Free cookie consent management tool by TermsFeed Policy Generator

Ignore:
Timestamp:
11/28/11 13:47:28 (13 years ago)
Author:
sforsten
Message:

#1669: branch has been merged with the trunk in revision 7081 and methods in RegressionBenchmark have been renamed.

Location:
branches/RegressionBenchmarks
Files:
12 edited
2 copied

Legend:

Unmodified
Added
Removed

  • branches/RegressionBenchmarks

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/ClassificationEnsembleProblemData.cs

    r6672 r7085  
    4343
    4444    private static readonly ClassificationEnsembleProblemData emptyProblemData;
    45     public static ClassificationEnsembleProblemData EmptyProblemData {
     45    public static new ClassificationEnsembleProblemData EmptyProblemData {
    4646      get { return emptyProblemData; }
    4747    }

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/ClassificationEnsembleSolution.cs

    r6666 r7085  
    201201
    202202    public IEnumerable<IEnumerable<double>> GetEstimatedClassValueVectors(Dataset dataset, IEnumerable<int> rows) {
     203      if (!Model.Models.Any()) yield break;
    203204      var estimatedValuesEnumerators = (from model in Model.Models
    204205                                        select model.GetEstimatedClassValues(dataset, rows).GetEnumerator())

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/ClassificationProblem.cs

    r7025 r7085  
    2828  [Item("Classification Problem", "A general classification problem.")]
    2929  [Creatable("Problems")]
    30   public class ClassificationProblem : DataAnalysisProblem<IClassificationProblemData>, IClassificationProblem {
     30  public class ClassificationProblem : DataAnalysisProblem<IClassificationProblemData>, IClassificationProblem, IStorableContent {
     31    public string Filename { get; set; }
     32
    3133    [StorableConstructor]
    3234    protected ClassificationProblem(bool deserializing) : base(deserializing) { }

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/ClassificationSolutionBase.cs

    r6913 r7085  
    7777    }
    7878
     79    [StorableHook(HookType.AfterDeserialization)]
     80    private void AfterDeserialization() {
     81      if (!this.ContainsKey(TrainingNormalizedGiniCoefficientResultName))
     82        Add(new Result(TrainingNormalizedGiniCoefficientResultName, "Normalized Gini coefficient of the model on the training partition.", new DoubleValue()));
     83      if (!this.ContainsKey(TestNormalizedGiniCoefficientResultName))
     84        Add(new Result(TestNormalizedGiniCoefficientResultName, "Normalized Gini coefficient of the model on the test partition.", new DoubleValue()));
     85    }
     86
    7987    protected void CalculateResults() {
    8088      double[] estimatedTrainingClassValues = EstimatedTrainingClassValues.ToArray(); // cache values
    … …  
    8492
    8593      OnlineCalculatorError errorState;
    86       double trainingAccuracy = OnlineAccuracyCalculator.Calculate(estimatedTrainingClassValues, originalTrainingClassValues, out errorState);
     94      double trainingAccuracy = OnlineAccuracyCalculator.Calculate(originalTrainingClassValues, estimatedTrainingClassValues, out errorState);
    8795      if (errorState != OnlineCalculatorError.None) trainingAccuracy = double.NaN;
    88       double testAccuracy = OnlineAccuracyCalculator.Calculate(estimatedTestClassValues, originalTestClassValues, out errorState);
     96      double testAccuracy = OnlineAccuracyCalculator.Calculate(originalTestClassValues, estimatedTestClassValues, out errorState);
    8997      if (errorState != OnlineCalculatorError.None) testAccuracy = double.NaN;
    9098

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/DiscriminantFunctionClassificationSolutionBase.cs

    r6913 r7085  
    108108
    109109      OnlineCalculatorError errorState;
    110       double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     110      double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    111111      TrainingMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingMSE : double.NaN;
    112       double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     112      double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    113113      TestMeanSquaredError = errorState == OnlineCalculatorError.None ? testMSE : double.NaN;
    114114
    115       double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     115      double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    116116      TrainingRSquared = errorState == OnlineCalculatorError.None ? trainingR2 : double.NaN;
    117       double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     117      double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    118118      TestRSquared = errorState == OnlineCalculatorError.None ? testR2 : double.NaN;
    119119

