[3532] | 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 |
|
---|
| 22 | using System;
|
---|
| 23 | using System.Collections.Generic;
|
---|
| 24 | using System.Linq;
|
---|
| 25 | using System.Drawing;
|
---|
| 26 | using HeuristicLab.Common;
|
---|
| 27 | using HeuristicLab.Core;
|
---|
| 28 | using HeuristicLab.Data;
|
---|
| 29 | using HeuristicLab.Optimization;
|
---|
| 30 | using HeuristicLab.Parameters;
|
---|
| 31 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
| 32 | using HeuristicLab.PluginInfrastructure;
|
---|
| 33 | using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
|
---|
| 34 | using HeuristicLab.Problems.DataAnalysis;
|
---|
| 35 | using HeuristicLab.Operators;
|
---|
| 36 | using HeuristicLab.Problems.DataAnalysis.Evaluators;
|
---|
| 37 | using HeuristicLab.Problems.DataAnalysis.Symbolic;
|
---|
| 38 |
|
---|
| 39 | namespace HeuristicLab.Problems.DataAnalysis.Regression.Symbolic {
|
---|
[4044] | 40 | [Item("SymbolicRegressionScaledMeanAndVarianceSquaredErrorEvaluator", "Calculates the mean and the variance of the squared errors of a linearly scaled symbolic regression solution.")]
|
---|
[3532] | 41 | [StorableClass]
|
---|
[4044] | 42 | public class SymbolicRegressionScaledMeanAndVarianceSquaredErrorEvaluator : SymbolicRegressionMeanSquaredErrorEvaluator {
|
---|
| 43 | private const string QualityVarianceParameterName = "QualityVariance";
|
---|
| 44 | private const string QualitySamplesParameterName = "QualitySamples";
|
---|
[3532] | 45 |
|
---|
| 46 | #region parameter properties
|
---|
| 47 | public ILookupParameter<DoubleValue> AlphaParameter {
|
---|
| 48 | get { return (ILookupParameter<DoubleValue>)Parameters["Alpha"]; }
|
---|
| 49 | }
|
---|
| 50 | public ILookupParameter<DoubleValue> BetaParameter {
|
---|
| 51 | get { return (ILookupParameter<DoubleValue>)Parameters["Beta"]; }
|
---|
| 52 | }
|
---|
[4044] | 53 | public ILookupParameter<DoubleValue> QualityVarianceParameter {
|
---|
| 54 | get { return (ILookupParameter<DoubleValue>)Parameters[QualityVarianceParameterName]; }
|
---|
| 55 | }
|
---|
| 56 | public ILookupParameter<IntValue> QualitySamplesParameter {
|
---|
| 57 | get { return (ILookupParameter<IntValue>)Parameters[QualitySamplesParameterName]; }
|
---|
| 58 | }
|
---|
| 59 |
|
---|
[3532] | 60 | #endregion
|
---|
| 61 | #region properties
|
---|
| 62 | public DoubleValue Alpha {
|
---|
| 63 | get { return AlphaParameter.ActualValue; }
|
---|
| 64 | set { AlphaParameter.ActualValue = value; }
|
---|
| 65 | }
|
---|
| 66 | public DoubleValue Beta {
|
---|
| 67 | get { return BetaParameter.ActualValue; }
|
---|
| 68 | set { BetaParameter.ActualValue = value; }
|
---|
| 69 | }
|
---|
[4044] | 70 | public DoubleValue QualityVariance {
|
---|
| 71 | get { return QualityVarianceParameter.ActualValue; }
|
---|
| 72 | set { QualityVarianceParameter.ActualValue = value; }
|
---|
| 73 | }
|
---|
| 74 | public IntValue QualitySamples {
|
---|
| 75 | get { return QualitySamplesParameter.ActualValue; }
|
---|
| 76 | set { QualitySamplesParameter.ActualValue = value; }
|
---|
| 77 | }
|
---|
[3532] | 78 | #endregion
|
---|
[4044] | 79 | public SymbolicRegressionScaledMeanAndVarianceSquaredErrorEvaluator()
|
---|
[3532] | 80 | : base() {
|
---|
| 81 | Parameters.Add(new LookupParameter<DoubleValue>("Alpha", "Alpha parameter for linear scaling of the estimated values."));
|
---|
| 82 | Parameters.Add(new LookupParameter<DoubleValue>("Beta", "Beta parameter for linear scaling of the estimated values."));
|
---|
[4044] | 83 | Parameters.Add(new LookupParameter<DoubleValue>(QualityVarianceParameterName, "A parameter which stores the variance of the squared errors."));
|
---|
| 84 | Parameters.Add(new LookupParameter<IntValue>(QualitySamplesParameterName, " The number of evaluated samples."));
