#region License Information /* HeuristicLab * Copyright (C) 2002-2010 Heuristic and Evolutionary Algorithms Laboratory (HEAL) * * This file is part of HeuristicLab. * * HeuristicLab is free software: you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation, either version 3 of the License, or * (at your option) any later version. * * HeuristicLab is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with HeuristicLab. If not, see . */ #endregion using System; using System.Collections.Generic; using System.Linq; using HeuristicLab.Common; using HeuristicLab.Core; using HeuristicLab.Data; using HeuristicLab.Encodings.RealVectorEncoding; using HeuristicLab.Optimization; using HeuristicLab.Parameters; using HeuristicLab.Persistence.Default.CompositeSerializers.Storable; using HeuristicLab.PluginInfrastructure; namespace HeuristicLab.Problems.DataAnalysis.SupportVectorMachine.ParameterAdjustmentProblem { [Item("Support Vector Machine Parameter Adjustment Problem", "Represents the problem of finding good parameter settings for support vector machines.")] [StorableClass] [Creatable("Problems")] public sealed class SupportVectorMachineParameterAdjustmentProblem : DataAnalysisProblem, ISingleObjectiveProblem { #region Parameter Properties public ValueParameter MaximizationParameter { get { return (ValueParameter)Parameters["Maximization"]; } } IParameter ISingleObjectiveProblem.MaximizationParameter { get { return MaximizationParameter; } } public ValueParameter BoundsParameter { get { return (ValueParameter)Parameters["Bounds"]; } } public ValueParameter ProblemSizeParameter { get { return (ValueParameter)Parameters["ProblemSize"]; } } public new ValueParameter SolutionCreatorParameter { get { return (ValueParameter)Parameters["SolutionCreator"]; } } IParameter IProblem.SolutionCreatorParameter { get { return SolutionCreatorParameter; } } public new ValueParameter EvaluatorParameter { get { return (ValueParameter)Parameters["Evaluator"]; } } IParameter IProblem.EvaluatorParameter { get { return EvaluatorParameter; } } public OptionalValueParameter BestKnownQualityParameter { get { return (OptionalValueParameter)Parameters["BestKnownQuality"]; } } IParameter ISingleObjectiveProblem.BestKnownQualityParameter { get { return BestKnownQualityParameter; } } public OptionalValueParameter BestKnownSolutionParameter { get { return (OptionalValueParameter)Parameters["BestKnownSolution"]; } } #endregion #region Properties public BoolValue Maximization { get { return MaximizationParameter.Value; } set { MaximizationParameter.Value = value; } } public DoubleMatrix Bounds { get { return BoundsParameter.Value; } } public IntValue ProblemSize { get { return ProblemSizeParameter.Value; } set { ProblemSizeParameter.Value = value; } } public new IRealVectorCreator SolutionCreator { get { return SolutionCreatorParameter.Value; } set { SolutionCreatorParameter.Value = value; } } ISolutionCreator IProblem.SolutionCreator { get { return SolutionCreatorParameter.Value; } } public new SupportVectorMachineParameterAdjustmentEvaluator Evaluator { get { return EvaluatorParameter.Value; } set { EvaluatorParameter.Value = value; } } ISingleObjectiveEvaluator ISingleObjectiveProblem.Evaluator { get { return EvaluatorParameter.Value; } } IEvaluator IProblem.Evaluator { get { return EvaluatorParameter.Value; } } public DoubleValue BestKnownQuality { get { return BestKnownQualityParameter.Value; } set { BestKnownQualityParameter.Value = value; } } private List operators; public override IEnumerable Operators { get { return operators; } } #endregion public IntValue TrainingSamplesStart { get { return new IntValue(DataAnalysisProblemData.TrainingSamplesStart.Value); } } public IntValue TrainingSamplesEnd { get { return new IntValue(DataAnalysisProblemData.TrainingSamplesEnd.Value); } } [Storable] private StdDevStrategyVectorCreator strategyVectorCreator; [Storable] private StdDevStrategyVectorCrossover strategyVectorCrossover; [Storable] private StdDevStrategyVectorManipulator strategyVectorManipulator; [StorableConstructor] private SupportVectorMachineParameterAdjustmentProblem(bool deserializing) : base() { } public SupportVectorMachineParameterAdjustmentProblem() : base() { UniformRandomRealVectorCreator creator = new UniformRandomRealVectorCreator(); SupportVectorMachineParameterAdjustmentEvaluator evaluator = new SupportVectorMachineParameterAdjustmentEvaluator(); var bounds = new DoubleMatrix(new double[,] { { 0.01, 1.0 }, { -7, 9}, { -7, 9} }); Parameters.Add(new ValueParameter("Maximization", "Set to false as we want to minimize the error.", new BoolValue(false))); Parameters.Add(new ValueParameter("Bounds", "The lower and upper bounds in each dimension.", bounds)); Parameters.Add(new ValueParameter("ProblemSize", "The dimension of the problem.", new IntValue(3))); Parameters.Add(new ValueParameter("SolutionCreator", "The operator which should be used to create new test function solutions.", creator)); Parameters.Add(new ValueParameter("Evaluator", "The operator which should be used to evaluate test function solutions.", evaluator)); Parameters.Add(new OptionalValueParameter("BestKnownQuality", "The quality of the best known solution of this test function.", new DoubleValue(0))); Parameters.Add(new OptionalValueParameter("BestKnownSolution", "The best known solution for this test