Changeset 16499 for trunk/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression/3.4/MultiObjective

Timestamp:

01/04/19 15:04:26 (6 years ago)

Author:

bburlacu

Message:

#2977: Implement Pearson R2 & Tree Similarity Evaluator.

File:

• trunk/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression/3.4/MultiObjective/PearsonRSquaredAverageSimilarityEvaluator.cs (copied) (copied from trunk/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression/3.4/MultiObjective/PearsonRSquaredTreeComplexityEvaluator.cs) (3 diffs)

trunk/HeuristicLab.Problems.DataAnalysis.Symbolic.Regression/3.4/MultiObjective/PearsonRSquaredAverageSimilarityEvaluator.cs

@@ -22 +22 @@
 using System;
 using System.Collections.Generic;
+using System.Diagnostics;
+using System.Linq;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
 using HeuristicLab.Data;
 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
+using HeuristicLab.Parameters;
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
 namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Regression {
-  [Item("Pearson R² & Tree Complexity Evaluator", "Calculates the Pearson R² and the tree complexity of a symbolic regression solution.")]
+  [Item("Pearson R² & Average Similarity Evaluator", "Calculates the Pearson R² and the average similarity of a symbolic regression solution candidate.")]
   [StorableClass]
-  public class PearsonRSquaredTreeComplexityEvaluator : SymbolicRegressionMultiObjectiveEvaluator {
+  public class PearsonRSquaredAverageSimilarityEvaluator : SymbolicRegressionMultiObjectiveEvaluator {
+    private const string StrictSimilarityParameterName = "StrictSimilarity";
+
+    private readonly object locker = new object();
+
+    public IFixedValueParameter<BoolValue> StrictSimilarityParameter {
+      get { return (IFixedValueParameter<BoolValue>)Parameters[StrictSimilarityParameterName]; }
+    }
+
+    public bool StrictSimilarity {
+      get { return StrictSimilarityParameter.Value.Value; }
+    }
+
     [StorableConstructor]
-    protected PearsonRSquaredTreeComplexityEvaluator(bool deserializing) : base(deserializing) { }
-    protected PearsonRSquaredTreeComplexityEvaluator(PearsonRSquaredTreeComplexityEvaluator original, Cloner cloner)
+    protected PearsonRSquaredAverageSimilarityEvaluator(bool deserializing) : base(deserializing) { }
+    protected PearsonRSquaredAverageSimilarityEvaluator(PearsonRSquaredAverageSimilarityEvaluator original, Cloner cloner)
       : base(original, cloner) {
     }
     public override IDeepCloneable Clone(Cloner cloner) {
-      return new PearsonRSquaredTreeComplexityEvaluator(this, cloner);
+      return new PearsonRSquaredAverageSimilarityEvaluator(this, cloner);
     }
 
-    public PearsonRSquaredTreeComplexityEvaluator() : base() { }
+    public PearsonRSquaredAverageSimilarityEvaluator() : base() {
+      Parameters.Add(new FixedValueParameter<BoolValue>(StrictSimilarityParameterName, "Use strict similarity calculation.", new BoolValue(false)));
+    }
 
     public override IEnumerable<bool> Maximization { get { return new bool[2] { true, false }; } } // maximize R² and minimize model complexity 
@@ -61 +78 @@
     }
 
-    public static double[] Calculate(ISymbolicDataAnalysisExpressionTreeInterpreter interpreter, ISymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, IRegressionProblemData problemData, IEnumerable<int> rows, bool applyLinearScaling, int decimalPlaces) {
+    public double[] Calculate(ISymbolicDataAnalysisExpressionTreeInterpreter interpreter, ISymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, IRegressionProblemData problemData, IEnumerable<int> rows, bool applyLinearScaling, int decimalPlaces) {
       double r2 = SymbolicRegressionSingleObjectivePearsonRSquaredEvaluator.Calculate(interpreter, solution, lowerEstimationLimit, upperEstimationLimit, problemData, rows, applyLinearScaling);
       if (decimalPlaces >= 0)
         r2 = Math.Round(r2, decimalPlaces);
-      return new double[2] { r2, SymbolicDataAnalysisModelComplexityCalculator.CalculateComplexity(solution) };
+
+      var variables = ExecutionContext.Scope.Variables;
+      if (!variables.ContainsKey("AverageSimilarity")) {
+        lock (locker) {
+          CalculateAverageSimilarities(ExecutionContext.Scope.Parent.SubScopes.Where(x => x.Variables.ContainsKey("SymbolicExpressionTree")).ToArray(), StrictSimilarity);
+
+        }
+      }
+
+      double avgSim = ((DoubleValue)variables["AverageSimilarity"].Value).Value;
+      return new double[2] { r2, avgSim };
     }
 
@@ -82 +109 @@
       return quality;
     }
+
+    private readonly Stopwatch sw = new Stopwatch();
+    public void CalculateAverageSimilarities(IScope[] treeScopes, bool strict) {
+      var trees = treeScopes.Select(x => (ISymbolicExpressionTree)x.Variables["SymbolicExpressionTree"].Value).ToArray();
+      var similarityMatrix = SymbolicExpressionTreeHash.ComputeSimilarityMatrix(trees, simplify: false, strict: strict);
+
+      for (int i = 0; i < treeScopes.Length; ++i) {
+        var scope = treeScopes[i];
+        var avgSimilarity = 0d;
+        for (int j = 0; j < trees.Length; ++j) {
+          if (i == j) continue;
+          avgSimilarity += similarityMatrix[i, j];
+        }
+        avgSimilarity /= trees.Length - 1;
+
+        if (scope.Variables.ContainsKey("AverageSimilarity")) {
+          ((DoubleValue)scope.Variables["AverageSimilarity"].Value).Value = avgSimilarity;
+        } else {
+          scope.Variables.Add(new Core.Variable("AverageSimilarity", new DoubleValue(avgSimilarity)));
+        }
+      }
+    }
   }
 }