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionEnsembleSolution.cs

    r6666 r7085  
    6464    private void AfterDeserialization() {
    6565      foreach (var model in Model.Models) {
    66         IRegressionProblemData problemData = (IRegressionProblemData) ProblemData.Clone();
     66        IRegressionProblemData problemData = (IRegressionProblemData)ProblemData.Clone();
    6767        problemData.TrainingPartition.Start = trainingPartitions[model].Start;
    6868        problemData.TrainingPartition.End = trainingPartitions[model].End;
    … …  
    202202
    203203    public IEnumerable<IEnumerable<double>> GetEstimatedValueVectors(Dataset dataset, IEnumerable<int> rows) {
     204      if (!Model.Models.Any()) yield break;
    204205      var estimatedValuesEnumerators = (from model in Model.Models
    205206                                        select model.GetEstimatedValues(dataset, rows).GetEnumerator())

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionProblem.cs

    r7025 r7085  
    2828  [Item("Regression Problem", "A general regression problem.")]
    2929  [Creatable("Problems")]
    30   public class RegressionProblem : DataAnalysisProblem<IRegressionProblemData>, IRegressionProblem {
     30  public class RegressionProblem : DataAnalysisProblem<IRegressionProblemData>, IRegressionProblem, IStorableContent {
     31    public string Filename { get; set; }
     32
    3133    [StorableConstructor]
    3234    protected RegressionProblem(bool deserializing) : base(deserializing) { }

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionSolutionBase.cs

    r6740 r7085  
    147147
    148148      OnlineCalculatorError errorState;
    149       double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     149      double trainingMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    150150      TrainingMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingMSE : double.NaN;
    151       double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     151      double testMSE = OnlineMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    152152      TestMeanSquaredError = errorState == OnlineCalculatorError.None ? testMSE : double.NaN;
    153153
    154       double trainingMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     154      double trainingMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    155155      TrainingMeanAbsoluteError = errorState == OnlineCalculatorError.None ? trainingMAE : double.NaN;
    156       double testMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     156      double testMAE = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    157157      TestMeanAbsoluteError = errorState == OnlineCalculatorError.None ? testMAE : double.NaN;
    158158
    159       double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     159      double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    160160      TrainingRSquared = errorState == OnlineCalculatorError.None ? trainingR2 : double.NaN;
    161       double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     161      double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    162162      TestRSquared = errorState == OnlineCalculatorError.None ? testR2 : double.NaN;
    163163
    164       double trainingRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     164      double trainingRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    165165      TrainingRelativeError = errorState == OnlineCalculatorError.None ? trainingRelError : double.NaN;
    166       double testRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     166      double testRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    167167      TestRelativeError = errorState == OnlineCalculatorError.None ? testRelError : double.NaN;
    168168
    169       double trainingNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     169      double trainingNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    170170      TrainingNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingNMSE : double.NaN;
    171       double testNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     171      double testNMSE = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    172172      TestNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? testNMSE : double.NaN;
    173173    }

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisProblemData.cs

    r6802 r7085  
    15651565      problemData.Parameters.Add(new FixedValueParameter<IntRange>(TestPartitionParameterName, "", (IntRange)new IntRange(0, 0).AsReadOnly()));
    15661566      problemData.Parameters.Add(new ConstrainedValueParameter<StringValue>(TargetVariableParameterName, new ItemSet<StringValue>()));
    1567       emptyProblemData = problemData;
     1567      emptyProblemData = problemData;     
    15681568    }
    15691569    #endregion