|
---|
[3532] | 85 | }
|
---|
| 86 |
|
---|
[4034] | 87 | protected override double Evaluate(ISymbolicExpressionTreeInterpreter interpreter, SymbolicExpressionTree solution, Dataset dataset, StringValue targetVariable, IEnumerable<int> rows) {
|
---|
[3532] | 88 | double alpha, beta;
|
---|
[4044] | 89 | double meanSE, varianceSE;
|
---|
| 90 | int count;
|
---|
| 91 | double mse = Calculate(interpreter, solution, LowerEstimationLimit.Value, UpperEstimationLimit.Value, dataset, targetVariable.Value, rows, out beta, out alpha, out meanSE, out varianceSE, out count);
|
---|
| 92 | Alpha = new DoubleValue(alpha);
|
---|
| 93 | Beta = new DoubleValue(beta);
|
---|
| 94 | QualityVariance = new DoubleValue(varianceSE);
|
---|
| 95 | QualitySamples = new IntValue(count);
|
---|
[3532] | 96 | return mse;
|
---|
| 97 | }
|
---|
| 98 |
|
---|
[4044] | 99 | public static double Calculate(ISymbolicExpressionTreeInterpreter interpreter, SymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, Dataset dataset, string targetVariable, IEnumerable<int> rows, out double beta, out double alpha, out double meanSE, out double varianceSE, out int count) {
|
---|
| 100 | IEnumerable<double> originalValues = dataset.GetEnumeratedVariableValues(targetVariable, rows);
|
---|
[4034] | 101 | IEnumerable<double> estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, dataset, rows);
|
---|
[3995] | 102 | CalculateScalingParameters(originalValues, estimatedValues, out beta, out alpha);
|
---|
| 103 |
|
---|
[4044] | 104 | return CalculateWithScaling(interpreter, solution, lowerEstimationLimit, upperEstimationLimit, dataset, targetVariable, rows, beta, alpha, out meanSE, out varianceSE, out count);
|
---|
[3532] | 105 | }
|
---|
| 106 |
|
---|
[4044] | 107 | public static double CalculateWithScaling(ISymbolicExpressionTreeInterpreter interpreter, SymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, Dataset dataset, string targetVariable, IEnumerable<int> rows, double beta, double alpha, out double meanSE, out double varianceSE, out int count) {
|
---|
[4034] | 108 | IEnumerable<double> estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, dataset, rows);
|
---|
| 109 | IEnumerable<double> originalValues = dataset.GetEnumeratedVariableValues(targetVariable, rows);
|
---|
[3995] | 110 | IEnumerator<double> originalEnumerator = originalValues.GetEnumerator();
|
---|
| 111 | IEnumerator<double> estimatedEnumerator = estimatedValues.GetEnumerator();
|
---|
[4044] | 112 | OnlineMeanAndVarianceCalculator seEvaluator = new OnlineMeanAndVarianceCalculator();
|
---|
[3532] | 113 |
|
---|
[3995] | 114 | while (originalEnumerator.MoveNext() & estimatedEnumerator.MoveNext()) {
|
---|
| 115 | double estimated = estimatedEnumerator.Current * beta + alpha;
|
---|
| 116 | double original = originalEnumerator.Current;
|
---|
| 117 | if (double.IsNaN(estimated))
|
---|
| 118 | estimated = upperEstimationLimit;
|
---|
[3996] | 119 | else
|
---|
| 120 | estimated = Math.Min(upperEstimationLimit, Math.Max(lowerEstimationLimit, estimated));
|
---|
[4044] | 121 | double error = estimated - original;
|
---|
| 122 | error *= error;
|
---|
| 123 | seEvaluator.Add(error);
|
---|
[3995] | 124 | }
|
---|
| 125 |
|
---|
| 126 | if (estimatedEnumerator.MoveNext() || originalEnumerator.MoveNext()) {
|
---|
| 127 | throw new ArgumentException("Number of elements in original and estimated enumeration doesn't match.");
|
---|
| 128 | } else {
|
---|
[4044] | 129 | meanSE = seEvaluator.Mean;
|
---|
| 130 | varianceSE = seEvaluator.Variance;
|
---|
| 131 | count = seEvaluator.Count;
|
---|
| 132 | return seEvaluator.Mean;
|
---|
[3995] | 133 | }
|
---|
[3532] | 134 | }
|
---|
| 135 |
|
---|
[3995] | 136 | /// <summary>
|
---|
| 137 | /// Calculates linear scaling parameters in one pass.
