Changeset 4555 for branches/DataAnalysis/HeuristicLab.Problems.DataAnalysis.Regression/3.3/Symbolic

Timestamp:

10/05/10 16:30:57 (14 years ago)

Author:

gkronber

Message:

Improved time series evaluators. #1142

Location:

branches/DataAnalysis/HeuristicLab.Problems.DataAnalysis.Regression/3.3/Symbolic/Evaluators

Files:

• SymbolicRegressionScaledMeanAndVarianceSquaredErrorEvaluator.cs (modified) (2 diffs)
• SymbolicRegressionScaledMeanSquaredErrorEvaluator.cs (modified) (3 diffs)

branches/DataAnalysis/HeuristicLab.Problems.DataAnalysis.Regression/3.3/Symbolic/Evaluators/SymbolicRegressionScaledMeanAndVarianceSquaredErrorEvaluator.cs

@@ -126 +126 @@
       IEnumerable<double> originalValues = dataset.GetEnumeratedVariableValues(targetVariable, rows);
       IEnumerable<double> estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, dataset, rows);
-      CalculateScalingParameters(originalValues, estimatedValues, out beta, out alpha);
+      SymbolicRegressionScaledMeanSquaredErrorEvaluator.CalculateScalingParameters(originalValues, estimatedValues, out beta, out alpha);
 
       return CalculateWithScaling(interpreter, solution, lowerEstimationLimit, upperEstimationLimit, dataset, targetVariable, rows, beta, alpha, out meanSE, out varianceSE, out count, out bias, out variance, out covariance);
@@ -174 +174 @@
       }
     }
-
-    /// <summary>
-    /// Calculates linear scaling parameters in one pass.
-    /// The formulas to calculate the scaling parameters were taken from Scaled Symblic Regression by Maarten Keijzer.
-    /// http://www.springerlink.com/content/x035121165125175/
-    /// </summary>
-    public static void CalculateScalingParameters(IEnumerable<double> original, IEnumerable<double> estimated, out double beta, out double alpha) {
-      IEnumerator<double> originalEnumerator = original.GetEnumerator();
-      IEnumerator<double> estimatedEnumerator = estimated.GetEnumerator();
-      OnlineMeanAndVarianceCalculator yVarianceCalculator = new OnlineMeanAndVarianceCalculator();
-      OnlineMeanAndVarianceCalculator tMeanCalculator = new OnlineMeanAndVarianceCalculator();
-      OnlineCovarianceEvaluator ytCovarianceEvaluator = new OnlineCovarianceEvaluator();
-      int cnt = 0;
-
-      while (originalEnumerator.MoveNext() & estimatedEnumerator.MoveNext()) {
-        double y = estimatedEnumerator.Current;
-        double t = originalEnumerator.Current;
-        if (IsValidValue(t) && IsValidValue(y)) {
-          tMeanCalculator.Add(t);
-          yVarianceCalculator.Add(y);
-          ytCovarianceEvaluator.Add(y, t);
-
-          cnt++;
-        }
-      }
-
-      if (estimatedEnumerator.MoveNext() || originalEnumerator.MoveNext())
-        throw new ArgumentException("Number of elements in original and estimated enumeration doesn't match.");
-      if (cnt < 2) {
-        alpha = 0;
-        beta = 1;
-      } else {
-        if (yVarianceCalculator.PopulationVariance.IsAlmost(0.0))
-          beta = 1;
-        else
-          beta = ytCovarianceEvaluator.Covariance / yVarianceCalculator.PopulationVariance;
-
-        alpha = tMeanCalculator.Mean - beta * yVarianceCalculator.Mean;
-      }
-    }
-
-    private static bool IsValidValue(double d) {
-      return !double.IsInfinity(d) && !double.IsNaN(d) && d > -1.0E07 && d < 1.0E07;  // don't consider very large or very small values for scaling
-    }
   }
 }

branches/DataAnalysis/HeuristicLab.Problems.DataAnalysis.Regression/3.3/Symbolic/Evaluators/SymbolicRegressionScaledMeanSquaredErrorEvaluator.cs

@@ -100 +100 @@
     }
 
-    /// <summary>
-    /// Calculates linear scaling parameters in one pass.
-    /// The formulas to calculate the scaling parameters were taken from Scaled Symblic Regression by Maarten Keijzer.
-    /// http://www.springerlink.com/content/x035121165125175/
-    /// </summary>
     public static void CalculateScalingParameters(IEnumerable<double> original, IEnumerable<double> estimated, out double beta, out double alpha) {
       IEnumerator<double> originalEnumerator = original.GetEnumerator();
       IEnumerator<double> estimatedEnumerator = estimated.GetEnumerator();
-      OnlineMeanAndVarianceCalculator yVarianceCalculator = new OnlineMeanAndVarianceCalculator();
-      OnlineMeanAndVarianceCalculator tMeanCalculator = new OnlineMeanAndVarianceCalculator();
-      OnlineCovarianceEvaluator ytCovarianceEvaluator = new OnlineCovarianceEvaluator();
-      int cnt = 0;
+      OnlineLinearScalingCalculator linearScalingCalculator = new OnlineLinearScalingCalculator();
 
       while (originalEnumerator.MoveNext() & estimatedEnumerator.MoveNext()) {
@@ -117 +109 @@
         double t = originalEnumerator.Current;
         if (IsValidValue(t) && IsValidValue(y)) {
-          tMeanCalculator.Add(t);
-          yVarianceCalculator.Add(y);
-          ytCovarianceEvaluator.Add(y, t);
-
-          cnt++;
+          linearScalingCalculator.Add(t, y);
         }
       }
@@ -127 +115 @@
       if (estimatedEnumerator.MoveNext() || originalEnumerator.MoveNext())
         throw new ArgumentException("Number of elements in original and estimated enumeration doesn't match.");
-      if (cnt < 2) {
-        alpha = 0;
-        beta = 1;
-      } else {
-        if (yVarianceCalculator.PopulationVariance.IsAlmost(0.0))
-          beta = 1;
-        else
-          beta = ytCovarianceEvaluator.Covariance / yVarianceCalculator.PopulationVariance;
-
-        alpha = tMeanCalculator.Mean - beta * yVarianceCalculator.Mean;
+      else {
+        beta = linearScalingCalculator.Beta;
+        alpha = linearScalingCalculator.Alpha;
       }
     }