleValue>)Parameters[UpperEstimationLimitParameterName]; }
|
|---|
| 58 | }
|
|---|
| 59 | public IValueLookupParameter<DoubleValue> LowerEstimationLimitParameter {
|
|---|
| 60 | get { return (IValueLookupParameter<DoubleValue>)Parameters[LowerEstimationLimitParameterName]; }
|
|---|
| 61 | }
|
|---|
| 62 | public IValueLookupParameter<IntValue> ValidationSamplesStartParameter {
|
|---|
| 63 | get { return (IValueLookupParameter<IntValue>)Parameters[ValidationSamplesStartParameterName]; }
|
|---|
| 64 | }
|
|---|
| 65 | public IValueLookupParameter<IntValue> ValidationSamplesEndParameter {
|
|---|
| 66 | get { return (IValueLookupParameter<IntValue>)Parameters[ValidationSamplesEndParameterName]; }
|
|---|
| 67 | }
|
|---|
| 68 |
|
|---|
| 69 | public ILookupParameter<ItemArray<SymbolicExpressionTree>> SymbolicExpressionTreeParameter {
|
|---|
| 70 | get { return (ILookupParameter<ItemArray<SymbolicExpressionTree>>)Parameters[SymbolicRegressionModelParameterName]; }
|
|---|
| 71 | }
|
|---|
| 72 | public ILookupParameter<DataAnalysisProblemData> DataAnalysisProblemDataParameter {
|
|---|
| 73 | get { return (ILookupParameter<DataAnalysisProblemData>)Parameters[DataAnalysisProblemDataParameterName]; }
|
|---|
| 74 | }
|
|---|
| 75 | public ILookupParameter<SymbolicRegressionSolution> BestValidationSolutionParameter {
|
|---|
| 76 | get { return (ILookupParameter<SymbolicRegressionSolution>)Parameters[BestValidationSolutionParameterName]; }
|
|---|
| 77 | }
|
|---|
| 78 | ILookupParameter ISolutionsVisualizer.VisualizationParameter {
|
|---|
| 79 | get { return BestValidationSolutionParameter; }
|
|---|
| 80 | }
|
|---|
| 81 |
|
|---|
| 82 | public ILookupParameter<ItemArray<DoubleValue>> QualityParameter {
|
|---|
| 83 | get { return (ILookupParameter<ItemArray<DoubleValue>>)Parameters[QualityParameterName]; }
|
|---|
| 84 | }
|
|---|
| 85 |
|
|---|
| 86 | public ILookupParameter<ResultCollection> ResultParameter {
|
|---|
| 87 | get { return (ILookupParameter<ResultCollection>)Parameters[ResultsParameterName]; }
|
|---|
| 88 | }
|
|---|
| 89 | #endregion
|
|---|
| 90 |
|
|---|
| 91 | #region properties
|
|---|
| 92 | public ISymbolicExpressionTreeInterpreter SymbolicExpressionTreeInterpreter {
|
|---|
| 93 | get { return SymbolicExpressionTreeInterpreterParameter.ActualValue; }
|
|---|
| 94 | }
|
|---|
| 95 | public DoubleValue UpperEstimationLimit {
|
|---|
| 96 | get { return UpperEstimationLimitParameter.ActualValue; }
|
|---|
| 97 | }
|
|---|
| 98 | public DoubleValue LowerEstimationLimit {
|
|---|
| 99 | get { return LowerEstimationLimitParameter.ActualValue; }
|
|---|
| 100 | }
|
|---|
| 101 | public IntValue ValidationSamplesStart {
|
|---|
| 102 | get { return ValidationSamplesStartParameter.ActualValue; }
|
|---|
| 103 | }
|
|---|
| 104 | public IntValue ValidationSamplesEnd {
|
|---|
| 105 | get { return ValidationSamplesEndParameter.ActualValue; }
|
|---|
| 106 | }
|
|---|
| 107 | #endregion
|
|---|
| 108 |
|
|---|
| 109 | public BestValidationSymbolicRegressionSolutionVisualizer()
|
|---|
| 110 | : base() {
|
|---|
| 111 | Parameters.Add(new SubScopesLookupParameter<SymbolicExpressionTree>(SymbolicRegressionModelParameterName, "The symbolic regression solutions from which the best solution should be visualized."));
|
|---|
| 112 | Parameters.Add(new SubScopesLookupParameter<DoubleValue>(QualityParameterName, "The quality of the symbolic regression solutions."));
|
|---|
| 113 | Parameters.Add(new LookupParameter<DataAnalysisProblemData>(DataAnalysisProblemDataParameterName, "The symbolic regression problme data on which the best solution should be evaluated."));
