Free cookie consent management tool by TermsFeed Policy Generator

source: branches/3.2/sources/HeuristicLab.GP.StructureIdentification.Classification/3.3/MulticlassModeller.cs @ 3580

Last change on this file since 3580 was 2440, checked in by gkronber, 15 years ago

Fixed #784 (ProblemInjector should be changed to read variable names instead of indexes for input and target variables)

File size: 8.1 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 HeuristicLab.Common;
25using HeuristicLab.Core;
26using HeuristicLab.Data;
27using HeuristicLab.DataAnalysis;
28
29namespace HeuristicLab.GP.StructureIdentification.Classification {
30  public class MulticlassModeller : OperatorBase {
31
32    private const string DATASET = "Dataset";
33    private const string TARGETVARIABLE = "TargetVariable";
34    private const string TARGETCLASSVALUES = "TargetClassValues";
35    private const string TRAININGSAMPLESSTART = "TrainingSamplesStart";
36    private const string TRAININGSAMPLESEND = "TrainingSamplesEnd";
37    private const string VALIDATIONSAMPLESSTART = "ValidationSamplesStart";
38    private const string VALIDATIONSAMPLESEND = "ValidationSamplesEnd";
39    private const string CLASSAVALUE = "ClassAValue";
40    private const string CLASSBVALUE = "ClassBValue";
41    private const double EPSILON = 1E-6;
42    public override string Description {
43      get { return @"TASK"; }
44    }
45
46    public MulticlassModeller()
47      : base() {
48      AddVariableInfo(new VariableInfo(DATASET, "The original dataset and the new dataset parts in the newly created subscopes", typeof(Dataset), VariableKind.In));
49      AddVariableInfo(new VariableInfo(TARGETVARIABLE, "TargetVariable", typeof(StringData), VariableKind.In));
50      AddVariableInfo(new VariableInfo(TARGETCLASSVALUES, "Class values of the target variable in the original dataset and in the new dataset parts", typeof(ItemList<DoubleData>), VariableKind.In | VariableKind.New));
51      AddVariableInfo(new VariableInfo(CLASSAVALUE, "The original class value of the new class A", typeof(DoubleData), VariableKind.New));
52      AddVariableInfo(new VariableInfo(CLASSBVALUE, "The original class value of the new class B", typeof(DoubleData), VariableKind.New));
53      AddVariableInfo(new VariableInfo(TRAININGSAMPLESSTART, "The start of training samples in the original dataset and starts of training samples in the new dataset parts", typeof(IntData), VariableKind.In | VariableKind.New));
54      AddVariableInfo(new VariableInfo(TRAININGSAMPLESEND, "The end of training samples in the original dataset and ends of training samples in the new dataset parts", typeof(IntData), VariableKind.In | VariableKind.New));
55      AddVariableInfo(new VariableInfo(VALIDATIONSAMPLESSTART, "The start of validation samples in the original dataset and starts of validation samples in the new dataset parts", typeof(IntData), VariableKind.In | VariableKind.New));
56      AddVariableInfo(new VariableInfo(VALIDATIONSAMPLESEND, "The end of validation samples in the original dataset and ends of validation samples in the new dataset parts", typeof(IntData), VariableKind.In | VariableKind.New));
57    }
58
59    public override IOperation Apply(IScope scope) {
60      Dataset origDataset = GetVariableValue<Dataset>(DATASET, scope, true);
61      int targetVariable = origDataset.GetVariableIndex(GetVariableValue<StringData>(TARGETVARIABLE, scope, true).Data);
62      ItemList<DoubleData> classValues = GetVariableValue<ItemList<DoubleData>>(TARGETCLASSVALUES, scope, true);
63      int origTrainingSamplesStart = GetVariableValue<IntData>(TRAININGSAMPLESSTART, scope, true).Data;
64      int origTrainingSamplesEnd = GetVariableValue<IntData>(TRAININGSAMPLESEND, scope, true).Data;
65      int origValidationSamplesStart = GetVariableValue<IntData>(VALIDATIONSAMPLESSTART, scope, true).Data;
66      int origValidationSamplesEnd = GetVariableValue<IntData>(VALIDATIONSAMPLESEND, scope, true).Data;
