Free cookie consent management tool by TermsFeed Policy Generator

source: branches/2845_EnhancedProgress/HeuristicLab.Algorithms.DataAnalysis/3.4/Linear/LinearDiscriminantAnalysis.cs @ 16307

Last change on this file since 16307 was 16307, checked in by pfleck, 5 years ago

#2845 Merged trunk changes before source move into branch

File size: 5.7 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2018 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.Threading;
26using HeuristicLab.Common;
27using HeuristicLab.Core;
28using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
29using HeuristicLab.Optimization;
30using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
31using HeuristicLab.Problems.DataAnalysis;
32using HeuristicLab.Problems.DataAnalysis.Symbolic;
33using HeuristicLab.Problems.DataAnalysis.Symbolic.Classification;
34
35namespace HeuristicLab.Algorithms.DataAnalysis {
36  /// <summary>
37  /// Linear discriminant analysis classification algorithm.
38  /// </summary>
39  [Item("Linear Discriminant Analysis (LDA)", "Linear discriminant analysis classification algorithm (wrapper for ALGLIB).")]
40  [Creatable(CreatableAttribute.Categories.DataAnalysisClassification, Priority = 100)]
41  [StorableClass]
42  public sealed class LinearDiscriminantAnalysis : FixedDataAnalysisAlgorithm<IClassificationProblem> {
43    private const string LinearDiscriminantAnalysisSolutionResultName = "Linear discriminant analysis solution";
44
45    [StorableConstructor]
46    private LinearDiscriminantAnalysis(bool deserializing) : base(deserializing) { }
47    private LinearDiscriminantAnalysis(LinearDiscriminantAnalysis original, Cloner cloner)
48      : base(original, cloner) {
49    }
50    public LinearDiscriminantAnalysis()
51      : base() {
52      Problem = new ClassificationProblem();
53    }
54    [StorableHook(HookType.AfterDeserialization)]
55    private void AfterDeserialization() { }
56
57    public override IDeepCloneable Clone(Cloner cloner) {
58      return new LinearDiscriminantAnalysis(this, cloner);
59    }
60
61    #region Fisher LDA
62    protected override void Run(CancellationToken cancellationToken) {
63      var solution = CreateLinearDiscriminantAnalysisSolution(Problem.ProblemData);
64      Results.Add(new Result(LinearDiscriminantAnalysisSolutionResultName, "The linear discriminant analysis.", solution));
65    }
66
67    public static IClassificationSolution CreateLinearDiscriminantAnalysisSolution(IClassificationProblemData problemData) {
68      var dataset = problemData.Dataset;
69      string targetVariable = problemData.TargetVariable;
70      IEnumerable<string> allowedInputVariables = problemData.AllowedInputVariables;
71      IEnumerable<int> rows = problemData.TrainingIndices;
72      int nClasses = problemData.ClassNames.Count();
73      var doubleVariableNames = allowedInputVariables.Where(dataset.VariableHasType<double>).ToArray();
74      var factorVariableNames = allowedInputVariables.Where(dataset.VariableHasType<string>).ToArray();
75      double[,] inputMatrix = dataset.ToArray(doubleVariableNames.Concat(new string[] { targetVariable }), rows);
76
77      var factorVariables = dataset.GetFactorVariableValues(factorVariableNames, rows);
78      var factorMatrix = dataset.ToArray(factorVariables, rows);
79
80      inputMatrix = factorMatrix.HorzCat(inputMatrix);
81
82      if (inputMatrix.Cast<double>().Any(x => double.IsNaN(x) || double.IsInfinity(x)))
83        throw new NotSupportedException("Linear discriminant analysis does not support NaN or infinity values in the input dataset.");
84
85      // change class values into class index
86      int targetVariableColumn = inputMatrix.GetLength(1) - 1;
87      List<double> classValues = problemData.ClassValues.OrderBy(x => x).ToList();
88      for (int row = 0; row < inputMatrix.GetLength(0); row++) {
89        inputMatrix[row, targetVariableColumn] = classValues.IndexOf(inputMatrix[row, targetVariableColumn]);
90      }
91      int info;
92      double[] w;
93      alglib.fisherlda(inputMatrix, inputMatrix.GetLength(0), inputMatrix.GetLength(1) - 1, nClasses, out info, out w);
94      if (info < 1) throw new ArgumentException("Error in calculation of linear discriminant analysis solution");
95
96      var nFactorCoeff = factorMatrix.GetLength(1);
97      var tree = LinearModelToTreeConverter.CreateTree(factorVariables, w.Take(nFactorCoeff).ToArray(),
98        doubleVariableNames, w.Skip(nFactorCoeff).Take(doubleVariableNames.Length).ToArray());
99
100      var model = CreateDiscriminantFunctionModel(tree, new SymbolicDataAnalysisExpressionTreeLinearInterpreter(), problemData, rows);
101      SymbolicDiscriminantFunctionClassificationSolution solution = new SymbolicDiscriminantFunctionClassificationSolution(model, (IClassificationProblemData)problemData.Clone());
102
103      return solution;
104    }
105    #endregion
106
107    private static SymbolicDiscriminantFunctionClassificationModel CreateDiscriminantFunctionModel(ISymbolicExpressionTree tree,
108      ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
109      IClassificationProblemData problemData,
110      IEnumerable<int> rows) {
111      var model = new SymbolicDiscriminantFunctionClassificationModel(problemData.TargetVariable, tree, interpreter, new AccuracyMaximizationThresholdCalculator());
112      model.RecalculateModelParameters(problemData, rows);
113      return model;
114    }
115  }
116}
Note: See TracBrowser for help on using the repository browser.