Changeset 12720 for trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification

Timestamp:

07/10/15 15:41:09 (9 years ago)

Author:

bburlacu

Message:

#2359: Changed the impact calculators so that the quality value necessary for impacts calculation is calculated with a separate method. Refactored the CalculateImpactAndReplacementValues method to return the new quality in an out-parameter (adjusted method signature in interface accordingly). Added Evaluate method to the regression and classification pruning operators that re-evaluates the tree using the problem evaluator after pruning was performed.

Location:

trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification/3.4

Files:

• SymbolicClassificationPruningAnalyzer.cs (modified) (1 diff)
• SymbolicClassificationPruningOperator.cs (modified) (6 diffs)
• SymbolicClassificationSolutionImpactValuesCalculator.cs (modified) (2 diffs)

trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification/3.4/SymbolicClassificationPruningAnalyzer.cs

@@ -45 +45 @@
 
     public SymbolicClassificationPruningAnalyzer() {
-      Parameters.Add(new ValueParameter<SymbolicDataAnalysisExpressionPruningOperator>(PruningOperatorParameterName, "The operator used to prune trees", new SymbolicClassificationPruningOperator(new SymbolicClassificationSolutionImpactValuesCalculator())));
+      Parameters.Add(new ValueParameter<SymbolicClassificationPruningOperator>(PruningOperatorParameterName, "The operator used to prune trees", new SymbolicClassificationPruningOperator(new SymbolicClassificationSolutionImpactValuesCalculator())));
     }
   }

trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification/3.4/SymbolicClassificationPruningOperator.cs

@@ -35 +35 @@
   public class SymbolicClassificationPruningOperator : SymbolicDataAnalysisExpressionPruningOperator {
     private const string ModelCreatorParameterName = "ModelCreator";
+    private const string EvaluatorParameterName = "Evaluator";
 
     #region parameter properties
     public ILookupParameter<ISymbolicClassificationModelCreator> ModelCreatorParameter {
       get { return (ILookupParameter<ISymbolicClassificationModelCreator>)Parameters[ModelCreatorParameterName]; }
+    }
+
+    public ILookupParameter<ISymbolicClassificationSingleObjectiveEvaluator> EvaluatorParameter {
+      get {
+        return (ILookupParameter<ISymbolicClassificationSingleObjectiveEvaluator>)Parameters[EvaluatorParameterName];
+      }
     }
     #endregion
@@ -51 +58 @@
       : base(impactValuesCalculator) {
       Parameters.Add(new LookupParameter<ISymbolicClassificationModelCreator>(ModelCreatorParameterName));
+      Parameters.Add(new LookupParameter<ISymbolicClassificationSingleObjectiveEvaluator>(EvaluatorParameterName));
     }
 
@@ -62 +70 @@
 
     protected override double Evaluate(IDataAnalysisModel model) {
-      var classificationModel = (IClassificationModel)model;
+      var evaluator = EvaluatorParameter.ActualValue;
+      var classificationModel = (ISymbolicClassificationModel)model;
       var classificationProblemData = (IClassificationProblemData)ProblemDataParameter.ActualValue;
       var rows = Enumerable.Range(FitnessCalculationPartitionParameter.ActualValue.Start, FitnessCalculationPartitionParameter.ActualValue.Size);
-
-      return Evaluate(classificationModel, classificationProblemData, rows);
-    }
-
-    private static double Evaluate(IClassificationModel model, IClassificationProblemData problemData, IEnumerable<int> rows) {
-      var estimatedValues = model.GetEstimatedClassValues(problemData.Dataset, rows);
-      var targetValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, rows);
-      OnlineCalculatorError errorState;
-      var quality = OnlineAccuracyCalculator.Calculate(targetValues, estimatedValues, out errorState);
-      if (errorState != OnlineCalculatorError.None) return double.NaN;
-      return quality;
+      return evaluator.Evaluate(this.ExecutionContext, classificationModel.SymbolicExpressionTree, classificationProblemData, rows);
     }
 
@@ -86 +85 @@
 
       var nodes = clonedTree.Root.GetSubtree(0).GetSubtree(0).IterateNodesPrefix().ToList();
-      double quality = Evaluate(model, problemData, rows);
+      double qualityForImpactsCalculation = double.NaN;
 
       for (int i = 0; i < nodes.Count; ++i) {
@@ -92 +91 @@
         if (node is ConstantTreeNode) continue;
 
-        double impactValue, replacementValue;
-        impactValuesCalculator.CalculateImpactAndReplacementValues(model, node, problemData, rows, out impactValue, out replacementValue, quality);
+        double impactValue, replacementValue, newQualityForImpactsCalculation;
+        impactValuesCalculator.CalculateImpactAndReplacementValues(model, node, problemData, rows, out impactValue, out replacementValue, out newQualityForImpactsCalculation, qualityForImpactsCalculation);
 
         if (pruneOnlyZeroImpactNodes && !impactValue.IsAlmost(0.0)) continue;
@@ -104 +103 @@
         i += node.GetLength() - 1; // skip subtrees under the node that was folded
 
