Free cookie consent management tool by TermsFeed Policy Generator

source: branches/plugins/HeuristicLab.GP.StructureIdentification.TimeSeries/3.2/TheilInequalityCoefficientEvaluator.cs @ 2950

Last change on this file since 2950 was 721, checked in by gkronber, 16 years ago

fixed a typo

File size: 6.0 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2008 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.Collections.Generic;
24using System.Linq;
25using System.Text;
26using HeuristicLab.Core;
27using HeuristicLab.Data;
28using HeuristicLab.GP.StructureIdentification;
29using HeuristicLab.DataAnalysis;
30
31namespace HeuristicLab.GP.StructureIdentification.TimeSeries {
32  public class TheilInequalityCoefficientEvaluator : GPEvaluatorBase {
33    public override string Description {
34      get {
35        return @"Evaluates 'FunctionTree' for all samples of 'Dataset' and calculates
36the 'Theil inequality coefficient (Theil's U2 not U1!)' of estimated values vs. real values of 'TargetVariable'.
37
38U2 = Sqrt(1/N * Sum(P_t - A_t)^2 ) / Sqrt(1/N * Sum(A_t)^2 )
39
40where P_t is the predicted change of the target variable and A_t is the measured (original) change.
41(P_t = y'_t - y_(t-1), A_t = y_t - y_(t-1)).
42
43U2 is 0 for a perfect prediction and 1 for the naive model y'_t = y_(t-1). An U2 > 1 means the
44model is worse than the naive model (=> model is useless).";
45      }
46    }
47
48    public TheilInequalityCoefficientEvaluator()
49      : base() {
50      AddVariableInfo(new VariableInfo("TheilInequalityCoefficient", "Theil's inequality coefficient (U2) of the model", typeof(DoubleData), VariableKind.New));
51      AddVariableInfo(new VariableInfo("TheilInequalityCoefficientBias", "Bias proportion of Theil's inequality coefficient", typeof(DoubleData), VariableKind.New));
52      AddVariableInfo(new VariableInfo("TheilInequalityCoefficientVariance", "Variance proportion of Theil's inequality coefficient", typeof(DoubleData), VariableKind.New));
53      AddVariableInfo(new VariableInfo("TheilInequalityCoefficientCovariance", "Covariance proportion of Theil's inequality coefficient", typeof(DoubleData), VariableKind.New));
54    }
55
56    public override void Evaluate(IScope scope, BakedTreeEvaluator evaluator, Dataset dataset, int targetVariable, int start, int end, bool updateTargetValues) {
57      #region create result variables
58      DoubleData theilInequaliy = GetVariableValue<DoubleData>("TheilInequalityCoefficient", scope, false, false);
59      if (theilInequaliy == null) {
60        theilInequaliy = new DoubleData();
61        scope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName("TheilInequalityCoefficient"), theilInequaliy));
62      }
63      DoubleData uBias = GetVariableValue<DoubleData>("TheilInequalityCoefficientBias", scope, false, false);
64      if (uBias == null) {
65        uBias = new DoubleData();
66        scope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName("TheilInequalityCoefficientBias"), uBias));
67      }
68      DoubleData uVariance = GetVariableValue<DoubleData>("TheilInequalityCoefficientVariance", scope, false, false);
69      if (uVariance == null) {
70        uVariance = new DoubleData();
71        scope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName("TheilInequalityCoefficientVariance"), uVariance));
72      }
73      DoubleData uCovariance = GetVariableValue<DoubleData>("TheilInequalityCoefficientCovariance", scope, false, false);
74      if (uCovariance == null) {
75        uCovariance = new DoubleData();
76        scope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName("TheilInequalityCoefficientCovariance"), uCovariance));
77      }
78      #endregion
79
80      double errorsSquaredSum = 0.0;
81      double originalSquaredSum = 0.0;
82      double[] estimatedChanges = new double[end - start];
83      double[] originalChanges = new double[end - start];
84      int nSamples = 0;
85      for (int sample = start; sample < end; sample++) {
86        double prevValue = dataset.GetValue(sample - 1, targetVariable);
87        double estimatedChange = evaluator.Evaluate(sample) - prevValue;
88        double originalChange = dataset.GetValue(sample, targetVariable) - prevValue;
89        if (updateTargetValues) {
90          dataset.SetValue(sample, targetVariable, estimatedChange + prevValue);
91        }
92        if (!double.IsNaN(originalChange) && !double.IsInfinity(originalChange)) {
93          double error = estimatedChange - originalChange;
94          errorsSquaredSum += error * error;
95          originalSquaredSum += originalChange * originalChange;
96          estimatedChanges[sample - start] = estimatedChange;
97          originalChanges[sample - start] = originalChange;
98          nSamples++;
99        }
100      }
101      double quality = Math.Sqrt(errorsSquaredSum / nSamples) / Math.Sqrt(originalSquaredSum / nSamples);
102      if (double.IsNaN(quality) || double.IsInfinity(quality))
103        quality = double.MaxValue;
104      theilInequaliy.Data = quality; // U2
105
106      // decomposition into U_bias + U_variance + U_covariance parts
107      double bias = Statistics.Mean(estimatedChanges) - Statistics.Mean(originalChanges);
108      bias *= bias; // squared
109      uBias.Data = bias / (errorsSquaredSum / nSamples);
110
111      double variance = Statistics.StandardDeviation(estimatedChanges) - Statistics.StandardDeviation(originalChanges);
112      variance *= variance; // squared
113      uVariance.Data = variance / (errorsSquaredSum / nSamples);
114
115      // all parts add up to one so I don't have to calculate the correlation coefficient for the covariance proportion
116      uCovariance.Data = 1.0 - uBias.Data - uVariance.Data;
117    }
118  }
119}
Note: See TracBrowser for help on using the repository browser.