Changeset 8486 for branches/HeuristicLab.TimeSeries

Timestamp:

08/14/12 13:59:47 (12 years ago)

Author:

mkommend

Message:

#1081: Corrected evaluators and time series models.

Location:

branches/HeuristicLab.TimeSeries

Files:

• HeuristicLab.Problems.DataAnalysis.Symbolic.TimeSeriesPrognosis/3.4/SingleObjective/SymbolicTimeSeriesPrognosisSingleObjectiveEvaluator.cs (modified) (4 diffs)
• HeuristicLab.Problems.DataAnalysis.Symbolic.TimeSeriesPrognosis/3.4/SingleObjective/SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator.cs (modified) (4 diffs)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionTreeInterpreter.cs (modified) (3 diffs)
• HeuristicLab.Problems.DataAnalysis.Views/3.4/HeuristicLab.Problems.DataAnalysis.Views-3.4.csproj (modified) (2 diffs)
• HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/Models/TimeSeriesPrognosisAutoRegressiveModel.cs (modified) (3 diffs)
• HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/Models/TimeSeriesPrognosisMovingAverageModel.cs (modified) (2 diffs)
• HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisProblemData.cs (modified) (2 diffs)

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis.Symbolic.TimeSeriesPrognosis/3.4/SingleObjective/SymbolicTimeSeriesPrognosisSingleObjectiveEvaluator.cs

@@ -23 +23 @@
 using System;
 using System.Collections.Generic;
+using System.Linq;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
@@ -61 +62 @@
     [ThreadStatic]
     private static double[] cache;
-    protected static void CalculateWithScaling(IEnumerable<double> targetValues, IEnumerable<double> estimatedValues, IOnlineCalculator calculator, int maxRows) {
+    protected static void CalculateWithScaling(IEnumerable<double> targetValues, IEnumerable<double> estimatedValues,
+      double lowerEstimationLimit, double upperEstimationLimit,
+      IOnlineCalculator calculator, int maxRows) {
       if (cache == null || cache.GetLength(0) < maxRows) {
         cache = new double[maxRows];
@@ -76 +79 @@
         double target = targetValuesEnumerator.Current;
         double estimated = estimatedValuesEnumerator.Current;
+        cache[i] = estimated;
         linearScalingCalculator.Add(estimated, target);
-        cache[i] = estimated;
         i++;
       }
@@ -85 +88 @@
       //calculate the quality by using the passed online calculator
       targetValuesEnumerator = targetValues.GetEnumerator();
-      i = 0;
-      while (targetValuesEnumerator.MoveNext()) {
-        double target = targetValuesEnumerator.Current;
-        double estimated = cache[i] * beta + alpha;
-        calculator.Add(target, estimated);
-        i++;
+      var scaledBoundedEstimatedValuesEnumerator = Enumerable.Range(0, i).Select(x => cache[x] * beta + alpha)
+        .LimitToRange(lowerEstimationLimit, upperEstimationLimit).GetEnumerator();
+
+      while (targetValuesEnumerator.MoveNext() & scaledBoundedEstimatedValuesEnumerator.MoveNext()) {
+        calculator.Add(targetValuesEnumerator.Current, scaledBoundedEstimatedValuesEnumerator.Current);
       }
     }

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis.Symbolic.TimeSeriesPrognosis/3.4/SingleObjective/SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator.cs

@@ -35 +35 @@
     [StorableConstructor]
     protected SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator(bool deserializing) : base(deserializing) { }
-    protected SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator(SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator original, Cloner cloner)
-      : base(original, cloner) {
-    }
+    protected SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator(SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator original, Cloner cloner) : base(original, cloner) { }
     public override IDeepCloneable Clone(Cloner cloner) {
       return new SymbolicTimeSeriesPrognosisSingleObjectiveMeanSquaredErrorEvaluator(this, cloner);
@@ -65 +63 @@
       IEnumerable<double> targetValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, rows.Zip(horizions, Enumerable.Range).SelectMany(r => r));
       IEnumerable<double> estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, problemData.Dataset, rows, horizions).SelectMany(x => x);
-      IEnumerable<double> boundedEstimatedValues = estimatedValues.LimitToRange(lowerEstimationLimit, upperEstimationLimit);
       OnlineCalculatorError errorState;
 