  • branches/RegressionBenchmarks/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisSolutionBase.cs

    r6802 r7085  
    4040    private const string TrainingNormalizedMeanSquaredErrorResultName = "Normalized mean squared error (training)";
    4141    private const string TestNormalizedMeanSquaredErrorResultName = "Normalized mean squared error (test)";
    42     private const string TrainingDirectionalSymmetryResultName = "Directional symmetry (training)";
    43     private const string TestDirectionalSymmetryResultName = "Directional symmetry (test)";
    44     private const string TrainingWeightedDirectionalSymmetryResultName = "Weighted directional symmetry (training)";
    45     private const string TestWeightedDirectionalSymmetryResultName = "Weighted directional symmetry (test)";
    46     private const string TrainingTheilsUStatisticResultName = "Theil's U (training)";
    47     private const string TestTheilsUStatisticResultName = "Theil's U (test)";
     42    private const string TrainingDirectionalSymmetryResultName = "Average directional symmetry (training)";
     43    private const string TestDirectionalSymmetryResultName = "Average directional symmetry (test)";
     44    private const string TrainingWeightedDirectionalSymmetryResultName = "Average weighted directional symmetry (training)";
     45    private const string TestWeightedDirectionalSymmetryResultName = "Average weighted directional symmetry (test)";
     46    private const string TrainingTheilsUStatisticResultName = "Average Theil's U (training)";
     47    private const string TestTheilsUStatisticResultName = "Average Theil's U (test)";
    4848
    4949    public new ITimeSeriesPrognosisModel Model {
    … …  
    146146      Add(new Result(TrainingNormalizedMeanSquaredErrorResultName, "Normalized mean of squared errors of the model on the training partition", new DoubleValue()));
    147147      Add(new Result(TestNormalizedMeanSquaredErrorResultName, "Normalized mean of squared errors of the model on the test partition", new DoubleValue()));
    148       Add(new Result(TrainingDirectionalSymmetryResultName, "The directional symmetry of the output of the model on the training partition", new DoubleValue()));
    149       Add(new Result(TestDirectionalSymmetryResultName, "The directional symmetry of the output of the model on the test partition", new DoubleValue()));
    150       Add(new Result(TrainingWeightedDirectionalSymmetryResultName, "The weighted directional symmetry of the output of the model on the training partition", new DoubleValue()));
    151       Add(new Result(TestWeightedDirectionalSymmetryResultName, "The weighted directional symmetry of the output of the model on the test partition", new DoubleValue()));
    152       Add(new Result(TrainingTheilsUStatisticResultName, "The Theil's U statistic of the output of the model on the training partition", new DoubleValue()));
    153       Add(new Result(TestTheilsUStatisticResultName, "The Theil's U statistic of the output of the model on the test partition", new DoubleValue()));
     148      Add(new Result(TrainingDirectionalSymmetryResultName, "The average directional symmetry of the forecasts of the model on the training partition", new PercentValue()));
     149      Add(new Result(TestDirectionalSymmetryResultName, "The average directional symmetry of the forecasts of the model on the test partition", new PercentValue()));
     150      Add(new Result(TrainingWeightedDirectionalSymmetryResultName, "The average weighted directional symmetry of the forecasts of the model on the training partition", new DoubleValue()));
     151      Add(new Result(TestWeightedDirectionalSymmetryResultName, "The average weighted directional symmetry of the forecasts of the model on the test partition", new DoubleValue()));
     152      Add(new Result(TrainingTheilsUStatisticResultName, "The average Theil's U statistic of the forecasts of the model on the training partition", new DoubleValue()));
     153      Add(new Result(TestTheilsUStatisticResultName, "The average Theil's U statistic of the forecasts of the model on the test partition", new DoubleValue()));
    154154    }
    155155
    … …  
    166166
    167167      OnlineCalculatorError errorState;
    168       double trainingMse = OnlineMeanSquaredErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     168      double trainingMse = OnlineMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    169169      TrainingMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingMse : double.NaN;
    170       double testMse = OnlineMeanSquaredErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     170      double testMse = OnlineMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    171171      TestMeanSquaredError = errorState == OnlineCalculatorError.None ? testMse : double.NaN;
    172172
    173       double trainingMae = OnlineMeanAbsoluteErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     173      double trainingMae = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    174174      TrainingMeanAbsoluteError = errorState == OnlineCalculatorError.None ? trainingMae : double.NaN;
    175       double testMae = OnlineMeanAbsoluteErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     175      double testMae = OnlineMeanAbsoluteErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    176176      TestMeanAbsoluteError = errorState == OnlineCalculatorError.None ? testMae : double.NaN;
