Free cookie consent management tool by TermsFeed Policy Generator

source: branches/DataAnalysis/HeuristicLab.Problems.DataAnalysis.MultiVariate.TimeSeriesPrognosis/3.3/Symbolic/Evaluators/SymbolicTimeSeriesPrognosisMahalanobisEvaluator.cs @ 5955

Last change on this file since 5955 was 5275, checked in by gkronber, 14 years ago

Merged changes from trunk to data analysis exploration branch and added fractional distance metric evaluator. #1142

File size: 7.8 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2010 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
4 *
5 * This file is part of HeuristicLab.
6 *
7 * HeuristicLab is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
11 *
12 * HeuristicLab is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15 * GNU General Public License for more details.
16 *
17 * You should have received a copy of the GNU General Public License
18 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
19 */
20#endregion
21
22using System;
23using System.Linq;
24using HeuristicLab.Common;
25using HeuristicLab.Core;
26using HeuristicLab.Data;
27using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
28using HeuristicLab.Problems.DataAnalysis.SupportVectorMachine;
29using HeuristicLab.Problems.DataAnalysis;
30using HeuristicLab.Problems.DataAnalysis.Evaluators;
31using HeuristicLab.Parameters;
32using HeuristicLab.Optimization;
33using HeuristicLab.Operators;
34using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic;
35using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
36using System.Collections.Generic;
37using HeuristicLab.Problems.DataAnalysis.Regression;
38using HeuristicLab.Problems.DataAnalysis.MultiVariate.TimeSeriesPrognosis.Symbolic.Interfaces;
39using HeuristicLab.Problems.DataAnalysis.MultiVariate.Evaluators;
40using HeuristicLab.Problems.DataAnalysis.Symbolic;
41
42namespace HeuristicLab.Problems.DataAnalysis.MultiVariate.TimeSeriesPrognosis.Symbolic.Evaluators {
43  [Item("SymbolicTimeSeriesPrognosisMahalanobisEvaluator", "")]
44  [StorableClass]
45  public class SymbolicTimeSeriesPrognosisMahalanobisEvaluator : SymbolicTimeSeriesPrognosisEvaluator {
46
47    [StorableConstructor]
48    protected SymbolicTimeSeriesPrognosisMahalanobisEvaluator(bool deserializing) : base(deserializing) { }
49    protected SymbolicTimeSeriesPrognosisMahalanobisEvaluator(SymbolicTimeSeriesPrognosisMahalanobisEvaluator original, Cloner cloner)
50      : base(original, cloner) {
51    }
52    public SymbolicTimeSeriesPrognosisMahalanobisEvaluator()
53      : base() {
54    }
55    public override IDeepCloneable Clone(Cloner cloner) {
56      return new SymbolicTimeSeriesPrognosisMahalanobisEvaluator(this, cloner);
57    }
58    public override double Evaluate(SymbolicExpressionTree tree, MultiVariateDataAnalysisProblemData problemData, ISymbolicTimeSeriesExpressionInterpreter interpreter, IEnumerable<int> rows, int predictionHorizon, DoubleArray lowerEstimationLimit, DoubleArray upperEstimationLimit) {
59      return Calculate(tree, problemData, interpreter, rows, predictionHorizon, lowerEstimationLimit, upperEstimationLimit);
60    }
61
62    public static double Calculate(SymbolicExpressionTree tree, MultiVariateDataAnalysisProblemData problemData,
63      ISymbolicTimeSeriesExpressionInterpreter interpreter,
64      IEnumerable<int> rows, int predictionHorizon,
65      DoubleArray lowerEstimationLimit, DoubleArray upperEstimationLimit) {
66      double[] alpha, beta;
67      double quality;
68
69      Dataset dataset = problemData.Dataset;
70      // calculate scaling parameters based on one-step-predictions
71      IEnumerable<string> selectedTargetVariables = (from item in problemData.TargetVariables
72                                                     where problemData.TargetVariables.ItemChecked(item)
73                                                     select item.Value).ToArray();
74      int dimension = selectedTargetVariables.Count();
75
76      IEnumerable<int> selectedTargetVariableIndexes = (from targetVariable in selectedTargetVariables
77                                                        select dataset.GetVariableIndex(targetVariable)).ToArray();