@@ -71 +68 @@
       if (applyLinearScaling && horizon == 1) { //perform normal evaluation and afterwards scale the solution and calculate the fitness value        
         var mseCalculator = new OnlineMeanSquaredErrorCalculator();
-        CalculateWithScaling(targetValues, boundedEstimatedValues, mseCalculator, problemData.Dataset.Rows * horizon);
+        CalculateWithScaling(targetValues, estimatedValues, lowerEstimationLimit, upperEstimationLimit, mseCalculator, problemData.Dataset.Rows * horizon);
         errorState = mseCalculator.ErrorState;
         mse = mseCalculator.MeanSquaredError;
@@ -79 +76 @@
         var scaledSolution = model.SymbolicExpressionTree;
         estimatedValues = interpreter.GetSymbolicExpressionTreeValues(scaledSolution, problemData.Dataset, rows, horizions).SelectMany(x => x);
-        boundedEstimatedValues = estimatedValues.LimitToRange(lowerEstimationLimit, upperEstimationLimit);
+        var boundedEstimatedValues = estimatedValues.LimitToRange(lowerEstimationLimit, upperEstimationLimit);
         mse = OnlineMeanSquaredErrorCalculator.Calculate(targetValues, boundedEstimatedValues, out errorState);
-      } else
+      } else {
+        var boundedEstimatedValues = estimatedValues.LimitToRange(lowerEstimationLimit, upperEstimationLimit);
         mse = OnlineMeanSquaredErrorCalculator.Calculate(targetValues, boundedEstimatedValues, out errorState);
+      }
 
       if (errorState != OnlineCalculatorError.None) return Double.NaN;
       else return mse;
-
     }
 

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionTreeInterpreter.cs

@@ -419 +419 @@
           }
         case OpCodes.Variable: {
-            if (row < 0 || row >= dataset.Rows)
-              return double.NaN;
+            if (row < 0 || row >= dataset.Rows) return double.NaN;
             var variableTreeNode = (VariableTreeNode)currentInstr.dynamicNode;
             return ((IList<double>)currentInstr.iArg0)[row] * variableTreeNode.Weight;
@@ -427 +426 @@
             var laggedVariableTreeNode = (LaggedVariableTreeNode)currentInstr.dynamicNode;
             int actualRow = row + laggedVariableTreeNode.Lag;
-            if (actualRow < 0 || actualRow >= dataset.Rows)
-              return double.NaN;
+            if (actualRow < 0 || actualRow >= dataset.Rows) return double.NaN;
             return ((IList<double>)currentInstr.iArg0)[actualRow] * laggedVariableTreeNode.Weight;
           }
@@ -439 +437 @@
         //to determine the relative amounts of the true and false branch see http://en.wikipedia.org/wiki/Logistic_function
         case OpCodes.VariableCondition: {
-            if (row < 0 || row >= dataset.Rows)
-              return double.NaN;
+            if (row < 0 || row >= dataset.Rows) return double.NaN;
             var variableConditionTreeNode = (VariableConditionTreeNode)currentInstr.dynamicNode;
             double variableValue = ((IList<double>)currentInstr.iArg0)[row];

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis.Views/3.4/HeuristicLab.Problems.DataAnalysis.Views-3.4.csproj

@@ -278 +278 @@
       <DependentUpon>RegressionSolutionView.cs</DependentUpon>
     </Compile>
-    <Compile Include="TimeSeriesPrognosis\TimeSeriesPrognosisSolutionLineChartView.cs">
-      <SubType>UserControl</SubType>
-    </Compile>
-    <Compile Include="TimeSeriesPrognosis\TimeSeriesPrognosisSolutionLineChartView.Designer.cs">
-      <DependentUpon>TimeSeriesPrognosisSolutionLineChartView.cs</DependentUpon>
-    </Compile>
-    <Compile Include="TimeSeriesPrognosis\TimeSeriesPrognosisSolutionPrognosedValuesView.cs">
-      <SubType>UserControl</SubType>
-    </Compile>
-    <Compile Include="TimeSeriesPrognosis\TimeSeriesPrognosisSolutionPrognosedValuesView.Designer.cs">
-      <DependentUpon>TimeSeriesPrognosisSolutionPrognosedValuesView.cs</DependentUpon>
-    </Compile>
     <None Include="HeuristicLab.snk" />
     <None Include="Plugin.cs.frame" />
@@ -312 +300 @@
     </BootstrapperPackage>
   </ItemGroup>
-  <ItemGroup />
+  <ItemGroup>
+    <Folder Include="TimeSeriesPrognosis\" />
+  </ItemGroup>
   <ItemGroup>
     <ProjectReference Include="..\..\HeuristicLab.Problems.DataAnalysis\3.4\HeuristicLab.Problems.DataAnalysis-3.4.csproj">