-        quality -= impactValue;
+        qualityForImpactsCalculation = newQualityForImpactsCalculation;
       }
       return model.SymbolicExpressionTree;

trunk/sources/HeuristicLab.Problems.DataAnalysis.Symbolic.Classification/3.4/SymbolicClassificationSolutionImpactValuesCalculator.cs

@@ -47 +47 @@
     }
 
-    public override double CalculateImpactValue(ISymbolicDataAnalysisModel model, ISymbolicExpressionTreeNode node, IDataAnalysisProblemData problemData, IEnumerable<int> rows, double originalQuality = double.NaN) {
+    public override double CalculateImpactValue(ISymbolicDataAnalysisModel model, ISymbolicExpressionTreeNode node, IDataAnalysisProblemData problemData, IEnumerable<int> rows, double qualityForImpactsCalculation = double.NaN) {
       double impactValue, replacementValue;
-      CalculateImpactAndReplacementValues(model, node, problemData, rows, out impactValue, out replacementValue, originalQuality);
+      double newQualityForImpactsCalculation;
+      CalculateImpactAndReplacementValues(model, node, problemData, rows, out impactValue, out replacementValue, out newQualityForImpactsCalculation, qualityForImpactsCalculation);
       return impactValue;
     }
 
     public override void CalculateImpactAndReplacementValues(ISymbolicDataAnalysisModel model, ISymbolicExpressionTreeNode node,
-      IDataAnalysisProblemData problemData, IEnumerable<int> rows, out double impactValue, out double replacementValue,
-      double originalQuality = Double.NaN) {
+      IDataAnalysisProblemData problemData, IEnumerable<int> rows, out double impactValue, out double replacementValue, out double newQualityForImpactsCalculation,
+      double qualityForImpactsCalculation = Double.NaN) {
       var classificationModel = (ISymbolicClassificationModel)model;
       var classificationProblemData = (IClassificationProblemData)problemData;
 
-      var dataset = classificationProblemData.Dataset;
-      var targetClassValues = dataset.GetDoubleValues(classificationProblemData.TargetVariable, rows);
-
-      OnlineCalculatorError errorState;
-      if (double.IsNaN(originalQuality)) {
-        var originalClassValues = classificationModel.GetEstimatedClassValues(dataset, rows);
-        originalQuality = OnlineAccuracyCalculator.Calculate(targetClassValues, originalClassValues, out errorState);
-        if (errorState != OnlineCalculatorError.None) originalQuality = 0.0;
-      }
+      if (double.IsNaN(qualityForImpactsCalculation))
+        qualityForImpactsCalculation = CalculateQualityForImpacts(classificationModel, classificationProblemData, rows);
 
       replacementValue = CalculateReplacementValue(classificationModel, node, classificationProblemData, rows);
@@ -81 +75 @@
       tempModelParentNode.InsertSubtree(i, constantNode);
 
+      OnlineCalculatorError errorState;
+      var dataset = classificationProblemData.Dataset;
+      var targetClassValues = dataset.GetDoubleValues(classificationProblemData.TargetVariable, rows);
       var estimatedClassValues = tempModel.GetEstimatedClassValues(dataset, rows);
-      double newQuality = OnlineAccuracyCalculator.Calculate(targetClassValues, estimatedClassValues, out errorState);
-      if (errorState != OnlineCalculatorError.None) newQuality = 0.0;
+      newQualityForImpactsCalculation = OnlineAccuracyCalculator.Calculate(targetClassValues, estimatedClassValues, out errorState);
+      if (errorState != OnlineCalculatorError.None) newQualityForImpactsCalculation = 0.0;
 
-      impactValue = originalQuality - newQuality;
+      impactValue = qualityForImpactsCalculation - newQualityForImpactsCalculation;
+    }
+
+    public static double CalculateQualityForImpacts(ISymbolicClassificationModel model, IClassificationProblemData problemData, IEnumerable<int> rows) {
+      OnlineCalculatorError errorState;
+      var dataset = problemData.Dataset;
+      var targetClassValues = dataset.GetDoubleValues(problemData.TargetVariable, rows);
+      var originalClassValues = model.GetEstimatedClassValues(dataset, rows);
+      var qualityForImpactsCalculation = OnlineAccuracyCalculator.Calculate(targetClassValues, originalClassValues, out errorState);
+      if (errorState != OnlineCalculatorError.None) qualityForImpactsCalculation = 0.0;
+
+      return qualityForImpactsCalculation;
     }
   }