| 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 System;
|
|---|
| 37 |
|
|---|
| 38 | using HeuristicLab.Problems.DataAnalysis.Evaluators;
|
|---|
| 39 | using HeuristicLab.Problems.DataAnalysis.Regression;
|
|---|
| 40 | using HeuristicLab.Analysis;
|
|---|
| 41 | using HeuristicLab.Problems.DataAnalysis.MultiVariate.TimeSeriesPrognosis.Symbolic.Evaluators;
|
|---|
| 42 | using HeuristicLab.Problems.DataAnalysis.MultiVariate.Regression.Symbolic;
|
|---|
| 43 | using HeuristicLab.Problems.DataAnalysis.MultiVariate.TimeSeriesPrognosis.Symbolic.Interfaces;
|
|---|
| 44 |
|
|---|
| 45 | namespace HeuristicLab.Problems.DataAnalysis.MultiVariate.TimeSeriesPrognosis.Symbolic.Analyzers {
|
|---|
| 46 | /// <summary>
|
|---|
| 47 | /// An operator that analyzes the validation best scaled symbolic time series prognosis solution.
|
|---|
| 48 | /// </summary>
|
|---|
| 49 | [Item("ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer", "An operator that analyzes the validation best scaled symbolic time series prognosis solution.")]
|
|---|
| 50 | [StorableClass]
|
|---|
| 51 | public sealed class ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer : SingleSuccessorOperator, IAnalyzer {
|
|---|
| 52 | private const string RandomParameterName = "Random";
|
|---|
| 53 | private const string SymbolicExpressionTreeParameterName = "SymbolicExpressionTree";
|
|---|
| 54 | private const string SymbolicExpressionTreeInterpreterParameterName = "SymbolicExpressionTreeInterpreter";
|
|---|
| 55 | private const string ProblemDataParameterName = "ProblemData";
|
|---|
| 56 | private const string ValidationSamplesStartParameterName = "SamplesStart";
|
|---|
| 57 | private const string ValidationSamplesEndParameterName = "SamplesEnd";
|
|---|
| 58 | private const string QualityParameterName = "Quality";
|
|---|
| 59 | private const string ScaledQualityParameterName = "ScaledQuality";
|
|---|
| 60 | private const string UpperEstimationLimitParameterName = "UpperEstimationLimit";
|
|---|
| 61 | private const string LowerEstimationLimitParameterName = "LowerEstimationLimit";
|
|---|
| 62 | private const string PredictionHorizonParameterName = "PredictionHorizon";
|
|---|
| 63 | private const string ConditionVariableParameterName = "ConditionVariableName";
|
|---|
| 64 | private const string AlphaParameterName = "Alpha";
|
|---|
| 65 | private const string BetaParameterName = "Beta";
|
|---|
| 66 | private const string BestSolutionParameterName = "Best solution (validation)";
|
|---|
| 67 | private const string BestSolutionQualityParameterName = "Best solution quality (validation)";
|
|---|
| 68 | private const string CurrentBestValidationQualityParameterName = "Current best validation quality";
|
|---|
| 69 | private const string BestSolutionQualityValuesParameterName = "Validation Quality";
|
|---|
| 70 | private const string ResultsParameterName = "Results";
|
|---|
| 71 | private const string VariableFrequenciesParameterName = "VariableFrequencies";
|
|---|
| 72 | private const string BestKnownQualityParameterName = "BestKnownQuality";
|
|---|
| 73 | private const string GenerationsParameterName = "Generations";
|
|---|
| 74 |
|
|---|
| 75 | private const string TrainingMeanSquaredErrorQualityParameterName = "Mean squared error (training)";
|
|---|
| 76 | private const string MinTrainingMeanSquaredErrorQualityParameterName = "Min mean squared error (training)";
|
|---|
| 77 | private const string MaxTrainingMeanSquaredErrorQualityParameterName = "Max mean squared error (training)";
|
|---|
| 78 | private const string AverageTrainingMeanSquaredErrorQualityParameterName = "Average mean squared error (training)";
|
|---|
