source: branches/2205_OptimizationNetworks/HeuristicLab.Networks.IntegratedOptimization.MachineLearning/FeatureSelectionNetwork.cs

#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2017 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 <http://www.gnu.org/licenses/>.
 */
#endregion

using System.Linq;
using System.Threading;
using HeuristicLab.Algorithms.DataAnalysis;
using HeuristicLab.Algorithms.OffspringSelectionGeneticAlgorithm;
using HeuristicLab.Common;
using HeuristicLab.Core;
using HeuristicLab.Core.Networks;
using HeuristicLab.Data;
using HeuristicLab.Encodings.BinaryVectorEncoding;
using HeuristicLab.Optimization;
using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;

namespace HeuristicLab.Networks.IntegratedOptimization.MachineLearning {
  [Item("Feature Selection Network 2", "")]
  [Creatable("Optimization Networks")]
  [StorableClass]
  public sealed class FeatureSelectionNetwork : Network {
    [StorableConstructor]
    private FeatureSelectionNetwork(bool deserializing) : base(deserializing) { }

    private FeatureSelectionNetwork(FeatureSelectionNetwork original, Cloner cloner)
      : base(original, cloner) {
    }
    public override IDeepCloneable Clone(Cloner cloner) {
      return new FeatureSelectionNetwork(this, cloner);
    }

    [Storable]
    private readonly OrchestratedAlgorithmNode FeatureSelectionAlgorithmNode;
    [Storable]
    private readonly OrchestratedAlgorithmNode RegressionAlgorithmNode;
    [Storable]
    private readonly FeatureSelectionOrchestrator Orchestrator;

    public FeatureSelectionNetwork()
      : base() {
      Orchestrator = new FeatureSelectionOrchestrator(this);
      Nodes.Add(Orchestrator);

      var featureSelectionAlgorithm = CreateFeatureSelectionAlgorithm();
      //TODO configure FeatureSelectionProblem

      FeatureSelectionAlgorithmNode = new OrchestratedAlgorithmNode("Feature Selection");
      FeatureSelectionAlgorithmNode.Algorithm = featureSelectionAlgorithm;
      FeatureSelectionAlgorithmNode.EvaluationPort.ConnectedPort = Orchestrator.FeatureSelectionEvaluationPort;
      FeatureSelectionAlgorithmNode.EvaluationPort.CloneParametersFromPort(Orchestrator.FeatureSelectionEvaluationPort);
      Nodes.Add(FeatureSelectionAlgorithmNode);

      var regressionAlgorithm = CreateRegressionAlgorithm();
      RegressionAlgorithmNode = new OrchestratedAlgorithmNode("Regression");
      RegressionAlgorithmNode.Algorithm = regressionAlgorithm;
      Orchestrator.RegressionOrchestrationPort.ConnectedPort = RegressionAlgorithmNode.OrchestrationPort;
      RegressionAlgorithmNode.OrchestrationPort.CloneParametersFromPort(Orchestrator.RegressionOrchestrationPort);
      Nodes.Add(RegressionAlgorithmNode);
    }

    private IAlgorithm CreateFeatureSelectionAlgorithm() {
      var problem = new OrchestratedBinaryProblem(Orchestrator, 0, false);
      problem.Encoding.SolutionCreator = new RandomBinaryVectorCreator() { TrueProbability = new DoubleValue(0.2) };

      var osga = new OffspringSelectionGeneticAlgorithm();
      osga.Problem = problem;
      osga.PopulationSize.Value = 100;
      osga.ComparisonFactorLowerBound.Value = 1;
      osga.ComparisonFactorUpperBound.Value = 1;
      osga.SuccessRatio.Value = 1.0;
      osga.MutationProbability.Value = 0.15;
      osga.Mutator = osga.MutatorParameter.ValidValues.OfType<SomePositionsBitflipManipulator>().First();
      return osga;
    }

    private static IAlgorithm CreateRegressionAlgorithm() {
      var linreg = new LinearRegression();
      return linreg;
    }


    public void Prepare() { Prepare(false); }
    public void Prepare(bool clearRuns) {
      var msg = FeatureSelectionAlgorithmNode.OrchestrationPort.PrepareMessage();
      if (clearRuns)
        msg["OrchestrationMessage"] = new EnumValue<OrchestrationMessage>(OrchestrationMessage.Prepare | OrchestrationMessage.ClearRuns);
      else
        msg["OrchestrationMessage"] = new EnumValue<OrchestrationMessage>(OrchestrationMessage.Prepare);
      FeatureSelectionAlgorithmNode.OrchestrationPort.ReceiveMessage(msg, new CancellationToken());
    }

    public void Start() {
      var problem = (OrchestratedBinaryProblem)FeatureSelectionAlgorithmNode.Algorithm.Problem;
      problem.Encoding.Length = Orchestrator.RegressionProblemData.AllowedInputVariables.Count();

      var msg = FeatureSelectionAlgorithmNode.OrchestrationPort.PrepareMessage();
      msg["OrchestrationMessage"] = new EnumValue<OrchestrationMessage>(OrchestrationMessage.Start);
      FeatureSelectionAlgorithmNode.OrchestrationPort.ReceiveMessage(msg, new CancellationToken());
    }
    public void Pause() {
      var msg = FeatureSelectionAlgorithmNode.OrchestrationPort.PrepareMessage();
      msg["OrchestrationMessage"] = new EnumValue<OrchestrationMessage>(OrchestrationMessage.Pause);
      FeatureSelectionAlgorithmNode.OrchestrationPort.ReceiveMessage(msg, new CancellationToken());
    }
    public void Stop() {
      var msg = FeatureSelectionAlgorithmNode.OrchestrationPort.PrepareMessage();
      msg["OrchestrationMessage"] = new EnumValue<OrchestrationMessage>(OrchestrationMessage.Stop);
      FeatureSelectionAlgorithmNode.OrchestrationPort.ReceiveMessage(msg, new CancellationToken());
      RegressionAlgorithmNode.Algorithm.Runs.Clear();
    }
  }
}