function instance.")); Parameters.Add(new OptionalValueParameter("ActualSamples", "The percentage of samples to use for cross validation.")); strategyVectorCreator = new StdDevStrategyVectorCreator(); strategyVectorCreator.LengthParameter.Value = ProblemSize; strategyVectorCrossover = new StdDevStrategyVectorCrossover(); strategyVectorManipulator = new StdDevStrategyVectorManipulator(); strategyVectorManipulator.LearningRateParameter.Value = new DoubleValue(0.5); strategyVectorManipulator.GeneralLearningRateParameter.Value = new DoubleValue(0.5); creator.RealVectorParameter.ActualName = "ParameterVector"; ParameterizeSolutionCreator(); ParameterizeEvaluator(); Initialize(); UpdateStrategyVectorBounds(); } public override IDeepCloneable Clone(Cloner cloner) { SupportVectorMachineParameterAdjustmentProblem clone = (SupportVectorMachineParameterAdjustmentProblem)base.Clone(cloner); clone.strategyVectorCreator = (StdDevStrategyVectorCreator)cloner.Clone(strategyVectorCreator); clone.strategyVectorCrossover = (StdDevStrategyVectorCrossover)cloner.Clone(strategyVectorCrossover); clone.strategyVectorManipulator = (StdDevStrategyVectorManipulator)cloner.Clone(strategyVectorManipulator); clone.Initialize(); return clone; } protected override void OnDataAnalysisProblemChanged(EventArgs e) { ParameterizeEvaluator(); base.OnDataAnalysisProblemChanged(e); } #region Events private void SolutionCreatorParameter_ValueChanged(object sender, EventArgs e) { ParameterizeSolutionCreator(); ParameterizeAnalyzers(); SolutionCreator_RealVectorParameter_ActualNameChanged(null, EventArgs.Empty); } private void SolutionCreator_RealVectorParameter_ActualNameChanged(object sender, EventArgs e) { ParameterizeEvaluator(); ParameterizeOperators(); ParameterizeAnalyzers(); } private void EvaluatorParameter_ValueChanged(object sender, EventArgs e) { ParameterizeEvaluator(); ParameterizeAnalyzers(); RaiseReset(EventArgs.Empty); } private void strategyVectorCreator_StrategyParameterParameter_ActualNameChanged(object sender, EventArgs e) { string name = strategyVectorCreator.StrategyParameterParameter.ActualName; strategyVectorCrossover.ParentsParameter.ActualName = name; strategyVectorCrossover.StrategyParameterParameter.ActualName = name; strategyVectorManipulator.StrategyParameterParameter.ActualName = name; } #endregion #region Helpers [StorableHook(HookType.AfterDeserialization)] private void Initialize() { InitializeOperators(); SolutionCreatorParameter.ValueChanged += new EventHandler(SolutionCreatorParameter_ValueChanged); SolutionCreator.RealVectorParameter.ActualNameChanged += new EventHandler(SolutionCreator_RealVectorParameter_ActualNameChanged); EvaluatorParameter.ValueChanged += new EventHandler(EvaluatorParameter_ValueChanged); strategyVectorCreator.StrategyParameterParameter.ActualNameChanged += new EventHandler(strategyVectorCreator_StrategyParameterParameter_ActualNameChanged); } private void InitializeOperators() { operators = new List(); operators.AddRange(ApplicationManager.Manager.GetInstances().Cast()); operators.Add(new SupportVectorMachineParameterAdjustmentBestSolutionAnalyzer()); operators.Add(strategyVectorCreator); operators.Add(strategyVectorCrossover); operators.Add(strategyVectorManipulator); ParameterizeOperators(); } private void ParameterizeSolutionCreator() { SolutionCreator.LengthParameter.Value = new IntValue(ProblemSize.Value); } private void ParameterizeEvaluator() { Evaluator.ParameterVectorParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; Evaluator.SamplesStartParameter.Value = TrainingSamplesStart; Evaluator.SamplesEndParameter.Value = TrainingSamplesEnd; Evaluator.NumberOfFoldsParameter.Value = new IntValue(5); } private void ParameterizeOperators() { foreach (IRealVectorCrossover op in Operators.OfType()) { op.ParentsParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; op.ChildParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; op.BoundsParameter.ActualName = BoundsParameter.Name; } foreach (IRealVectorManipulator op in Operators.OfType()) { op.RealVectorParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; op.BoundsParameter.ActualName = BoundsParameter.Name; } foreach (IRealVectorMoveOperator op in Operators.OfType()) { op.RealVectorParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; } foreach (IRealVectorMoveGenerator op in Operators.OfType()) { op.BoundsParameter.ActualName = BoundsParameter.Name; } foreach (SupportVectorMachineParameterAdjustmentEvaluator op in Operators.OfType()) { op.QualityParameter.ActualName = Evaluator.QualityParameter.ActualName; op.ParameterVectorParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; } } private void ParameterizeAnalyzers() { foreach (SupportVectorMachineParameterAdjustmentBestSolutionAnalyzer analyzer in Operators.OfType()) { analyzer.DataAnalysisProblemDataParameter.ActualName = DataAnalysisProblemDataParameter.Name; analyzer.ParameterVectorParameter.ActualName = SolutionCreator.RealVectorParameter.ActualName; analyzer.QualityParameter.ActualName = Evaluator.QualityParameter.ActualName; } } private void UpdateStrategyVectorBounds() { DoubleMatrix strategyBounds = (DoubleMatrix)Bounds.Clone(); for (int i = 0; i < strategyBounds.Rows; i++) if (strategyBounds[i, 0] < 0) strategyBounds[i, 0] = 0; strategyVectorCreator.BoundsParameter.Value = strategyBounds; } #endregion } }