| 79 | private const string BestTrainingMeanSquaredErrorQualityParameterName = "Best mean squared error (training)";
|
|---|
| 80 |
|
|---|
| 81 | private const string TrainingAverageRelativeErrorQualityParameterName = "Average relative error (training)";
|
|---|
| 82 | private const string MinTrainingAverageRelativeErrorQualityParameterName = "Min average relative error (training)";
|
|---|
| 83 | private const string MaxTrainingAverageRelativeErrorQualityParameterName = "Max average relative error (training)";
|
|---|
| 84 | private const string AverageTrainingAverageRelativeErrorQualityParameterName = "Average average relative error (training)";
|
|---|
| 85 | private const string BestTrainingAverageRelativeErrorQualityParameterName = "Best average relative error (training)";
|
|---|
| 86 |
|
|---|
| 87 | private const string TrainingRSquaredQualityParameterName = "R² (training)";
|
|---|
| 88 | private const string MinTrainingRSquaredQualityParameterName = "Min R² (training)";
|
|---|
| 89 | private const string MaxTrainingRSquaredQualityParameterName = "Max R² (training)";
|
|---|
| 90 | private const string AverageTrainingRSquaredQualityParameterName = "Average R² (training)";
|
|---|
| 91 | private const string BestTrainingRSquaredQualityParameterName = "Best R² (training)";
|
|---|
| 92 |
|
|---|
| 93 | private const string TestMeanSquaredErrorQualityParameterName = "Mean squared error (test)";
|
|---|
| 94 | private const string MinTestMeanSquaredErrorQualityParameterName = "Min mean squared error (test)";
|
|---|
| 95 | private const string MaxTestMeanSquaredErrorQualityParameterName = "Max mean squared error (test)";
|
|---|
| 96 | private const string AverageTestMeanSquaredErrorQualityParameterName = "Average mean squared error (test)";
|
|---|
| 97 | private const string BestTestMeanSquaredErrorQualityParameterName = "Best mean squared error (test)";
|
|---|
| 98 |
|
|---|
| 99 | private const string TestAverageRelativeErrorQualityParameterName = "Average relative error (test)";
|
|---|
| 100 | private const string MinTestAverageRelativeErrorQualityParameterName = "Min average relative error (test)";
|
|---|
| 101 | private const string MaxTestAverageRelativeErrorQualityParameterName = "Max average relative error (test)";
|
|---|
| 102 | private const string AverageTestAverageRelativeErrorQualityParameterName = "Average average relative error (test)";
|
|---|
| 103 | private const string BestTestAverageRelativeErrorQualityParameterName = "Best average relative error (test)";
|
|---|
| 104 |
|
|---|
| 105 | private const string TestRSquaredQualityParameterName = "R² (test)";
|
|---|
| 106 | private const string MinTestRSquaredQualityParameterName = "Min R² (test)";
|
|---|
| 107 | private const string MaxTestRSquaredQualityParameterName = "Max R² (test)";
|
|---|
| 108 | private const string AverageTestRSquaredQualityParameterName = "Average R² (test)";
|
|---|
| 109 | private const string BestTestRSquaredQualityParameterName = "Best R² (test)";
|
|---|
| 110 |
|
|---|
| 111 | private const string RSquaredValuesParameterName = "R²";
|
|---|
| 112 | private const string MeanSquaredErrorValuesParameterName = "Mean squared error";
|
|---|
| 113 | private const string RelativeErrorValuesParameterName = "Average relative error";
|
|---|
| 114 | private const string BestSolutionResultName = "Best solution (on validiation set)";
|
|---|
| 115 | private const string RelativeNumberOfEvaluatedSamplesParameterName = "RelativeNumberOfEvaluatedSamples";
|
|---|
| 116 |
|
|---|
| 117 | #region parameter properties
|
|---|
| 118 | public ILookupParameter<IRandom> RandomParameter {
|
|---|
| 119 | get { return (ILookupParameter<IRandom>)Parameters[RandomParameterName]; }
