[6589] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[16565] | 3 | * Copyright (C) 2002-2019 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[6589] | 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 |
|
---|
[16244] | 22 | using System;
|
---|
[6589] | 23 | using System.Collections.Generic;
|
---|
| 24 | using System.Linq;
|
---|
| 25 | using HeuristicLab.Common;
|
---|
| 26 | using HeuristicLab.Data;
|
---|
| 27 | using HeuristicLab.Optimization;
|
---|
[16565] | 28 | using HEAL.Attic;
|
---|
[13100] | 29 | using HeuristicLab.Problems.DataAnalysis.OnlineCalculators;
|
---|
[6589] | 30 |
|
---|
| 31 | namespace HeuristicLab.Problems.DataAnalysis {
|
---|
[16565] | 32 | [StorableType("60599497-EAF0-4DB0-B2E4-D58F34458D8F")]
|
---|
[6589] | 33 | public abstract class ClassificationSolutionBase : DataAnalysisSolution, IClassificationSolution {
|
---|
| 34 | private const string TrainingAccuracyResultName = "Accuracy (training)";
|
---|
| 35 | private const string TestAccuracyResultName = "Accuracy (test)";
|
---|
[6913] | 36 | private const string TrainingNormalizedGiniCoefficientResultName = "Normalized Gini Coefficient (training)";
|
---|
| 37 | private const string TestNormalizedGiniCoefficientResultName = "Normalized Gini Coefficient (test)";
|
---|
[11763] | 38 | private const string ClassificationPerformanceMeasuresResultName = "Classification Performance Measures";
|
---|
[6589] | 39 |
|
---|
| 40 | public new IClassificationModel Model {
|
---|
| 41 | get { return (IClassificationModel)base.Model; }
|
---|
| 42 | protected set { base.Model = value; }
|
---|
| 43 | }
|
---|
| 44 |
|
---|
| 45 | public new IClassificationProblemData ProblemData {
|
---|
| 46 | get { return (IClassificationProblemData)base.ProblemData; }
|
---|
[16244] | 47 | set {
|
---|
| 48 | if (value == null) throw new ArgumentNullException("The problemData must not be null.");
|
---|
| 49 | string errorMessage = string.Empty;
|
---|
| 50 | if (!Model.IsProblemDataCompatible(value, out errorMessage)) throw new ArgumentException(errorMessage);
|
---|
| 51 |
|
---|
| 52 | base.ProblemData = value;
|
---|
| 53 | }
|
---|
[6589] | 54 | }
|
---|
| 55 |
|
---|
| 56 | #region Results
|
---|
| 57 | public double TrainingAccuracy {
|
---|
| 58 | get { return ((DoubleValue)this[TrainingAccuracyResultName].Value).Value; }
|
---|
| 59 | private set { ((DoubleValue)this[TrainingAccuracyResultName].Value).Value = value; }
|
---|
| 60 | }
|
---|
| 61 | public double TestAccuracy {
|
---|
| 62 | get { return ((DoubleValue)this[TestAccuracyResultName].Value).Value; }
|
---|
| 63 | private set { ((DoubleValue)this[TestAccuracyResultName].Value).Value = value; }
|
---|
| 64 | }
|
---|
[6913] | 65 | public double TrainingNormalizedGiniCoefficient {
|
---|
| 66 | get { return ((DoubleValue)this[TrainingNormalizedGiniCoefficientResultName].Value).Value; }
|
---|
| 67 | protected set { ((DoubleValue)this[TrainingNormalizedGiniCoefficientResultName].Value).Value = value; }
|
---|
| 68 | }
|
---|
| 69 | public double TestNormalizedGiniCoefficient {
|
---|
| 70 | get { return ((DoubleValue)this[TestNormalizedGiniCoefficientResultName].Value).Value; }
|
---|
| 71 | protected set { ((DoubleValue)this[TestNormalizedGiniCoefficientResultName].Value).Value = value; }
|
---|
| 72 | }
|
---|
[11763] | 73 | public ClassificationPerformanceMeasuresResultCollection ClassificationPerformanceMeasures {
|
---|
| 74 | get { return ((ClassificationPerformanceMeasuresResultCollection)this[ClassificationPerformanceMeasuresResultName].Value); }
|
---|
| 75 | protected set { (this[ClassificationPerformanceMeasuresResultName].Value) = value; }
|
---|
| 76 | }
|
---|
[6589] | 77 | #endregion
|
---|
| 78 |
|
---|
| 79 | [StorableConstructor]
|
---|
[16565] | 80 | protected ClassificationSolutionBase(StorableConstructorFlag _) : base(_) { }
|
---|
