Changeset 332 for trunk/sources/HeuristicLab.StructureIdentification/Evaluation

Timestamp:

06/23/08 19:13:17 (16 years ago)

Author:

gkronber

Message:

implemented #172

Location:

trunk/sources/HeuristicLab.StructureIdentification/Evaluation

Files:

• EarlyStoppingMeanSquaredErrorEvaluator.cs (modified) (4 diffs)
• MeanSquaredErrorEvaluator.cs (modified) (5 diffs)

trunk/sources/HeuristicLab.StructureIdentification/Evaluation/EarlyStoppingMeanSquaredErrorEvaluator.cs

@@ -45 +45 @@
     }
 
+    // evaluates the function-tree for the given target-variable and the whole dataset and returns the MSE
     public override double Evaluate(IScope scope, IFunctionTree functionTree, int targetVariable, Dataset dataset) {
       double qualityLimit = GetVariableValue<DoubleData>("QualityLimit", scope, false).Data;
+      bool useEstimatedValues = GetVariableValue<BoolData>("UseEstimatedTargetValue", scope, false).Data;
+      if(useEstimatedValues && backupValues == null) {
+        backupValues = new double[dataset.Rows];
+        for(int i = 0; i < dataset.Rows; i++) {
+          backupValues[i] = dataset.GetValue(i, targetVariable);
+        }
+      }
       double errorsSquaredSum = 0;
       double targetMean = dataset.GetMean(targetVariable);
@@ -59 +67 @@
           estimated = targetMean - maximumPunishment;
         }
+
         double error = estimated - original;
         errorsSquaredSum += error * error;
@@ -65 +74 @@
         if(errorsSquaredSum / dataset.Rows >= qualityLimit) {
           scope.GetVariableValue<DoubleData>("TotalEvaluatedNodes", true).Data = totalEvaluatedNodes + treeSize * sample+1;
+          if(useEstimatedValues) RestoreDataset(dataset, targetVariable, 0, sample);
           return errorsSquaredSum / (sample + 1); // return estimated MSE (when the remaining errors are on average the same)
         }
+        if(useEstimatedValues) {
+          dataset.SetValue(sample, targetVariable, estimated);
+        }
       }
+      if(useEstimatedValues) RestoreDataset(dataset, targetVariable, 0, dataset.Rows);
       errorsSquaredSum /= dataset.Rows;
       if(double.IsNaN(errorsSquaredSum) || double.IsInfinity(errorsSquaredSum)) {
@@ -75 +89 @@
       return errorsSquaredSum;
     }
+
+    private void RestoreDataset(Dataset dataset, int targetVariable, int from, int to) {
+      for(int i = from; i < to; i++) {
+        dataset.SetValue(i, targetVariable, backupValues[i]);
+      }
+    }
   }
 }

trunk/sources/HeuristicLab.StructureIdentification/Evaluation/MeanSquaredErrorEvaluator.cs

@@ -32 +32 @@
 namespace HeuristicLab.StructureIdentification {
   public class MeanSquaredErrorEvaluator : GPEvaluatorBase {
+    protected double[] backupValues;
     public override string Description {
       get {
@@ -41 +42 @@
     public MeanSquaredErrorEvaluator()
       : base() {
+      AddVariableInfo(new VariableInfo("UseEstimatedTargetValue", "Wether to use the original (measured) or the estimated (calculated) value for the targat variable when doing autoregressive modelling", typeof(BoolData), VariableKind.In));
+      GetVariableInfo("UseEstimatedTargetValue").Local = true;
+      AddVariable(new HeuristicLab.Core.Variable("UseEstimatedTargetValue", new BoolData(false)));
     }
 
@@ -46 +50 @@
       double errorsSquaredSum = 0;
       double targetMean = dataset.GetMean(targetVariable);
+      bool useEstimatedValues = GetVariableValue<BoolData>("UseEstimatedTargetValue", scope, false).Data;
+      if(useEstimatedValues && backupValues == null) {
+        backupValues = new double[dataset.Rows];
+        for(int i = 0; i < dataset.Rows; i++) {
+          backupValues[i] = dataset.GetValue(i, targetVariable);
+        }
+      }
+
       for(int sample = 0; sample < dataset.Rows; sample++) {
-
         double estimated = functionTree.Evaluate(dataset, sample);
         double original = dataset.GetValue(sample, targetVariable);
@@ -59 +70 @@
         double error = estimated - original;
         errorsSquaredSum += error * error;
+        if(useEstimatedValues) {
+          dataset.SetValue(sample, targetVariable, estimated);
+        }
       }
+
+      if(useEstimatedValues) RestoreDataset(dataset, targetVariable);
       errorsSquaredSum /= dataset.Rows;
       if(double.IsNaN(errorsSquaredSum) || double.IsInfinity(errorsSquaredSum)) {
@@ -67 +83 @@
       return errorsSquaredSum;
     }
+
+    private void RestoreDataset(Dataset dataset, int targetVariable) {
+      for(int i = 0; i < dataset.Rows; i++) {
+        dataset.SetValue(i, targetVariable, backupValues[i]);
+      }
+    }
   }
 }