|
|---|
| 120 | }
|
|---|
| 121 | public ScopeTreeLookupParameter<SymbolicExpressionTree> SymbolicExpressionTreeParameter {
|
|---|
| 122 | get { return (ScopeTreeLookupParameter<SymbolicExpressionTree>)Parameters[SymbolicExpressionTreeParameterName]; }
|
|---|
| 123 | }
|
|---|
| 124 | public OptionalValueParameter<StringValue> ConditionVariableNameParameter {
|
|---|
| 125 | get { return (OptionalValueParameter<StringValue>)Parameters[ConditionVariableParameterName]; }
|
|---|
| 126 | }
|
|---|
| 127 | public IValueLookupParameter<ISymbolicTimeSeriesExpressionInterpreter> SymbolicExpressionTreeInterpreterParameter {
|
|---|
| 128 | get { return (IValueLookupParameter<ISymbolicTimeSeriesExpressionInterpreter>)Parameters[SymbolicExpressionTreeInterpreterParameterName]; }
|
|---|
| 129 | }
|
|---|
| 130 | public IValueLookupParameter<MultiVariateDataAnalysisProblemData> ProblemDataParameter {
|
|---|
| 131 | get { return (IValueLookupParameter<MultiVariateDataAnalysisProblemData>)Parameters[ProblemDataParameterName]; }
|
|---|
| 132 | }
|
|---|
| 133 | public ILookupParameter<ISingleObjectiveSymbolicTimeSeriesPrognosisEvaluator> EvaluatorParameter {
|
|---|
| 134 | get { return (ILookupParameter<ISingleObjectiveSymbolicTimeSeriesPrognosisEvaluator>)Parameters["Evaluator"]; }
|
|---|
| 135 | }
|
|---|
| 136 | public IValueLookupParameter<IntValue> ValidationSamplesStartParameter {
|
|---|
| 137 | get { return (IValueLookupParameter<IntValue>)Parameters[ValidationSamplesStartParameterName]; }
|
|---|
| 138 | }
|
|---|
| 139 | public IValueLookupParameter<IntValue> ValidationSamplesEndParameter {
|
|---|
| 140 | get { return (IValueLookupParameter<IntValue>)Parameters[ValidationSamplesEndParameterName]; }
|
|---|
| 141 | }
|
|---|
| 142 | public IValueLookupParameter<DoubleArray> UpperEstimationLimitParameter {
|
|---|
| 143 | get { return (IValueLookupParameter<DoubleArray>)Parameters[UpperEstimationLimitParameterName]; }
|
|---|
| 144 | }
|
|---|
| 145 | public IValueLookupParameter<DoubleArray> LowerEstimationLimitParameter {
|
|---|
| 146 | get { return (IValueLookupParameter<DoubleArray>)Parameters[LowerEstimationLimitParameterName]; }
|
|---|
| 147 | }
|
|---|
| 148 | public IValueLookupParameter<IntValue> PredictionHorizonParameter {
|
|---|
| 149 | get { return (IValueLookupParameter<IntValue>)Parameters[PredictionHorizonParameterName]; }
|
|---|
| 150 | }
|
|---|
| 151 | public ILookupParameter<SymbolicTimeSeriesPrognosisSolution> BestSolutionParameter {
|
|---|
| 152 | get { return (ILookupParameter<SymbolicTimeSeriesPrognosisSolution>)Parameters[BestSolutionParameterName]; }
|
|---|
| 153 | }
|
|---|
| 154 | public ILookupParameter<IntValue> GenerationsParameter {
|
|---|
| 155 | get { return (ILookupParameter<IntValue>)Parameters[GenerationsParameterName]; }
|
|---|
| 156 | }
|
|---|
| 157 | public ILookupParameter<DoubleValue> BestSolutionQualityParameter {
|
|---|
| 158 | get { return (ILookupParameter<DoubleValue>)Parameters[BestSolutionQualityParameterName]; }
|
|---|
| 159 | }
|
|---|
| 160 | public ILookupParameter<ResultCollection> ResultsParameter {
|
|---|
| 161 | get { return (ILookupParameter<ResultCollection>)Parameters[ResultsParameterName]; }
|
|---|
| 162 | }
|
|---|
| 163 | public ILookupParameter<DoubleValue> BestKnownQualityParameter {
|
|---|
| 164 | get { return (ILookupParameter<DoubleValue>)Parameters[BestKnownQualityParameterName]; }
|
|---|
| 165 | }
|
|---|
| 166 | public ILookupParameter<DataTable> VariableFrequenciesParameter {
|
|---|
| 167 | get { return (ILookupParameter<DataTable>)Parameters[VariableFrequenciesParameterName]; }
|
|---|
| 168 | }
|
|---|
| 169 | public IValueParameter<PercentValue> RelativeNumberOfEvaluatedSamplesParameter {
