- Timestamp:
- 06/20/14 17:08:41 (10 years ago)
- Location:
- trunk/sources/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis
- Files:
-
- 3 edited
Legend:
- Unmodified
- Added
- Removed
-
trunk/sources/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisProblemData.cs
r10540 r11031 1621 1621 } 1622 1622 1623 protected override bool IsProblemDataCompatible(IDataAnalysisProblemData problemData, out string errorMessage) { 1624 if (problemData == null) throw new ArgumentNullException("problemData", "The provided problemData is null."); 1625 ITimeSeriesPrognosisProblemData timeseriesProblemData = problemData as ITimeSeriesPrognosisProblemData; 1626 if (timeseriesProblemData == null) 1627 throw new ArgumentException("The problem data is not a time-series problem data. Instead a " + problemData.GetType().GetPrettyName() + " was provided.", "problemData"); 1628 1629 var returnValue = base.IsProblemDataCompatible(problemData, out errorMessage); 1630 //check targetVariable 1631 if (problemData.InputVariables.All(var => var.Value != TargetVariable)) { 1632 errorMessage = string.Format("The target variable {0} is not present in the new problem data.", TargetVariable) 1633 + Environment.NewLine + errorMessage; 1634 return false; 1635 } 1636 return returnValue; 1637 } 1638 1623 1639 public override void AdjustProblemDataProperties(IDataAnalysisProblemData problemData) { 1624 1640 TimeSeriesPrognosisProblemData timeSeriesProblemData = problemData as TimeSeriesPrognosisProblemData; … … 1626 1642 throw new ArgumentException("The problem data is not a timeseries problem data. Instead a " + problemData.GetType().GetPrettyName() + " was provided.", "problemData"); 1627 1643 1644 var trainingDataStart = TrainingIndices.First(); 1645 1628 1646 base.AdjustProblemDataProperties(problemData); 1647 1648 TestPartition.Start = trainingDataStart; 1629 1649 1630 1650 TrainingHorizon = timeSeriesProblemData.TrainingHorizon; -
trunk/sources/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisResults.cs
r9456 r11031 369 369 OnlineCalculatorError errorState; 370 370 var problemData = Solution.ProblemData; 371 if (!problemData.TrainingIndices.Any()) return; 371 372 var model = Solution.Model; 372 373 //mean model … … 415 416 OnlineCalculatorError errorState; 416 417 var problemData = Solution.ProblemData; 418 if (!problemData.TestIndices.Any()) return; 417 419 var model = Solution.Model; 418 //mean model419 double trainingMean = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TrainingIndices).Average();420 var meanModel = new ConstantTimeSeriesPrognosisModel(trainingMean);421 422 //AR1 model423 double alpha, beta;424 IEnumerable<double> trainingStartValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TrainingIndices.Select(r => r - 1).Where(r => r > 0)).ToList();425 OnlineLinearScalingParameterCalculator.Calculate(problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TrainingIndices.Where(x => x > 0)), trainingStartValues, out alpha, out beta, out errorState);426 var AR1model = new TimeSeriesPrognosisAutoRegressiveModel(problemData.TargetVariable, new double[] { beta }, alpha);427 428 420 var testHorizions = problemData.TestIndices.Select(r => Math.Min(testHorizon, problemData.TestPartition.End - r)).ToList(); 429 421 IEnumerable<IEnumerable<double>> testTargetValues = problemData.TestIndices.Zip(testHorizions, Enumerable.Range).Select(r => problemData.Dataset.GetDoubleValues(problemData.TargetVariable, r)).ToList(); 430 422 IEnumerable<IEnumerable<double>> testEstimatedValues = model.GetPrognosedValues(problemData.Dataset, problemData.TestIndices, testHorizions).ToList(); 431 423 IEnumerable<double> testStartValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TestIndices.Select(r => r - 1).Where(r => r > 0)).ToList(); 432 IEnumerable<IEnumerable<double>> testMeanModelPredictions = meanModel.GetPrognosedValues(problemData.Dataset, problemData.TestIndices, testHorizions).ToList();433 IEnumerable<IEnumerable<double>> testAR1ModelPredictions = AR1model.GetPrognosedValues(problemData.Dataset, problemData.TestIndices, testHorizions).ToList();434 424 435 425 IEnumerable<double> originalTestValues = testTargetValues.SelectMany(x => x).ToList(); … … 453 443 PrognosisTestWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(testStartValues, testTargetValues, testEstimatedValues, out errorState); 454 444 PrognosisTestWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? PrognosisTestWeightedDirectionalSymmetry : 0.0; 455 PrognosisTestTheilsUStatisticAR1 = OnlineTheilsUStatisticCalculator.Calculate(testStartValues, testTargetValues, testAR1ModelPredictions, testEstimatedValues, out errorState); 456 PrognosisTestTheilsUStatisticAR1 = errorState == OnlineCalculatorError.None ? PrognosisTestTheilsUStatisticAR1 : double.PositiveInfinity; 457 PrognosisTestTheilsUStatisticMean = OnlineTheilsUStatisticCalculator.Calculate(testStartValues, testTargetValues, testMeanModelPredictions, testEstimatedValues, out errorState); 458 PrognosisTestTheilsUStatisticMean = errorState == OnlineCalculatorError.None ? PrognosisTestTheilsUStatisticMean : double.PositiveInfinity; 445 446 447 if (problemData.TrainingIndices.Any()) { 448 //mean model 449 double trainingMean = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TrainingIndices).Average(); 450 var meanModel = new ConstantTimeSeriesPrognosisModel(trainingMean); 451 452 //AR1 model 453 double alpha, beta; 454 IEnumerable<double> trainingStartValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TrainingIndices.Select(r => r - 1).Where(r => r > 0)).ToList(); 455 OnlineLinearScalingParameterCalculator.Calculate(problemData.Dataset.GetDoubleValues(problemData.TargetVariable, problemData.TrainingIndices.Where(x => x > 0)), trainingStartValues, out alpha, out beta, out errorState); 456 var AR1model = new TimeSeriesPrognosisAutoRegressiveModel(problemData.TargetVariable, new double[] { beta }, alpha); 457 458 IEnumerable<IEnumerable<double>> testMeanModelPredictions = meanModel.GetPrognosedValues(problemData.Dataset, problemData.TestIndices, testHorizions).ToList(); 459 IEnumerable<IEnumerable<double>> testAR1ModelPredictions = AR1model.GetPrognosedValues(problemData.Dataset, problemData.TestIndices, testHorizions).ToList(); 460 461 PrognosisTestTheilsUStatisticAR1 = OnlineTheilsUStatisticCalculator.Calculate(testStartValues, testTargetValues, testAR1ModelPredictions, testEstimatedValues, out errorState); 462 PrognosisTestTheilsUStatisticAR1 = errorState == OnlineCalculatorError.None ? PrognosisTestTheilsUStatisticAR1 : double.PositiveInfinity; 463 PrognosisTestTheilsUStatisticMean = OnlineTheilsUStatisticCalculator.Calculate(testStartValues, testTargetValues, testMeanModelPredictions, testEstimatedValues, out errorState); 464 PrognosisTestTheilsUStatisticMean = errorState == OnlineCalculatorError.None ? PrognosisTestTheilsUStatisticMean : double.PositiveInfinity; 465 } 459 466 } 460 467 } -
trunk/sources/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/TimeSeriesPrognosis/TimeSeriesPrognosisSolutionBase.cs
r9462 r11031 149 149 protected void CalculateTimeSeriesResults() { 150 150 OnlineCalculatorError errorState; 151 double trainingMean = ProblemData. Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices).Average();151 double trainingMean = ProblemData.TrainingIndices.Any() ? ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices).Average() : double.NaN; 152 152 var meanModel = new ConstantTimeSeriesPrognosisModel(trainingMean); 153 153 … … 159 159 160 160 #region Calculate training quality measures 161 IEnumerable<double> trainingTargetValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices).ToList(); 162 IEnumerable<double> trainingEstimatedValues = EstimatedTrainingValues.ToList(); 163 IEnumerable<double> trainingMeanModelPredictions = meanModel.GetEstimatedValues(ProblemData.Dataset, ProblemData.TrainingIndices).ToList(); 164 IEnumerable<double> trainingAR1ModelPredictions = AR1model.GetEstimatedValues(ProblemData.Dataset, ProblemData.TrainingIndices).ToList(); 165 166 TrainingDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingEstimatedValues, out errorState); 167 TrainingDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TrainingDirectionalSymmetry : 0.0; 168 TrainingWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingEstimatedValues, out errorState); 169 TrainingWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TrainingWeightedDirectionalSymmetry : 0.0; 170 TrainingTheilsUStatisticAR1 = OnlineTheilsUStatisticCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingAR1ModelPredictions, trainingEstimatedValues, out errorState); 171 TrainingTheilsUStatisticAR1 = errorState == OnlineCalculatorError.None ? TrainingTheilsUStatisticAR1 : double.PositiveInfinity; 172 TrainingTheilsUStatisticMean = OnlineTheilsUStatisticCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingMeanModelPredictions, trainingEstimatedValues, out errorState); 173 TrainingTheilsUStatisticMean = errorState == OnlineCalculatorError.None ? TrainingTheilsUStatisticMean : double.PositiveInfinity; 161 if (ProblemData.TrainingIndices.Any()) { 162 IEnumerable<double> trainingTargetValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices).ToList(); 163 IEnumerable<double> trainingEstimatedValues = EstimatedTrainingValues.ToList(); 164 IEnumerable<double> trainingMeanModelPredictions = meanModel.GetEstimatedValues(ProblemData.Dataset, ProblemData.TrainingIndices).ToList(); 165 IEnumerable<double> trainingAR1ModelPredictions = AR1model.GetEstimatedValues(ProblemData.Dataset, ProblemData.TrainingIndices).ToList(); 166 167 TrainingDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingEstimatedValues, out errorState); 168 TrainingDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TrainingDirectionalSymmetry : 0.0; 169 TrainingWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingEstimatedValues, out errorState); 170 TrainingWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TrainingWeightedDirectionalSymmetry : 0.0; 171 TrainingTheilsUStatisticAR1 = OnlineTheilsUStatisticCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingAR1ModelPredictions, trainingEstimatedValues, out errorState); 172 TrainingTheilsUStatisticAR1 = errorState == OnlineCalculatorError.None ? TrainingTheilsUStatisticAR1 : double.PositiveInfinity; 173 TrainingTheilsUStatisticMean = OnlineTheilsUStatisticCalculator.Calculate(trainingTargetValues.First(), trainingTargetValues, trainingMeanModelPredictions, trainingEstimatedValues, out errorState); 174 TrainingTheilsUStatisticMean = errorState == OnlineCalculatorError.None ? TrainingTheilsUStatisticMean : double.PositiveInfinity; 175 } 174 176 #endregion 175 177 176 178 #region Calculate test quality measures 177 IEnumerable<double> testTargetValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices).ToList(); 178 IEnumerable<double> testEstimatedValues = EstimatedTestValues.ToList(); 179 IEnumerable<double> testMeanModelPredictions = meanModel.GetEstimatedValues(ProblemData.Dataset, ProblemData.TestIndices).ToList(); 180 IEnumerable<double> testAR1ModelPredictions = AR1model.GetEstimatedValues(ProblemData.Dataset, ProblemData.TestIndices).ToList(); 181 182 TestDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(testTargetValues.First(), testTargetValues, testEstimatedValues, out errorState); 183 TestDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TestDirectionalSymmetry : 0.0; 184 TestWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(testTargetValues.First(), testTargetValues, testEstimatedValues, out errorState); 185 TestWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TestWeightedDirectionalSymmetry : 0.0; 186 TestTheilsUStatisticAR1 = OnlineTheilsUStatisticCalculator.Calculate(testTargetValues.First(), testTargetValues, testAR1ModelPredictions, testEstimatedValues, out errorState); 187 TestTheilsUStatisticAR1 = errorState == OnlineCalculatorError.None ? TestTheilsUStatisticAR1 : double.PositiveInfinity; 188 TestTheilsUStatisticMean = OnlineTheilsUStatisticCalculator.Calculate(testTargetValues.First(), testTargetValues, testMeanModelPredictions, testEstimatedValues, out errorState); 189 TestTheilsUStatisticMean = errorState == OnlineCalculatorError.None ? TestTheilsUStatisticMean : double.PositiveInfinity; 179 if (ProblemData.TestIndices.Any()) { 180 IEnumerable<double> testTargetValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices).ToList(); 181 IEnumerable<double> testEstimatedValues = EstimatedTestValues.ToList(); 182 IEnumerable<double> testMeanModelPredictions = meanModel.GetEstimatedValues(ProblemData.Dataset, ProblemData.TestIndices).ToList(); 183 IEnumerable<double> testAR1ModelPredictions = AR1model.GetEstimatedValues(ProblemData.Dataset, ProblemData.TestIndices).ToList(); 184 185 TestDirectionalSymmetry = OnlineDirectionalSymmetryCalculator.Calculate(testTargetValues.First(), testTargetValues, testEstimatedValues, out errorState); 186 TestDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TestDirectionalSymmetry : 0.0; 187 TestWeightedDirectionalSymmetry = OnlineWeightedDirectionalSymmetryCalculator.Calculate(testTargetValues.First(), testTargetValues, testEstimatedValues, out errorState); 188 TestWeightedDirectionalSymmetry = errorState == OnlineCalculatorError.None ? TestWeightedDirectionalSymmetry : 0.0; 189 TestTheilsUStatisticAR1 = OnlineTheilsUStatisticCalculator.Calculate(testTargetValues.First(), testTargetValues, testAR1ModelPredictions, testEstimatedValues, out errorState); 190 TestTheilsUStatisticAR1 = errorState == OnlineCalculatorError.None ? TestTheilsUStatisticAR1 : double.PositiveInfinity; 191 TestTheilsUStatisticMean = OnlineTheilsUStatisticCalculator.Calculate(testTargetValues.First(), testTargetValues, testMeanModelPredictions, testEstimatedValues, out errorState); 192 TestTheilsUStatisticMean = errorState == OnlineCalculatorError.None ? TestTheilsUStatisticMean : double.PositiveInfinity; 193 } 190 194 #endregion 191 195 }
Note: See TracChangeset
for help on using the changeset viewer.