source: stable/HeuristicLab.Algorithms.DataAnalysis/3.4/NearestNeighbour/NearestNeighbourClassification.cs @ 17164

Last change on this file since 17164 was 17164, checked in by gkronber, 2 months ago

#2942: merged r16491 from trunk to stable

File size: 6.0 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2019 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 System.Threading;
25using HeuristicLab.Common;
26using HeuristicLab.Core;
27using HeuristicLab.Data;
28using HeuristicLab.Optimization;
29using HeuristicLab.Parameters;
30using HEAL.Attic;
31using HeuristicLab.Problems.DataAnalysis;
32
33namespace HeuristicLab.Algorithms.DataAnalysis {
34  /// <summary>
35  /// Nearest neighbour classification data analysis algorithm.
36  /// </summary>
37  [Item("Nearest Neighbour Classification (kNN)", "Nearest neighbour classification data analysis algorithm (wrapper for ALGLIB).")]
38  [Creatable(CreatableAttribute.Categories.DataAnalysisClassification, Priority = 150)]
39  [StorableType("98161D6F-D977-45EA-B899-E47EE017865E")]
40  public sealed class NearestNeighbourClassification : FixedDataAnalysisAlgorithm<IClassificationProblem> {
41    private const string KParameterName = "K";
42    private const string NearestNeighbourClassificationModelResultName = "Nearest neighbour classification solution";
43    private const string WeightsParameterName = "Weights";
44    private const string SelfMatchParameterName = "SelfMatch";
45
46    #region parameter properties
47    public IFixedValueParameter<IntValue> KParameter {
48      get { return (IFixedValueParameter<IntValue>)Parameters[KParameterName]; }
49    }
50    public IFixedValueParameter<BoolValue> SelfMatchParameter {
51      get { return (IFixedValueParameter<BoolValue>)Parameters[SelfMatchParameterName]; }
52    }
53    public IValueParameter<DoubleArray> WeightsParameter {
54      get { return (IValueParameter<DoubleArray>)Parameters[WeightsParameterName]; }
55    }
56    #endregion
57    #region properties
58    public bool SelfMatch {
59      get { return SelfMatchParameter.Value.Value; }
60      set { SelfMatchParameter.Value.Value = value; }
61    }
62    public int K {
63      get { return KParameter.Value.Value; }
64      set {
65        if (value <= 0) throw new ArgumentException("K must be larger than zero.", "K");
66        else KParameter.Value.Value = value;
67      }
68    }
69    public DoubleArray Weights {
70      get { return WeightsParameter.Value; }
71      set { WeightsParameter.Value = value; }
72    }
73    #endregion
74
75    [StorableConstructor]
76    private NearestNeighbourClassification(StorableConstructorFlag _) : base(_) { }
77    private NearestNeighbourClassification(NearestNeighbourClassification original, Cloner cloner)
78      : base(original, cloner) {
79    }
80    public NearestNeighbourClassification()
81      : base() {
82      Parameters.Add(new FixedValueParameter<BoolValue>(SelfMatchParameterName, "Should we use equal points for classification?", new BoolValue(false)));
83      Parameters.Add(new FixedValueParameter<IntValue>(KParameterName, "The number of nearest neighbours to consider for regression.", new IntValue(3)));
84      Parameters.Add(new OptionalValueParameter<DoubleArray>(WeightsParameterName, "Optional: use weights to specify individual scaling values for all features. If not set the weights are calculated automatically (each feature is scaled to unit variance)"));
85      Problem = new ClassificationProblem();
86    }
87    [StorableHook(HookType.AfterDeserialization)]
88    private void AfterDeserialization() {
89      // BackwardsCompatibility3.3
90      #region Backwards compatible code, remove with 3.4
91      if (!Parameters.ContainsKey(WeightsParameterName)) {
92        Parameters.Add(new OptionalValueParameter<DoubleArray>(WeightsParameterName, "Optional: use weights to specify individual scaling values for all features. If not set the weights are calculated automatically (each feature is scaled to unit variance)"));
93      }
94      if (!Parameters.ContainsKey(SelfMatchParameterName)) {
95        Parameters.Add(new FixedValueParameter<BoolValue>(SelfMatchParameterName, "Should we use equal points for classification?", new BoolValue(false)));
96      }
97      #endregion
98    }
99
100    public override IDeepCloneable Clone(Cloner cloner) {
101      return new NearestNeighbourClassification(this, cloner);
102    }
103
104    #region nearest neighbour
105    protected override void Run(CancellationToken cancellationToken) {
106      double[] weights = null;
107      if (Weights != null) weights = Weights.CloneAsArray();
108      var solution = CreateNearestNeighbourClassificationSolution(Problem.ProblemData, K, SelfMatch, weights);
109      Results.Add(new Result(NearestNeighbourClassificationModelResultName, "The nearest neighbour classification solution.", solution));
110    }
111
112    public static IClassificationSolution CreateNearestNeighbourClassificationSolution(IClassificationProblemData problemData, int k, bool selfMatch = false, double[] weights = null) {
113      var problemDataClone = (IClassificationProblemData)problemData.Clone();
114      return new NearestNeighbourClassificationSolution(Train(problemDataClone, k, selfMatch, weights), problemDataClone);
115    }
116
117    public static INearestNeighbourModel Train(IClassificationProblemData problemData, int k, bool selfMatch = false, double[] weights = null) {
118      return new NearestNeighbourModel(problemData.Dataset,
119        problemData.TrainingIndices,
120        k,
121        selfMatch,
122        problemData.TargetVariable,
123        problemData.AllowedInputVariables,
124        weights,
125        problemData.ClassValues.ToArray());
126    }
127    #endregion
128  }
129}
Note: See TracBrowser for help on using the repository browser.