[6589] | 81 | protected ClassificationSolutionBase(ClassificationSolutionBase original, Cloner cloner)
|
---|
| 82 | : base(original, cloner) {
|
---|
| 83 | }
|
---|
| 84 | protected ClassificationSolutionBase(IClassificationModel model, IClassificationProblemData problemData)
|
---|
| 85 | : base(model, problemData) {
|
---|
| 86 | Add(new Result(TrainingAccuracyResultName, "Accuracy of the model on the training partition (percentage of correctly classified instances).", new PercentValue()));
|
---|
| 87 | Add(new Result(TestAccuracyResultName, "Accuracy of the model on the test partition (percentage of correctly classified instances).", new PercentValue()));
|
---|
[6913] | 88 | Add(new Result(TrainingNormalizedGiniCoefficientResultName, "Normalized Gini coefficient of the model on the training partition.", new DoubleValue()));
|
---|
| 89 | Add(new Result(TestNormalizedGiniCoefficientResultName, "Normalized Gini coefficient of the model on the test partition.", new DoubleValue()));
|
---|
[11763] | 90 | Add(new Result(ClassificationPerformanceMeasuresResultName, @"Classification performance measures.\n
|
---|
| 91 | In a multiclass classification all misclassifications of the negative class will be treated as true negatives except on positive class estimations.",
|
---|
| 92 | new ClassificationPerformanceMeasuresResultCollection()));
|
---|
[6589] | 93 | }
|
---|
| 94 |
|
---|
[7011] | 95 | [StorableHook(HookType.AfterDeserialization)]
|
---|
| 96 | private void AfterDeserialization() {
|
---|
[14290] | 97 | if (string.IsNullOrEmpty(Model.TargetVariable))
|
---|
| 98 | Model.TargetVariable = this.ProblemData.TargetVariable;
|
---|
| 99 |
|
---|
[7011] | 100 | if (!this.ContainsKey(TrainingNormalizedGiniCoefficientResultName))
|
---|
| 101 | Add(new Result(TrainingNormalizedGiniCoefficientResultName, "Normalized Gini coefficient of the model on the training partition.", new DoubleValue()));
|
---|
| 102 | if (!this.ContainsKey(TestNormalizedGiniCoefficientResultName))
|
---|
| 103 | Add(new Result(TestNormalizedGiniCoefficientResultName, "Normalized Gini coefficient of the model on the test partition.", new DoubleValue()));
|
---|
[11763] | 104 | if (!this.ContainsKey(ClassificationPerformanceMeasuresResultName)) {
|
---|
| 105 | Add(new Result(ClassificationPerformanceMeasuresResultName, @"Classification performance measures.\n
|
---|
| 106 | In a multiclass classification all misclassifications of the negative class will be treated as true negatives except on positive class estimations.",
|
---|
| 107 | new ClassificationPerformanceMeasuresResultCollection()));
|
---|
[11766] | 108 | CalculateClassificationResults();
|
---|
[11763] | 109 | }
|
---|
[7011] | 110 | }
|
---|
| 111 |
|
---|
[8723] | 112 | protected void CalculateClassificationResults() {
|
---|
[6589] | 113 | double[] estimatedTrainingClassValues = EstimatedTrainingClassValues.ToArray(); // cache values
|
---|
[8139] | 114 | double[] originalTrainingClassValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TrainingIndices).ToArray();
|
---|
[11763] | 115 |
|
---|
[6589] | 116 | double[] estimatedTestClassValues = EstimatedTestClassValues.ToArray(); // cache values
|
---|
[8139] | 117 | double[] originalTestClassValues = ProblemData.Dataset.GetDoubleValues(ProblemData.TargetVariable, ProblemData.TestIndices).ToArray();
|
---|
[6589] | 118 |
|
---|
[11766] | 119 | var positiveClassName = ProblemData.PositiveClass;
|
---|
[11763] | 120 | double positiveClassValue = ProblemData.GetClassValue(positiveClassName);
|
---|
| 121 | ClassificationPerformanceMeasuresCalculator trainingPerformanceCalculator = new ClassificationPerformanceMeasuresCalculator(positiveClassName, positiveClassValue);
|
---|
| 122 | ClassificationPerformanceMeasuresCalculator testPerformanceCalculator = new ClassificationPerformanceMeasuresCalculator(positiveClassName, positiveClassValue);