    177177
    178       double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     178      double trainingR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    179179      TrainingRSquared = errorState == OnlineCalculatorError.None ? trainingR2 : double.NaN;
    180       double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     180      double testR2 = OnlinePearsonsRSquaredCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    181181      TestRSquared = errorState == OnlineCalculatorError.None ? testR2 : double.NaN;
    182182
    183       double trainingRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     183      double trainingRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    184184      TrainingRelativeError = errorState == OnlineCalculatorError.None ? trainingRelError : double.NaN;
    185       double testRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     185      double testRelError = OnlineMeanAbsolutePercentageErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    186186      TestRelativeError = errorState == OnlineCalculatorError.None ? testRelError : double.NaN;
    187187
    188       double trainingNmse = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     188      double trainingNmse = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTrainingValues, estimatedTrainingValues, out errorState);
    189189      TrainingNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? trainingNmse : double.NaN;
    190       double testNmse = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     190      double testNmse = OnlineNormalizedMeanSquaredErrorCalculator.Calculate(originalTestValues, estimatedTestValues, out errorState);
    191191      TestNormalizedMeanSquaredError = errorState == OnlineCalculatorError.None ? testNmse : double.NaN;
    192192
    193       double trainingDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     193      var startTrainingValues = originalTrainingValues;
     194      // each continuation is only one element long
     195      var actualContinuationsTraining = from x in originalTrainingValues.Skip(1)
     196                                        select Enumerable.Repeat(x, 1);
     197      // each forecast is only one elemnt long
     198      // disregards the first estimated value (we could include this again by extending the list of original values by one step to the left
     199      // this is the easier way
     200      var predictedContinuationsTraining = from x in estimatedTrainingValues.Skip(1)
     201                                           select Enumerable.Repeat(x, 1);
     202
     203      var startTestValues = originalTestValues;
     204      var actualContinuationsTest = from x in originalTestValues.Skip(1)
     205                                    select Enumerable.Repeat(x, 1);
     206      var predictedContinuationsTest = from x in estimatedTestValues.Skip(1)
     207                                       select Enumerable.Repeat(x, 1);
     208
     209      double trainingDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(startTrainingValues, actualContinuationsTraining, predictedContinuationsTraining, out errorState);
    194210      TrainingDirectionalSymmetry = errorState == OnlineCalculatorError.None ? trainingDirectionalSymmetry : double.NaN;
    195       double testDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     211      double testDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(startTestValues, actualContinuationsTest, predictedContinuationsTest, out errorState);
    196212      TestDirectionalSymmetry = errorState == OnlineCalculatorError.None ? testDirectionalSymmetry : double.NaN;
    197213
    198       double trainingWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     214      double trainingWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(startTrainingValues, actualContinuationsTraining, predictedContinuationsTraining, out errorState);
    199215      TrainingWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? trainingWeightedDirectionalSymmetry : double.NaN;
    200       double testWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     216      double testWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(startTestValues, actualContinuationsTest, predictedContinuationsTest, out errorState);
    201217      TestWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? testWeightedDirectionalSymmetry : double.NaN;
    202218
    203       double trainingTheilsU = OnlineTheilsUStatisticCalculator.Calculate(estimatedTrainingValues, originalTrainingValues, out errorState);
     219      double trainingTheilsU = OnlineTheilsUStatisticCalculator.Calculate(startTrainingValues, actualContinuationsTraining, predictedContinuationsTraining, out errorState);
    204220      TrainingTheilsUStatistic = errorState == OnlineCalculatorError.None ? trainingTheilsU : double.NaN;
    205       double testTheilsU = OnlineTheilsUStatisticCalculator.Calculate(estimatedTestValues, originalTestValues, out errorState);
     221      double testTheilsU = OnlineTheilsUStatisticCalculator.Calculate(startTestValues, actualContinuationsTest, predictedContinuationsTest, out errorState);
    206222      TestTheilsUStatistic = errorState == OnlineCalculatorError.None ? testTheilsU : double.NaN;
    207 
    208 
    209223    }
    210224  }
Note: See TracChangeset for help on using the changeset viewer.