|
|---|
| 170 | get { return (IValueParameter<PercentValue>)Parameters[RelativeNumberOfEvaluatedSamplesParameterName]; }
|
|---|
| 171 | }
|
|---|
| 172 | public ILookupParameter<BoolValue> MaximizationParameter {
|
|---|
| 173 | get { return (ILookupParameter<BoolValue>)Parameters["Maximization"]; }
|
|---|
| 174 | }
|
|---|
| 175 | #endregion
|
|---|
| 176 | #region properties
|
|---|
| 177 | public IRandom Random {
|
|---|
| 178 | get { return RandomParameter.ActualValue; }
|
|---|
| 179 | }
|
|---|
| 180 | public ItemArray<SymbolicExpressionTree> SymbolicExpressionTree {
|
|---|
| 181 | get { return SymbolicExpressionTreeParameter.ActualValue; }
|
|---|
| 182 | }
|
|---|
| 183 | public ISymbolicTimeSeriesExpressionInterpreter SymbolicExpressionTreeInterpreter {
|
|---|
| 184 | get { return SymbolicExpressionTreeInterpreterParameter.ActualValue; }
|
|---|
| 185 | }
|
|---|
| 186 | public MultiVariateDataAnalysisProblemData ProblemData {
|
|---|
| 187 | get { return ProblemDataParameter.ActualValue; }
|
|---|
| 188 | }
|
|---|
| 189 | public IntValue ValidiationSamplesStart {
|
|---|
| 190 | get { return ValidationSamplesStartParameter.ActualValue; }
|
|---|
| 191 | }
|
|---|
| 192 | public IntValue ValidationSamplesEnd {
|
|---|
| 193 | get { return ValidationSamplesEndParameter.ActualValue; }
|
|---|
| 194 | }
|
|---|
| 195 | public DoubleArray UpperEstimationLimit {
|
|---|
| 196 | get { return UpperEstimationLimitParameter.ActualValue; }
|
|---|
| 197 | }
|
|---|
| 198 | public DoubleArray LowerEstimationLimit {
|
|---|
| 199 | get { return LowerEstimationLimitParameter.ActualValue; }
|
|---|
| 200 | }
|
|---|
| 201 | public IntValue PredictionHorizon {
|
|---|
| 202 | get { return PredictionHorizonParameter.ActualValue; }
|
|---|
| 203 | }
|
|---|
| 204 | public StringValue ConditionVariableName {
|
|---|
| 205 | get { return ConditionVariableNameParameter.Value; }
|
|---|
| 206 | }
|
|---|
| 207 | public ResultCollection Results {
|
|---|
| 208 | get { return ResultsParameter.ActualValue; }
|
|---|
| 209 | }
|
|---|
| 210 | public DataTable VariableFrequencies {
|
|---|
| 211 | get { return VariableFrequenciesParameter.ActualValue; }
|
|---|
| 212 | }
|
|---|
| 213 | public IntValue Generations {
|
|---|
| 214 | get { return GenerationsParameter.ActualValue; }
|
|---|
| 215 | }
|
|---|
| 216 | public PercentValue RelativeNumberOfEvaluatedSamples {
|
|---|
| 217 | get { return RelativeNumberOfEvaluatedSamplesParameter.Value; }
|
|---|
| 218 | }
|
|---|
| 219 | public BoolValue Maximization {
|
|---|
| 220 | get { return MaximizationParameter.ActualValue; }
|
|---|
| 221 | }
|
|---|
| 222 | #endregion
|
|---|
| 223 |
|
|---|
| 224 | [StorableConstructor]
|
|---|
| 225 | protected ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer(bool deserializing) : base(deserializing) { }
|
|---|
| 226 | protected ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer(ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer original, Cloner cloner)
|
|---|
| 227 | : base(original, cloner) {
|
|---|
| 228 | }
|
|---|
| 229 | public ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer()
|
|---|
| 230 | : base() {
|
|---|
| 231 | Parameters.Add(new LookupParameter<IRandom>(RandomParameterName, "A random number generator."));
|
|---|
| 232 | Parameters.Add(new ScopeTreeLookupParameter<SymbolicExpressionTree>(SymbolicExpressionTreeParameterName, "The symbolic expression trees to analyze."));
|
|---|
| 233 | Parameters.Add(new OptionalValueParameter<StringValue>(ConditionVariableParameterName, "The name of the condition variable indicating if a row should be considered for evaluation or not."));
|
|---|