|
---|
| 138 | /// The formulas to calculate the scaling parameters were taken from Scaled Symblic Regression by Maarten Keijzer.
|
---|
| 139 | /// http://www.springerlink.com/content/x035121165125175/
|
---|
| 140 | /// </summary>
|
---|
[3532] | 141 | public static void CalculateScalingParameters(IEnumerable<double> original, IEnumerable<double> estimated, out double beta, out double alpha) {
|
---|
[3995] | 142 | IEnumerator<double> originalEnumerator = original.GetEnumerator();
|
---|
| 143 | IEnumerator<double> estimatedEnumerator = estimated.GetEnumerator();
|
---|
[4027] | 144 | OnlineMeanAndVarianceCalculator yVarianceCalculator = new OnlineMeanAndVarianceCalculator();
|
---|
| 145 | OnlineMeanAndVarianceCalculator tMeanCalculator = new OnlineMeanAndVarianceCalculator();
|
---|
| 146 | OnlineCovarianceEvaluator ytCovarianceEvaluator = new OnlineCovarianceEvaluator();
|
---|
[3995] | 147 | int cnt = 0;
|
---|
| 148 |
|
---|
| 149 | while (originalEnumerator.MoveNext() & estimatedEnumerator.MoveNext()) {
|
---|
| 150 | double y = estimatedEnumerator.Current;
|
---|
| 151 | double t = originalEnumerator.Current;
|
---|
| 152 | if (IsValidValue(t) && IsValidValue(y)) {
|
---|
[4027] | 153 | tMeanCalculator.Add(t);
|
---|
| 154 | yVarianceCalculator.Add(y);
|
---|
| 155 | ytCovarianceEvaluator.Add(y, t);
|
---|
| 156 |
|
---|
[3995] | 157 | cnt++;
|
---|
[3532] | 158 | }
|
---|
[3995] | 159 | }
|
---|
| 160 |
|
---|
| 161 | if (estimatedEnumerator.MoveNext() || originalEnumerator.MoveNext())
|
---|
| 162 | throw new ArgumentException("Number of elements in original and estimated enumeration doesn't match.");
|
---|
| 163 | if (cnt < 2) {
|
---|
| 164 | alpha = 0;
|
---|
| 165 | beta = 1;
|
---|
| 166 | } else {
|
---|
[4027] | 167 | if (yVarianceCalculator.Variance.IsAlmost(0.0))
|
---|
[3807] | 168 | beta = 1;
|
---|
[3995] | 169 | else
|
---|
[4027] | 170 | beta = ytCovarianceEvaluator.Covariance / yVarianceCalculator.Variance;
|
---|
[3995] | 171 |
|
---|
[4027] | 172 | alpha = tMeanCalculator.Mean - beta * yVarianceCalculator.Mean;
|
---|
[3532] | 173 | }
|
---|
| 174 | }
|
---|
| 175 |
|
---|
| 176 | private static bool IsValidValue(double d) {
|
---|
[3807] | 177 | return !double.IsInfinity(d) && !double.IsNaN(d) && d > -1.0E07 && d < 1.0E07; // don't consider very large or very small values for scaling
|
---|
[3532] | 178 | }
|
---|
| 179 | }
|
---|
| 180 | }
|
---|