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/Models/TimeSeriesPrognosisAutoRegressiveModel.cs

@@ -22 +22 @@
 using System;
 using System.Collections.Generic;
+using System.Linq;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
@@ -65 +66 @@
         int row = rowsEnumerator.Current;
         int horizon = horizonsEnumerator.Current;
+        if (row - TimeOffset < 0) {
+          yield return Enumerable.Repeat(double.NaN, horizon);
+          continue;
+        }
+
         double[] prognosis = new double[horizon];
-
         for (int h = 0; h < horizon; h++) {
           double estimatedValue = 0.0;
-          for (int i = 1; i < TimeOffset; i++) {
+          for (int i = 1; i <= TimeOffset; i++) {
             int offset = h - i;
             if (offset >= 0) estimatedValue += prognosis[offset] * Phi[i - 1];
@@ -90 +95 @@
       foreach (int row in rows) {
         double estimatedValue = 0.0;
+        if (row - TimeOffset < 0) {
+          yield return double.NaN;
+          continue;
+        }
+
         for (int i = 1; i <= TimeOffset; i++) {
           estimatedValue += targetVariables[row - i] * Phi[i - 1];

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/Models/TimeSeriesPrognosisMovingAverageModel.cs

@@ -60 +60 @@
         int horizon = horizonsEnumerator.Current;
         int startIndex = row - WindowSize;
+        if (startIndex < 0) {
+          yield return Enumerable.Repeat(double.NaN, horizon);
+          continue;
+        }
 
-        if (startIndex < 0) startIndex = 0;
-        int count = row - startIndex - 1;
-
-        List<double> targetValues = dataset.GetDoubleValues(TargetVariable, Enumerable.Range(startIndex, count)).ToList();
+        List<double> targetValues = dataset.GetDoubleValues(TargetVariable, Enumerable.Range(startIndex, WindowSize)).ToList();
         int position = 0;
         for (int i = 0; i < horizon; i++) {
-          double prognosis = targetValues.GetRange(position, count).Average();
+          double prognosis = targetValues.GetRange(position, WindowSize).Average();
           targetValues.Add(prognosis);
-          if (count < WindowSize) count++;
-          else position++;
+          position++;
         }
         yield return targetValues.GetRange(targetValues.Count - horizon, horizon);
@@ -80 +80 @@
 
     public IEnumerable<double> GetEstimatedValues(Dataset dataset, IEnumerable<int> rows) {
-      return GetPrognosedValues(dataset, rows, rows.Select(r => 1)).SelectMany(e => e);
-    }
-    public IEnumerable<double> GetEstimatedValues(Dataset dataset, IEnumerable<int> rows, int x) {
       var targetValues = dataset.GetReadOnlyDoubleValues(TargetVariable).ToList();
       foreach (int row in rows) {
-        yield return targetValues.GetRange(row - WindowSize, WindowSize).Average();
+        if (row - WindowSize < 0) yield return double.NaN;
+        else yield return targetValues.GetRange(row - WindowSize, WindowSize).Average();
       }
     }
-
 
     public ITimeSeriesPrognosisSolution CreateTimeSeriesPrognosisSolution(ITimeSeriesPrognosisProblemData problemData) {

branches/HeuristicLab.TimeSeries/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisProblemData.cs

@@ -1578 +1578 @@
     }
 
-    public TimeSeriesPrognosisProblemData() : this(defaultDataset, defaultAllowedInputVariables, defaultTargetVariable) { }
+    public TimeSeriesPrognosisProblemData()
+      : this(defaultDataset, defaultAllowedInputVariables, defaultTargetVariable) {
+      TrainingPartition.Start = 50;
+    }
     public TimeSeriesPrognosisProblemData(Dataset dataset, IEnumerable<string> allowedInputVariables, string targetVariable)
       : base(dataset, allowedInputVariables, targetVariable) {
@@ -1587 +1590 @@
       TestHorizonParameter.Hidden = true;
 
+      TrainingPartition.Start = Math.Min((int)(TrainingPartition.Size * 0.2), 50);
+
       RegisterParameterEventHandlers();
     }