| 234 | Parameters.Add(new ValueLookupParameter<ISymbolicTimeSeriesExpressionInterpreter>(SymbolicExpressionTreeInterpreterParameterName, "The interpreter that should be used for the analysis of symbolic expression trees."));
|
|---|
| 235 | Parameters.Add(new ValueLookupParameter<MultiVariateDataAnalysisProblemData>(ProblemDataParameterName, "The problem data for which the symbolic expression tree is a solution."));
|
|---|
| 236 | Parameters.Add(new ValueLookupParameter<IntValue>(ValidationSamplesStartParameterName, "The first index of the validation partition of the data set."));
|
|---|
| 237 | Parameters.Add(new ValueLookupParameter<IntValue>(ValidationSamplesEndParameterName, "The last index of the validation partition of the data set."));
|
|---|
| 238 | Parameters.Add(new ValueLookupParameter<IntValue>(PredictionHorizonParameterName, "The number of time steps for which to create a forecast."));
|
|---|
| 239 | Parameters.Add(new LookupParameter<ISingleObjectiveSymbolicTimeSeriesPrognosisEvaluator>("Evaluator", ""));
|
|---|
| 240 | Parameters.Add(new ValueLookupParameter<DoubleArray>(UpperEstimationLimitParameterName, "The upper estimation limit that was set for the evaluation of the symbolic expression trees."));
|
|---|
| 241 | Parameters.Add(new ValueLookupParameter<DoubleArray>(LowerEstimationLimitParameterName, "The lower estimation limit that was set for the evaluation of the symbolic expression trees."));
|
|---|
| 242 | Parameters.Add(new LookupParameter<SymbolicTimeSeriesPrognosisSolution>(BestSolutionParameterName, "The best symbolic time series prognosis solution."));
|
|---|
| 243 | Parameters.Add(new LookupParameter<IntValue>(GenerationsParameterName, "The number of generations calculated so far."));
|
|---|
| 244 | Parameters.Add(new LookupParameter<DoubleValue>(BestSolutionQualityParameterName, "The quality of the best symbolic regression solution."));
|
|---|
| 245 | Parameters.Add(new LookupParameter<ResultCollection>(ResultsParameterName, "The result collection where the best symbolic regression solution should be stored."));
|
|---|
| 246 | Parameters.Add(new LookupParameter<DoubleValue>(BestKnownQualityParameterName, "The best known (validation) quality achieved on the data set."));
|
|---|
| 247 | Parameters.Add(new LookupParameter<DataTable>(VariableFrequenciesParameterName, "The variable frequencies table to use for the calculation of variable impacts"));
|
|---|
| 248 | Parameters.Add(new ValueParameter<PercentValue>(RelativeNumberOfEvaluatedSamplesParameterName, "The relative number of samples of the dataset partition, which should be randomly chosen for evaluation between the start and end index.", new PercentValue(1)));
|
|---|
| 249 | Parameters.Add(new LookupParameter<BoolValue>("Maximization"));
|
|---|
| 250 | }
|
|---|
| 251 |
|
|---|
| 252 | public override IDeepCloneable Clone(Cloner cloner) {
|
|---|
| 253 | return new ValidationBestScaledSymbolicTimeSeriesPrognosisSolutionAnalyzer(this, cloner);
|
|---|
| 254 | }
|
|---|
| 255 | [StorableHook(Persistence.Default.CompositeSerializers.Storable.HookType.AfterDeserialization)]
|
|---|
| 256 | private void AfterDeserialization() {
|
|---|
| 257 | //if (!Parameters.ContainsKey("Evaluator")) {
|
|---|
| 258 | // Parameters.Add(new LookupParameter<ISingleObjectiveSymbolicTimeSeriesPrognosisEvaluator>("Evaluator", ""));
|
|---|
| 259 | //}
|
|---|
| 260 | }
|
|---|
| 261 |
|
|---|
| 262 | public override IOperation Apply() {
|
|---|
| 263 | var trees = SymbolicExpressionTree;
|
|---|
| 264 |
|
|---|
| 265 | ISingleObjectiveSymbolicTimeSeriesPrognosisEvaluator evaluator = EvaluatorParameter.ActualValue;
|
|---|
| 266 |
|
|---|