|
---|
| 123 |
|
---|
[6589] | 124 | OnlineCalculatorError errorState;
|
---|
[6961] | 125 | double trainingAccuracy = OnlineAccuracyCalculator.Calculate(originalTrainingClassValues, estimatedTrainingClassValues, out errorState);
|
---|
[6589] | 126 | if (errorState != OnlineCalculatorError.None) trainingAccuracy = double.NaN;
|
---|
[6961] | 127 | double testAccuracy = OnlineAccuracyCalculator.Calculate(originalTestClassValues, estimatedTestClassValues, out errorState);
|
---|
[6589] | 128 | if (errorState != OnlineCalculatorError.None) testAccuracy = double.NaN;
|
---|
| 129 |
|
---|
| 130 | TrainingAccuracy = trainingAccuracy;
|
---|
| 131 | TestAccuracy = testAccuracy;
|
---|
[6913] | 132 |
|
---|
| 133 | double trainingNormalizedGini = NormalizedGiniCalculator.Calculate(originalTrainingClassValues, estimatedTrainingClassValues, out errorState);
|
---|
| 134 | if (errorState != OnlineCalculatorError.None) trainingNormalizedGini = double.NaN;
|
---|
| 135 | double testNormalizedGini = NormalizedGiniCalculator.Calculate(originalTestClassValues, estimatedTestClassValues, out errorState);
|
---|
| 136 | if (errorState != OnlineCalculatorError.None) testNormalizedGini = double.NaN;
|
---|
| 137 |
|
---|
| 138 | TrainingNormalizedGiniCoefficient = trainingNormalizedGini;
|
---|
| 139 | TestNormalizedGiniCoefficient = testNormalizedGini;
|
---|
[11763] | 140 |
|
---|
[13801] | 141 | ClassificationPerformanceMeasures.Reset();
|
---|
| 142 |
|
---|
[11763] | 143 | trainingPerformanceCalculator.Calculate(originalTrainingClassValues, estimatedTrainingClassValues);
|
---|
| 144 | if (trainingPerformanceCalculator.ErrorState == OnlineCalculatorError.None)
|
---|
| 145 | ClassificationPerformanceMeasures.SetTrainingResults(trainingPerformanceCalculator);
|
---|
| 146 |
|
---|
| 147 | testPerformanceCalculator.Calculate(originalTestClassValues, estimatedTestClassValues);
|
---|
| 148 | if (testPerformanceCalculator.ErrorState == OnlineCalculatorError.None)
|
---|
| 149 | ClassificationPerformanceMeasures.SetTestResults(testPerformanceCalculator);
|
---|
[13100] | 150 |
|
---|
[13102] | 151 | if (ProblemData.Classes == 2) {
|
---|
| 152 | var f1Training = FOneScoreCalculator.Calculate(originalTrainingClassValues, estimatedTrainingClassValues, out errorState);
|
---|
| 153 | if (errorState == OnlineCalculatorError.None) ClassificationPerformanceMeasures.TrainingF1Score = f1Training;
|
---|
| 154 | var f1Test = FOneScoreCalculator.Calculate(originalTestClassValues, estimatedTestClassValues, out errorState);
|
---|
| 155 | if (errorState == OnlineCalculatorError.None) ClassificationPerformanceMeasures.TestF1Score = f1Test;
|
---|
| 156 | }
|
---|
[13100] | 157 |
|
---|
| 158 | var mccTraining = MatthewsCorrelationCoefficientCalculator.Calculate(originalTrainingClassValues, estimatedTrainingClassValues, out errorState);
|
---|
| 159 | if (errorState == OnlineCalculatorError.None) ClassificationPerformanceMeasures.TrainingMatthewsCorrelation = mccTraining;
|
---|
| 160 | var mccTest = MatthewsCorrelationCoefficientCalculator.Calculate(originalTestClassValues, estimatedTestClassValues, out errorState);
|
---|
| 161 | if (errorState == OnlineCalculatorError.None) ClassificationPerformanceMeasures.TestMatthewsCorrelation = mccTest;
|
---|
[6589] | 162 | }
|
---|
| 163 |
|
---|
| 164 | public abstract IEnumerable<double> EstimatedClassValues { get; }
|
---|
| 165 | public abstract IEnumerable<double> EstimatedTrainingClassValues { get; }
|
---|
| 166 | public abstract IEnumerable<double> EstimatedTestClassValues { get; }
|
---|
| 167 |
|
---|
| 168 | public abstract IEnumerable<double> GetEstimatedClassValues(IEnumerable<int> rows);
|
---|
[8723] | 169 |
|
---|
| 170 | protected override void RecalculateResults() {
|
---|
| 171 | CalculateClassificationResults();
|
---|
| 172 | }
|
---|
[6589] | 173 | }
|
---|
| 174 | }
|
---|