|
|---|
| 114 | Parameters.Add(new LookupParameter<ISymbolicExpressionTreeInterpreter>(SymbolicExpressionTreeInterpreterParameterName, "The interpreter that should be used to calculate the output values of symbolic expression trees."));
|
|---|
| 115 | Parameters.Add(new ValueLookupParameter<DoubleValue>(UpperEstimationLimitParameterName, "The upper limit that should be used as cut off value for the output values of symbolic expression trees."));
|
|---|
| 116 | Parameters.Add(new ValueLookupParameter<DoubleValue>(LowerEstimationLimitParameterName, "The lower limit that should be used as cut off value for the output values of symbolic expression trees."));
|
|---|
| 117 | Parameters.Add(new ValueLookupParameter<IntValue>(ValidationSamplesStartParameterName, "The start index of the validation partition (part of the training partition)."));
|
|---|
| 118 | Parameters.Add(new ValueLookupParameter<IntValue>(ValidationSamplesEndParameterName, "The end index of the validation partition (part of the training partition)."));
|
|---|
| 119 | Parameters.Add(new LookupParameter<SymbolicRegressionSolution>(BestValidationSolutionParameterName, "The best symbolic expression tree based on the validation data for the symbolic regression problem."));
|
|---|
| 120 | Parameters.Add(new LookupParameter<ResultCollection>(ResultsParameterName, "The result collection of the algorithm."));
|
|---|
| 121 | }
|
|---|
| 122 |
|
|---|
| 123 | public override IOperation Apply() {
|
|---|
| 124 | ItemArray<SymbolicExpressionTree> expressions = SymbolicExpressionTreeParameter.ActualValue;
|
|---|
| 125 | DataAnalysisProblemData problemData = DataAnalysisProblemDataParameter.ActualValue;
|
|---|
| 126 |
|
|---|
| 127 | int validationSamplesStart = ValidationSamplesStart.Value;
|
|---|
| 128 | int validationSamplesEnd = ValidationSamplesEnd.Value;
|
|---|
| 129 | var validationValues = problemData.Dataset.GetVariableValues(problemData.TargetVariable.Value, validationSamplesStart, validationSamplesEnd);
|
|---|
| 130 | double upperEstimationLimit = UpperEstimationLimit.Value;
|
|---|
| 131 | double lowerEstimationLimit = LowerEstimationLimit.Value;
|
|---|
| 132 | var currentBestExpression = (from expression in expressions
|
|---|
| 133 | let validationQuality =
|
|---|
| 134 | SymbolicRegressionMeanSquaredErrorEvaluator.Calculate(
|
|---|
| 135 | SymbolicExpressionTreeInterpreter, expression,
|
|---|
| 136 | lowerEstimationLimit, upperEstimationLimit,
|
|---|
| 137 | problemData.Dataset, problemData.TargetVariable.Value,
|
|---|
| 138 | validationSamplesStart, validationSamplesEnd)
|
|---|
| 139 | select new { Expression = expression, ValidationQuality = validationQuality })
|
|---|
| 140 | .OrderBy(x => x.ValidationQuality)
|
|---|
| 141 | .First();
|
|---|
| 142 |
|
|---|
| 143 | SymbolicRegressionSolution bestOfRunSolution = BestValidationSolutionParameter.ActualValue;
|
|---|
| 144 | if (bestOfRunSolution == null) {
|
|---|
| 145 | // no best of run solution yet -> make a solution from the currentBestExpression
|
|---|
| 146 | UpdateBestOfRunSolution(problemData, currentBestExpression.Expression, SymbolicExpressionTreeInterpreter, lowerEstimationLimit, upperEstimationLimit);
|
|---|
| 147 | } else {
|
|---|
| 148 | // compare quality of current best with best of run solution
|
|---|
| 149 | var estimatedValidationValues = bestOfRunSolution.EstimatedValues.Skip(validationSamplesStart).Take(validationSamplesEnd - validationSamplesStart);
|
|---|