| 267 | int trainingStart = ProblemData.TrainingSamplesStart.Value;
|
|---|
| 268 | int trainingEnd = ProblemData.TrainingSamplesEnd.Value;
|
|---|
| 269 | int testStart = ProblemData.TestSamplesStart.Value;
|
|---|
| 270 | int testEnd = ProblemData.TestSamplesEnd.Value;
|
|---|
| 271 |
|
|---|
| 272 | #region validation best model
|
|---|
| 273 | int validationStart = ValidiationSamplesStart.Value;
|
|---|
| 274 | int validationEnd = ValidationSamplesEnd.Value;
|
|---|
| 275 | int rowCount = (int)Math.Ceiling((validationEnd - validationStart) * RelativeNumberOfEvaluatedSamples.Value);
|
|---|
| 276 | IEnumerable<int> rows = RandomEnumerable.SampleRandomNumbers(Random.Next(), validationStart, validationEnd, rowCount);
|
|---|
| 277 | double bestValidationQuality = Maximization.Value ? double.MinValue : double.MaxValue;
|
|---|
| 278 | SymbolicExpressionTree bestTree = null;
|
|---|
| 279 | string conditionalVariableName = ConditionVariableName == null ? null : ConditionVariableName.Value;
|
|---|
| 280 | if (conditionalVariableName != null) {
|
|---|
| 281 | rows = from row in rows
|
|---|
| 282 | where !ProblemData.Dataset[conditionalVariableName, row].IsAlmost(0.0)
|
|---|
| 283 | select row;
|
|---|
| 284 | }
|
|---|
| 285 |
|
|---|
| 286 | foreach (var tree in trees) {
|
|---|
| 287 | double validationQuality;
|
|---|
| 288 | validationQuality = evaluator.Evaluate(tree, ProblemData,
|
|---|
| 289 | SymbolicExpressionTreeInterpreter, rows, PredictionHorizon.Value, LowerEstimationLimit, UpperEstimationLimit);
|
|---|
| 290 | if ((Maximization.Value && validationQuality > bestValidationQuality) ||
|
|---|
| 291 | (!Maximization.Value && validationQuality < bestValidationQuality)) {
|
|---|
| 292 | bestValidationQuality = validationQuality;
|
|---|
| 293 | bestTree = tree;
|
|---|
| 294 | }
|
|---|
| 295 | }
|
|---|
| 296 |
|
|---|
| 297 |
|
|---|
| 298 | if (BestSolutionQualityParameter.ActualValue == null ||
|
|---|
| 299 | (Maximization.Value && BestSolutionQualityParameter.ActualValue.Value < bestValidationQuality) ||
|
|---|
| 300 | (!Maximization.Value && BestSolutionQualityParameter.ActualValue.Value > bestValidationQuality)) {
|
|---|
| 301 | var scaledTree = GetScaledTree(bestTree);
|
|---|
| 302 | var model = new SymbolicTimeSeriesPrognosisModel((ISymbolicTimeSeriesExpressionInterpreter)SymbolicExpressionTreeInterpreter.Clone(), scaledTree);
|
|---|
| 303 | model.Name = "Time series prognosis model";
|
|---|
| 304 | model.Description = "Best solution on validation partition found over the whole run.";
|
|---|
| 305 |
|
|---|
| 306 | var solution = new SymbolicTimeSeriesPrognosisSolution((MultiVariateDataAnalysisProblemData)ProblemData.Clone(), model, PredictionHorizon.Value, conditionalVariableName, LowerEstimationLimit.ToArray(), UpperEstimationLimit.ToArray());
|
|---|
| 307 | solution.Name = BestSolutionParameterName;
|
|---|
| 308 | solution.Description = "Best solution on validation partition found over the whole run.";
|
|---|
| 309 |
|
|---|
| 310 | BestSolutionParameter.ActualValue = solution;
|
|---|
| 311 | BestSolutionQualityParameter.ActualValue = new DoubleValue(bestValidationQuality);
|
|---|
| 312 |
|
|---|
| 313 | // BestSymbolicTimeSeriesPrognosisSolutionAnalyzer.UpdateBestSolutionResults(solution, ProblemData, Results, Generations, VariableFrequencies);
|
|---|
| 314 | }
|
|---|
| 315 |
|
|---|
| 316 | if (!Results.ContainsKey(BestSolutionQualityValuesParameterName)) {
|
|---|
| 317 | Results.Add(new Result(BestSolutionResultName, BestSolutionParameter.ActualValue));
|
|---|
| 318 | Results.Add(new Result(BestSolutionQualityValuesParameterName, new DataTable(BestSolutionQualityValuesParameterName, BestSolutionQualityValuesParameterName)));