67      ItemList<DoubleData> binaryClassValues = new ItemList<DoubleData>();
68      binaryClassValues.Add(new DoubleData(0.0));
69      binaryClassValues.Add(new DoubleData(1.0));
70      for (int i = 0; i < classValues.Count - 1; i++) {
71        for (int j = i + 1; j < classValues.Count; j++) {
72          Dataset dataset = new Dataset();
73          dataset.Columns = origDataset.Columns;
74          double classAValue = classValues[i].Data;
75          double classBValue = classValues[j].Data;
76          int trainingSamplesStart;
77          int trainingSamplesEnd;
78          int validationSamplesStart;
79          int validationSamplesEnd;
80
81          trainingSamplesStart = 0;
82          List<double[]> rows = new List<double[]>();
83          for (int k = origTrainingSamplesStart; k < origTrainingSamplesEnd; k++) {
84            double[] row = new double[dataset.Columns];
85            double targetValue = origDataset.GetValue(k, targetVariable);
86            if (targetValue.IsAlmost(classAValue)) {
87              for (int l = 0; l < row.Length; l++) {
88                row[l] = origDataset.GetValue(k, l);
89              }
90              row[targetVariable] = 0;
91              rows.Add(row);
92            } else if (targetValue.IsAlmost(classBValue)) {
93              for (int l = 0; l < row.Length; l++) {
94                row[l] = origDataset.GetValue(k, l);
95              }
96              row[targetVariable] = 1.0;
97              rows.Add(row);
98            }
99          }
100          trainingSamplesEnd = rows.Count;
101          validationSamplesStart = rows.Count;
102          for (int k = origValidationSamplesStart; k < origValidationSamplesEnd; k++) {
103            double[] row = new double[dataset.Columns];
104            double targetValue = origDataset.GetValue(k, targetVariable);
105            if (targetValue.IsAlmost(classAValue)) {
106              for (int l = 0; l < row.Length; l++) {
107                row[l] = origDataset.GetValue(k, l);
108              }
109              row[targetVariable] = 0;
110              rows.Add(row);
111            } else if (targetValue.IsAlmost(classBValue)) {
112              for (int l = 0; l < row.Length; l++) {
113                row[l] = origDataset.GetValue(k, l);
114              }
115              row[targetVariable] = 1.0;
116              rows.Add(row);
117            }
118          }
119          validationSamplesEnd = rows.Count;
120
121          dataset.Rows = rows.Count;
122          dataset.Samples = new double[dataset.Rows * dataset.Columns];
123          for (int k = 0; k < dataset.Rows; k++) {
124            for (int l = 0; l < dataset.Columns; l++) {
125              dataset.SetValue(k, l, rows[k][l]);
126            }
127          }
128
129          Scope childScope = new Scope(classAValue + " vs. " + classBValue);
130
131          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(TARGETCLASSVALUES), binaryClassValues));
132          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(CLASSAVALUE), new DoubleData(classAValue)));
133          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(CLASSBVALUE), new DoubleData(classBValue)));
134          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(TRAININGSAMPLESSTART), new IntData(trainingSamplesStart)));
135          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(TRAININGSAMPLESEND), new IntData(trainingSamplesEnd)));
136          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(VALIDATIONSAMPLESSTART), new IntData(validationSamplesStart)));
137          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(VALIDATIONSAMPLESEND), new IntData(validationSamplesEnd)));
138          childScope.AddVariable(new HeuristicLab.Core.Variable(scope.TranslateName(DATASET), dataset));
139          scope.AddSubScope(childScope);
140        }
141      }
142      return null;
143    }
144  }
145}
Note: See TracBrowser for help on using the repository browser.