| 150 | var bestOfRunValidationQuality = SimpleMSEEvaluator.Calculate(validationValues, estimatedValidationValues);
|
|---|
| 151 | if (bestOfRunValidationQuality > currentBestExpression.ValidationQuality) {
|
|---|
| 152 | UpdateBestOfRunSolution(problemData, currentBestExpression.Expression, SymbolicExpressionTreeInterpreter, lowerEstimationLimit, upperEstimationLimit);
|
|---|
| 153 | }
|
|---|
| 154 | }
|
|---|
| 155 |
|
|---|
| 156 |
|
|---|
| 157 | return base.Apply();
|
|---|
| 158 | }
|
|---|
| 159 |
|
|---|
| 160 | private void UpdateBestOfRunSolution(DataAnalysisProblemData problemData, SymbolicExpressionTree tree, ISymbolicExpressionTreeInterpreter interpreter,
|
|---|
| 161 | double lowerEstimationLimit, double upperEstimationLimit) {
|
|---|
| 162 | var newBestSolution = CreateDataAnalysisSolution(problemData, tree, interpreter, lowerEstimationLimit, upperEstimationLimit);
|
|---|
| 163 | if (BestValidationSolutionParameter.ActualValue == null)
|
|---|
| 164 | BestValidationSolutionParameter.ActualValue = newBestSolution;
|
|---|
| 165 | else
|
|---|
| 166 | // only update model
|
|---|
| 167 | BestValidationSolutionParameter.ActualValue.Model = newBestSolution.Model;
|
|---|
| 168 |
|
|---|
| 169 | var trainingValues = problemData.Dataset.GetVariableValues(problemData.TargetVariable.Value, problemData.TrainingSamplesStart.Value, problemData.TrainingSamplesEnd.Value);
|
|---|
| 170 | var testValues = problemData.Dataset.GetVariableValues(problemData.TargetVariable.Value, problemData.TestSamplesStart.Value, problemData.TestSamplesEnd.Value);
|
|---|
| 171 |
|
|---|
| 172 | AddResult("MeanSquaredError (Training)", new DoubleValue(SimpleMSEEvaluator.Calculate(trainingValues, newBestSolution.EstimatedTrainingValues)));
|
|---|
| 173 | AddResult("MeanRelativeError (Training)", new PercentValue(SimpleMeanAbsolutePercentageErrorEvaluator.Calculate(trainingValues, newBestSolution.EstimatedTrainingValues)));
|
|---|
| 174 | AddResult("RSquared (Training)", new DoubleValue(SimpleRSquaredEvaluator.Calculate(trainingValues, newBestSolution.EstimatedTrainingValues)));
|
|---|
| 175 |
|
|---|
| 176 | AddResult("MeanSquaredError (Test)", new DoubleValue(SimpleMSEEvaluator.Calculate(testValues, newBestSolution.EstimatedTestValues)));
|
|---|
| 177 | AddResult("MeanRelativeError (Test)", new PercentValue(SimpleMeanAbsolutePercentageErrorEvaluator.Calculate(testValues, newBestSolution.EstimatedTestValues)));
|
|---|
| 178 | AddResult("RSquared (Test)", new DoubleValue(SimpleRSquaredEvaluator.Calculate(testValues, newBestSolution.EstimatedTestValues)));
|
|---|
| 179 | }
|
|---|
| 180 |
|
|---|
| 181 | private void AddResult(string resultName, IItem value) {
|
|---|
| 182 | var resultCollection = ResultParameter.ActualValue;
|
|---|
| 183 | if (resultCollection.ContainsKey(resultName)) {
|
|---|
| 184 | resultCollection[resultName].Value = value;
|
|---|
| 185 | } else {
|
|---|
| 186 | resultCollection.Add(new Result(resultName, value));
|
|---|
| 187 | }
|
|---|
| 188 | }
|
|---|
| 189 |
|
|---|
| 190 | private SymbolicRegressionSolution CreateDataAnalysisSolution(DataAnalysisProblemData problemData, SymbolicExpressionTree expression, ISymbolicExpressionTreeInterpreter interpreter,
|
|---|
| 191 | double lowerEstimationLimit, double upperEstimationLimit) {
|
|---|
| 192 | var model = new SymbolicRegressionModel(interpreter, expression, problemData.InputVariables.Select(s => s.Value));
|
|---|
| 193 | return new SymbolicRegressionSolution(problemData, model, lowerEstimationLimit, upperEstimationLimit);
|
|---|
| 194 | }
|
|---|
| 195 | }
|
|---|
| 196 | }
|
|---|