|
|---|
| 319 | Results.Add(new Result(BestSolutionQualityParameterName, new DoubleValue()));
|
|---|
| 320 | Results.Add(new Result(CurrentBestValidationQualityParameterName, new DoubleValue()));
|
|---|
| 321 | }
|
|---|
| 322 | Results[BestSolutionResultName].Value = BestSolutionParameter.ActualValue;
|
|---|
| 323 | Results[BestSolutionQualityParameterName].Value = new DoubleValue(BestSolutionQualityParameter.ActualValue.Value);
|
|---|
| 324 | Results[CurrentBestValidationQualityParameterName].Value = new DoubleValue(bestValidationQuality);
|
|---|
| 325 |
|
|---|
| 326 | DataTable validationValues = (DataTable)Results[BestSolutionQualityValuesParameterName].Value;
|
|---|
| 327 | AddValue(validationValues, BestSolutionQualityParameter.ActualValue.Value, BestSolutionQualityParameterName, BestSolutionQualityParameterName);
|
|---|
| 328 | AddValue(validationValues, bestValidationQuality, CurrentBestValidationQualityParameterName, CurrentBestValidationQualityParameterName);
|
|---|
| 329 | #endregion
|
|---|
| 330 | return base.Apply();
|
|---|
| 331 | }
|
|---|
| 332 |
|
|---|
| 333 | private SymbolicExpressionTree GetScaledTree(SymbolicExpressionTree tree) {
|
|---|
| 334 | double[] alpha, beta;
|
|---|
| 335 | int trainingStart = ProblemData.TrainingSamplesStart.Value;
|
|---|
| 336 | int trainingEnd = ProblemData.TrainingSamplesEnd.Value;
|
|---|
| 337 | IEnumerable<int> trainingRows = Enumerable.Range(trainingStart, trainingEnd - trainingStart);
|
|---|
| 338 | string conditionalVariableName = ConditionVariableName == null ? null : ConditionVariableName.Value;
|
|---|
| 339 | if (conditionalVariableName != null) {
|
|---|
| 340 | trainingRows = from row in trainingRows
|
|---|
| 341 | where !ProblemData.Dataset[conditionalVariableName, row].IsAlmost(0.0)
|
|---|
| 342 | select row;
|
|---|
| 343 | }
|
|---|
| 344 |
|
|---|
| 345 | // calculate scaling parameters based on one-step-predictions
|
|---|
| 346 | IEnumerable<string> selectedTargetVariables = from item in ProblemData.TargetVariables
|
|---|
| 347 | where ProblemData.TargetVariables.ItemChecked(item)
|
|---|
| 348 | select item.Value;
|
|---|
| 349 | int dimension = selectedTargetVariables.Count();
|
|---|
| 350 |
|
|---|
| 351 | IEnumerable<IEnumerable<double>> oneStepPredictions =
|
|---|
| 352 | SymbolicExpressionTreeInterpreter.GetSymbolicExpressionTreeValues(tree, ProblemData.Dataset, selectedTargetVariables, trainingRows, 1)
|
|---|
| 353 | .Cast<IEnumerable<double>>();
|
|---|
| 354 | IEnumerable<IEnumerable<double>> originalValues = from row in trainingRows
|
|---|
| 355 | select (from targetVariable in selectedTargetVariables
|
|---|
| 356 | select ProblemData.Dataset[targetVariable, row]);
|
|---|
| 357 | alpha = new double[dimension];
|
|---|
| 358 | beta = new double[dimension];
|
|---|
| 359 |
|
|---|
| 360 | SymbolicTimeSeriesPrognosisScaledNormalizedMseEvaluator.CalculateScalingParameters(originalValues, oneStepPredictions, ref beta, ref alpha);
|
|---|
| 361 |
|
|---|
| 362 | // scale tree for solution
|
|---|
| 363 | return SymbolicVectorRegressionSolutionLinearScaler.Scale(tree, beta, alpha);
|
|---|
| 364 | }
|
|---|
| 365 |
|
|---|
| 366 |
|
|---|
| 367 |
|
|---|
| 368 | private static void AddValue(DataTable table, double data, string name, string description) {
|
|---|
| 369 | DataRow row;
|
|---|
| 370 | table.Rows.TryGetValue(name, out row);
|
|---|
| 371 | if (row == null) {
|
|---|
| 372 | row = new DataRow(name, description);
|
|---|
| 373 | row.Values.Add(data);
|
|---|
| 374 | table.Rows.Add(row);
|
|---|
| 375 | } else {
|
|---|
| 376 | row.Values.Add(data);
|
|---|
| 377 | }
|
|---|
| 378 | }
|
|---|
| 379 | }
|
|---|
| 380 | }
|
|---|