78      IEnumerable<IEnumerable<double>> oneStepPredictions =
79        interpreter.GetSymbolicExpressionTreeValues(tree, problemData.Dataset, selectedTargetVariables, rows, 1)
80        .Cast<IEnumerable<double>>();
81      IEnumerable<IEnumerable<double>> originalValues = from row in rows
82                                                        select (from targetVariableIndex in selectedTargetVariableIndexes
83                                                                select dataset[row, targetVariableIndex]);
84      alpha = new double[dimension];
85      beta = new double[dimension];
86
87      SymbolicTimeSeriesPrognosisScaledNormalizedMseEvaluator.CalculateScalingParameters(originalValues, oneStepPredictions, ref beta, ref alpha);
88
89      // calculate the quality for the full horizon
90      quality = CalculateWithScaling(tree, problemData, interpreter,
91        rows, predictionHorizon,
92        lowerEstimationLimit, upperEstimationLimit,
93        beta, alpha);
94      return quality;
95
96    }
97
98    public static double CalculateWithScaling(SymbolicExpressionTree tree, MultiVariateDataAnalysisProblemData problemData,
99      ISymbolicTimeSeriesExpressionInterpreter interpreter,
100      IEnumerable<int> rows, int predictionHorizon,
101      DoubleArray lowerEstimationLimit, DoubleArray upperEstimationLimit,
102      double[] beta, double[] alpha) {
103
104      Dataset dataset = problemData.Dataset;
105
106      IEnumerable<string> selectedTargetVariables = (from targetVariable in problemData.TargetVariables
107                                                     where problemData.TargetVariables.ItemChecked(targetVariable)
108                                                     select targetVariable.Value).ToArray();
109
110      IEnumerable<int> selectedTargetVariableIndexes = (from targetVariable in selectedTargetVariables
111                                                        select dataset.GetVariableIndex(targetVariable)).ToArray();
112      IEnumerable<double[]> estimatedValues =
113        interpreter.GetScaledSymbolicExpressionTreeValues(tree, dataset, selectedTargetVariables,
114        rows, predictionHorizon, beta, alpha);
115
116      IEnumerable<IEnumerable<double>> originalValues = from row in rows
117                                                        from step in Enumerable.Range(0, predictionHorizon)
118                                                        select (from targetVariableIndex in selectedTargetVariableIndexes
119                                                                select dataset[row + step, targetVariableIndex]);
120
121      OnlineMeanMahalanobisDistanceEvaluator evaluator = new OnlineMeanMahalanobisDistanceEvaluator();
122      // for covariance calculation: array of variables
123      IEnumerable<double>[] targetValues = (from targetVariableIndex in selectedTargetVariableIndexes
124                                            select dataset.GetEnumeratedVariableValues(targetVariableIndex, rows))
125                                           .ToArray();
126      evaluator.InitializeCovarianceMatrixFromSamples(targetValues);
127
128      var estimatedValuesEnumerator = estimatedValues.GetEnumerator();
129      var originalValuesEnumerator = originalValues.GetEnumerator();
130      while (originalValuesEnumerator.MoveNext() & estimatedValuesEnumerator.MoveNext()) {
131        IEnumerable<double> currentOriginal = originalValuesEnumerator.Current;
132        double[] currentEstimated = estimatedValuesEnumerator.Current;
133        // limit estimated values to bounds
134        for (int i = 0; i < currentEstimated.Length; i++) {
135          if (double.IsNaN(currentEstimated[i])) currentEstimated[i] = upperEstimationLimit[i];
136          else currentEstimated[i] = Math.Min(upperEstimationLimit[i], Math.Max(lowerEstimationLimit[i], currentEstimated[i]));
137        }
138
139        evaluator.Add(currentOriginal, currentEstimated);
140      }
141
142      return evaluator.MeanGeneralizedSquaredInterpointDistance;
143    }
144  }
145}
Note: See TracBrowser for help on using the repository browser.