source: branches/HeuristicLab.TreeSimplifier/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification/3.4/SymbolicDiscriminantFunctionClassificationSolutionValuesCalculator.cs @ 8388

using System.Collections.Generic;
using System.Linq;
using HeuristicLab.Common;
using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;

namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Classification {
  public class SymbolicDiscriminantFunctionClassificationSolutionValuesCalculator : SymbolicDataAnalysisSolutionValuesCalculator {
    public override Dictionary<ISymbolicExpressionTreeNode, double> CalculateReplacementValues(ISymbolicExpressionTree tree,
                                                                                      ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
                                                                                      IDataAnalysisProblemData problemData) {
      var replacementValues = new Dictionary<ISymbolicExpressionTreeNode, double>();
      foreach (ISymbolicExpressionTreeNode node in tree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPrefix()) {
        replacementValues[node] = CalculateReplacementValue(node, tree, interpreter, problemData);
      }
      return replacementValues;
    }
    public override Dictionary<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;
      Dictionary<ISymbolicExpressionTreeNode, double> impactValues = new Dictionary<ISymbolicExpressionTreeNode, double>();
      List<ISymbolicExpressionTreeNode> nodes = tree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPostfix().ToList();

      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;

      foreach (ISymbolicExpressionTreeNode node in nodes) {
        var parent = node.Parent;
        constantNode.Value = CalculateReplacementValue(node, tree, interpreter, classificationProblemData);
        ISymbolicExpressionTreeNode replacementNode = constantNode;
        SwitchNode(parent, node, replacementNode);
        var newOutput = interpreter.GetSymbolicExpressionTreeValues(tree, dataset, rows)
          .LimitToRange(lowerEstimationLimit, upperEstimationLimit)
          .ToArray();
        double newGini = NormalizedGiniCalculator.Calculate(targetClassValues, newOutput, out errorState);
        if (errorState != OnlineCalculatorError.None) newGini = 0.0;

        // impact = 0 if no change
        // impact < 0 if new solution is better
        // impact > 0 if new solution is worse
        impactValues[node] = originalGini - newGini;
        SwitchNode(parent, replacementNode, node);
      }
      return impactValues;

    }
  }
}