source: branches/HeuristicLab.TreeSimplifier/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification/3.4/SymbolicDiscriminantFunctionClassificationSolutionImpactValuesCalculator.cs @ 8935

#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2012 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;
using System.Collections.Generic;
using System.Linq;
using HeuristicLab.Common;
using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;

namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Classification {
  public class SymbolicDiscriminantFunctionClassificationSolutionImpactValuesCalculator : SymbolicDataAnalysisSolutionImpactValuesCalculator {
    public override IEnumerable<Tuple<ISymbolicExpressionTreeNode, double>> CalculateReplacementValues(ISymbolicExpressionTree tree,
                                                                                                       ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
                                                                                                       IDataAnalysisProblemData problemData) {
      return from node in tree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPrefix()
             select new Tuple<ISymbolicExpressionTreeNode, double>(node, CalculateReplacementValue(node, tree, interpreter, problemData));
    }

    public override IEnumerable<Tuple<ISymbolicExpressionTreeNode, double>> CalculateImpactValues(ISymbolicExpressionTree tree,
                                                                                                  ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
                                                                                                  IDataAnalysisProblemData classificationProblemData,
                                                                                                  double lowerEstimationLimit, double upperEstimationLimit) {
      var problemData = (IClassificationProblemData)classificationProblemData;
      var dataset = problemData.Dataset;
      var rows = problemData.TrainingIndices;
      string targetVariable = problemData.TargetVariable;
      var targetClassValues = dataset.GetDoubleValues(targetVariable, rows);
      var originalOutput = interpreter.GetSymbolicExpressionTreeValues(tree, dataset, rows).LimitToRange(lowerEstimationLimit, upperEstimationLimit).ToArray();
      OnlineCalculatorError errorState;
      double originalGini = NormalizedGiniCalculator.Calculate(targetClassValues, originalOutput, out errorState);
      if (errorState != OnlineCalculatorError.None) originalGini = 0.0;

      return from node in tree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPostfix()
             select new Tuple<ISymbolicExpressionTreeNode, double>(node, CalculateImpact(tree, originalGini, node, interpreter, problemData, lowerEstimationLimit, upperEstimationLimit));
    }

    private static double CalculateImpact(ISymbolicExpressionTree tree, double originalQuality, ISymbolicExpressionTreeNode node,
                                          ISymbolicDataAnalysisExpressionTreeInterpreter interpreter, IClassificationProblemData problemData,
                                          double lowerEstimationLimit, double upperEstimationLimit) {
      var dataset = problemData.Dataset;
      var rows = problemData.TrainingIndices.ToList();
      string targetVariable = problemData.TargetVariable;
      var targetValues = dataset.GetDoubleValues(targetVariable, rows).ToList();

      var parent = node.Parent;
      var constantNode = (ConstantTreeNode)new Constant().CreateTreeNode();
      constantNode.Value = CalculateReplacementValue(node, tree, interpreter, problemData);
      SwitchNode(parent, node, constantNode);
      var newOutput = interpreter.GetSymbolicExpressionTreeValues(tree, dataset, rows)
                                 .LimitToRange(lowerEstimationLimit, upperEstimationLimit)
                                 .ToArray();
      OnlineCalculatorError errorState;
      double quality = NormalizedGiniCalculator.Calculate(targetValues, newOutput, out errorState);
      if (errorState != OnlineCalculatorError.None) quality = 0.0;
      SwitchNode(parent, constantNode, node);
      // impact = 0 if no change
      // impact < 0 if new solution is better
      // impact > 0 if new solution is worse
      return originalQuality - quality;
    }
  }
}