- Timestamp:
- 09/14/10 18:46:12 (14 years ago)
- Location:
- branches/HeuristicLab.Classification
- Files:
-
- 1 added
- 1 deleted
- 7 edited
- 2 moved
Legend:
- Unmodified
- Added
- Removed
-
branches/HeuristicLab.Classification/HeuristicLab.Classification.sln
r4366 r4391 7 7 EndProjectSection 8 8 EndProject 9 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "HeuristicLab.Problems.DataAnalysis.Classification ", "HeuristicLab.Problems.DataAnalysis.Classification\3.3\HeuristicLab.Problems.DataAnalysis.Classification.csproj", "{D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}"9 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "HeuristicLab.Problems.DataAnalysis.Classification-3.3", "HeuristicLab.Problems.DataAnalysis.Classification\3.3\HeuristicLab.Problems.DataAnalysis.Classification-3.3.csproj", "{D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}" 10 10 EndProject 11 11 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Solution Items", "Solution Items", "{08118029-0DA6-4105-BF1F-9453C0F3AC75}" 12 12 EndProject 13 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "HeuristicLab.Problems.DataAnalysis.Classification.Views ", "HeuristicLab.Problems.DataAnalysis.Classification.Views\3.3\HeuristicLab.Problems.DataAnalysis.Classification.Views.csproj", "{7F038E9A-B91E-4F49-8C27-D47BC86F1220}"13 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "HeuristicLab.Problems.DataAnalysis.Classification.Views-3.3", "HeuristicLab.Problems.DataAnalysis.Classification.Views\3.3\HeuristicLab.Problems.DataAnalysis.Classification.Views-3.3.csproj", "{7F038E9A-B91E-4F49-8C27-D47BC86F1220}" 14 14 EndProject 15 15 Global … … 36 36 {3AB5A7E5-9F0C-4920-B18D-E06B9577618C}.Release|Any CPU.Build.0 = Release|Any CPU 37 37 {3AB5A7E5-9F0C-4920-B18D-E06B9577618C}.Release|x64.ActiveCfg = Release|Any CPU 38 {3AB5A7E5-9F0C-4920-B18D-E06B9577618C}.Release|x86.ActiveCfg = Release|Any CPU 38 {3AB5A7E5-9F0C-4920-B18D-E06B9577618C}.Release|x86.ActiveCfg = Release|x86 39 {3AB5A7E5-9F0C-4920-B18D-E06B9577618C}.Release|x86.Build.0 = Release|x86 39 40 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Debug|Any CPU.ActiveCfg = Debug|Any CPU 40 41 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Debug|Any CPU.Build.0 = Debug|Any CPU … … 46 47 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Release|Any CPU.Build.0 = Release|Any CPU 47 48 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Release|x64.ActiveCfg = Release|Any CPU 48 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Release|x86.ActiveCfg = Release|Any CPU 49 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Release|x86.ActiveCfg = Release|x86 50 {D35FD6DB-D704-4BEA-BDAE-C40ABFB5B13E}.Release|x86.Build.0 = Release|x86 49 51 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Debug|Any CPU.ActiveCfg = Debug|Any CPU 50 52 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Debug|Any CPU.Build.0 = Debug|Any CPU … … 56 58 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Release|Any CPU.Build.0 = Release|Any CPU 57 59 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Release|x64.ActiveCfg = Release|Any CPU 58 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Release|x86.ActiveCfg = Release|Any CPU 60 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Release|x86.ActiveCfg = Release|x86 61 {7F038E9A-B91E-4F49-8C27-D47BC86F1220}.Release|x86.Build.0 = Release|x86 59 62 EndGlobalSection 60 63 GlobalSection(SolutionProperties) = preSolution -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/ClassificationProblemData.cs
r4366 r4391 34 34 [StorableClass] 35 35 public class ClassificationProblemData : DataAnalysisProblemData { 36 #region default data 37 private static string[] defaultInputs = new string[] { "sample", "clump thickness", "cell size", "cell shape", "marginal adhesion", "epithelial cell size", "bare nuclei", "chromatin", "nucleoli", "mitoses", "class" }; 38 private static double[,] defaultData = new double[,]{ 39 {1000025,5,1,1,1,2,1,3,1,1,2 }, 40 {1002945,5,4,4,5,7,10,3,2,1,2 }, 41 {1015425,3,1,1,1,2,2,3,1,1,2 }, 42 {1016277,6,8,8,1,3,4,3,7,1,2 }, 43 {1017023,4,1,1,3,2,1,3,1,1,2 }, 44 {1017122,8,10,10,8,7,10,9,7,1,4 }, 45 {1018099,1,1,1,1,2,10,3,1,1,2 }, 46 {1018561,2,1,2,1,2,1,3,1,1,2 }, 47 {1033078,2,1,1,1,2,1,1,1,5,2 }, 48 {1033078,4,2,1,1,2,1,2,1,1,2 }, 49 {1035283,1,1,1,1,1,1,3,1,1,2 }, 50 {1036172,2,1,1,1,2,1,2,1,1,2 }, 51 {1041801,5,3,3,3,2,3,4,4,1,4 }, 52 {1043999,1,1,1,1,2,3,3,1,1,2 }, 53 {1044572,8,7,5,10,7,9,5,5,4,4 }, 54 {1047630,7,4,6,4,6,1,4,3,1,4 }, 55 {1048672,4,1,1,1,2,1,2,1,1,2 }, 56 {1049815,4,1,1,1,2,1,3,1,1,2 }, 57 {1050670,10,7,7,6,4,10,4,1,2,4 }, 58 {1050718,6,1,1,1,2,1,3,1,1,2 }, 59 {1054590,7,3,2,10,5,10,5,4,4,4 }, 60 {1054593,10,5,5,3,6,7,7,10,1,4 }, 61 {1056784,3,1,1,1,2,1,2,1,1,2 }, 62 {1057013,8,4,5,1,2,2,7,3,1,4 }, 63 {1059552,1,1,1,1,2,1,3,1,1,2 }, 64 {1065726,5,2,3,4,2,7,3,6,1,4 }, 65 {1066373,3,2,1,1,1,1,2,1,1,2 }, 66 {1066979,5,1,1,1,2,1,2,1,1,2 }, 67 {1067444,2,1,1,1,2,1,2,1,1,2 }, 68 {1070935,1,1,3,1,2,1,1,1,1,2 }, 69 {1070935,3,1,1,1,1,1,2,1,1,2 }, 70 {1071760,2,1,1,1,2,1,3,1,1,2 }, 71 {1072179,10,7,7,3,8,5,7,4,3,4 }, 72 {1074610,2,1,1,2,2,1,3,1,1,2 }, 73 {1075123,3,1,2,1,2,1,2,1,1,2 }, 74 {1079304,2,1,1,1,2,1,2,1,1,2 }, 75 {1080185,10,10,10,8,6,1,8,9,1,4 }, 76 {1081791,6,2,1,1,1,1,7,1,1,2 }, 77 {1084584,5,4,4,9,2,10,5,6,1,4 }, 78 {1091262,2,5,3,3,6,7,7,5,1,4 }, 79 {1096800,6,6,6,9,6,4,7,8,1,2 }, 80 {1099510,10,4,3,1,3,3,6,5,2,4 }, 81 {1100524,6,10,10,2,8,10,7,3,3,4 }, 82 {1102573,5,6,5,6,10,1,3,1,1,4 }, 83 {1103608,10,10,10,4,8,1,8,10,1,4 }, 84 {1103722,1,1,1,1,2,1,2,1,2,2 }, 85 {1105257,3,7,7,4,4,9,4,8,1,4 }, 86 {1105524,1,1,1,1,2,1,2,1,1,2 }, 87 {1106095,4,1,1,3,2,1,3,1,1,2 }, 88 {1106829,7,8,7,2,4,8,3,8,2,4 }, 89 {1108370,9,5,8,1,2,3,2,1,5,4 }, 90 {1108449,5,3,3,4,2,4,3,4,1,4 }, 91 {1110102,10,3,6,2,3,5,4,10,2,4 }, 92 {1110503,5,5,5,8,10,8,7,3,7,4 }, 93 {1110524,10,5,5,6,8,8,7,1,1,4 }, 94 {1111249,10,6,6,3,4,5,3,6,1,4 }, 95 {1112209,8,10,10,1,3,6,3,9,1,4 }, 96 {1113038,8,2,4,1,5,1,5,4,4,4 }, 97 {1113483,5,2,3,1,6,10,5,1,1,4 }, 98 {1113906,9,5,5,2,2,2,5,1,1,4 }, 99 {1115282,5,3,5,5,3,3,4,10,1,4 }, 100 {1115293,1,1,1,1,2,2,2,1,1,2 }, 101 {1116116,9,10,10,1,10,8,3,3,1,4 }, 102 {1116132,6,3,4,1,5,2,3,9,1,4 }, 103 {1116192,1,1,1,1,2,1,2,1,1,2 }, 104 {1116998,10,4,2,1,3,2,4,3,10,4 }, 105 {1117152,4,1,1,1,2,1,3,1,1,2 }, 106 {1118039,5,3,4,1,8,10,4,9,1,4 }, 107 {1120559,8,3,8,3,4,9,8,9,8,4 }, 108 {1121732,1,1,1,1,2,1,3,2,1,2 }, 109 {1121919,5,1,3,1,2,1,2,1,1,2 }, 110 {1123061,6,10,2,8,10,2,7,8,10,4 }, 111 {1124651,1,3,3,2,2,1,7,2,1,2 }, 112 {1125035,9,4,5,10,6,10,4,8,1,4 }, 113 {1126417,10,6,4,1,3,4,3,2,3,4 }, 114 {1131294,1,1,2,1,2,2,4,2,1,2 }, 115 {1132347,1,1,4,1,2,1,2,1,1,2 }, 116 {1133041,5,3,1,2,2,1,2,1,1,2 }, 117 {1133136,3,1,1,1,2,3,3,1,1,2 }, 118 {1136142,2,1,1,1,3,1,2,1,1,2 }, 119 {1137156,2,2,2,1,1,1,7,1,1,2 }, 120 {1143978,4,1,1,2,2,1,2,1,1,2 }, 121 {1143978,5,2,1,1,2,1,3,1,1,2 }, 122 {1147044,3,1,1,1,2,2,7,1,1,2 }, 123 {1147699,3,5,7,8,8,9,7,10,7,4 }, 124 {1147748,5,10,6,1,10,4,4,10,10,4 }, 125 {1148278,3,3,6,4,5,8,4,4,1,4 }, 126 {1148873,3,6,6,6,5,10,6,8,3,4 }, 127 {1152331,4,1,1,1,2,1,3,1,1,2 }, 128 {1155546,2,1,1,2,3,1,2,1,1,2 }, 129 {1156272,1,1,1,1,2,1,3,1,1,2 }, 130 {1156948,3,1,1,2,2,1,1,1,1,2 }, 131 {1157734,4,1,1,1,2,1,3,1,1,2 }, 132 {1158247,1,1,1,1,2,1,2,1,1,2 }, 133 {1160476,2,1,1,1,2,1,3,1,1,2 }, 134 {1164066,1,1,1,1,2,1,3,1,1,2 }, 135 {1165297,2,1,1,2,2,1,1,1,1,2 }, 136 {1165790,5,1,1,1,2,1,3,1,1,2 }, 137 {1165926,9,6,9,2,10,6,2,9,10,4 }, 138 {1166630,7,5,6,10,5,10,7,9,4,4 }, 139 {1166654,10,3,5,1,10,5,3,10,2,4 }, 140 {1167439,2,3,4,4,2,5,2,5,1,4 }, 141 {1167471,4,1,2,1,2,1,3,1,1,2 }, 142 {1168359,8,2,3,1,6,3,7,1,1,4 }, 143 {1168736,10,10,10,10,10,1,8,8,8,4 }, 144 {1169049,7,3,4,4,3,3,3,2,7,4 }, 145 {1170419,10,10,10,8,2,10,4,1,1,4 }, 146 {1170420,1,6,8,10,8,10,5,7,1,4 }, 147 {1171710,1,1,1,1,2,1,2,3,1,2 }, 148 {1171710,6,5,4,4,3,9,7,8,3,4 }, 149 {1171795,1,3,1,2,2,2,5,3,2,2 }, 150 {1171845,8,6,4,3,5,9,3,1,1,4 }, 151 {1172152,10,3,3,10,2,10,7,3,3,4 }, 152 {1173216,10,10,10,3,10,8,8,1,1,4 }, 153 {1173235,3,3,2,1,2,3,3,1,1,2 }, 154 {1173347,1,1,1,1,2,5,1,1,1,2 }, 155 {1173347,8,3,3,1,2,2,3,2,1,2 }, 156 {1173509,4,5,5,10,4,10,7,5,8,4 }, 157 {1173514,1,1,1,1,4,3,1,1,1,2 }, 158 {1173681,3,2,1,1,2,2,3,1,1,2 }, 159 {1174057,1,1,2,2,2,1,3,1,1,2 }, 160 {1174057,4,2,1,1,2,2,3,1,1,2 }, 161 {1174131,10,10,10,2,10,10,5,3,3,4 }, 162 {1174428,5,3,5,1,8,10,5,3,1,4 }, 163 {1175937,5,4,6,7,9,7,8,10,1,4 }, 164 {1176406,1,1,1,1,2,1,2,1,1,2 }, 165 {1176881,7,5,3,7,4,10,7,5,5,4 } 166 }; 167 #endregion 168 36 169 private const int MaximumClasses = 20; 37 170 private const string ClassNamesParameterName = "ClassNames"; … … 64 197 65 198 public ClassificationProblemData() 66 : base( ) {199 : base(new Dataset(defaultInputs, defaultData), defaultInputs, defaultInputs[defaultInputs.Length - 1], 0, 60, 60, 120) { 67 200 Parameters.Add(new ValueParameter<StringArray>(ClassNamesParameterName, "An array of the names for all class values.")); 68 201 Parameters.Add(new ValueParameter<DoubleMatrix>(MisclassificationMatrixParameterName, "A matrix that describles the penalties for misclassifaction between the single classes.")); 69 202 sortedClassValues = new List<double>(); 70 203 204 InputVariables.SetItemCheckedState(InputVariables[InputVariables.Count - 1], false); 71 205 RegisterParameterEvents(); 72 206 UpdateClassValues(); … … 173 307 174 308 private void UpdateMisclassifciationMatrixHeaders() { 175 MisclassificationMatrix.RowNames = ClassNames ;176 MisclassificationMatrix.ColumnNames = ClassNames ;309 MisclassificationMatrix.RowNames = ClassNames.Select(name => "Estimated " + name); 310 MisclassificationMatrix.ColumnNames = ClassNames.Select(name => "Actual " + name); 177 311 } 178 312 -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/HeuristicLab.Problems.DataAnalysis.Classification-3.3.csproj
r4388 r4391 156 156 <Compile Include="Interfaces\ISymbolicClassificationAnalyzer.cs" /> 157 157 <Compile Include="Symbolic\Analyzer\ValidationBestSymbolicClassificationSolutionAnalyzer.cs" /> 158 <Compile Include="Symbolic\Evaluators\SymbolicClassificationPearsonRSquaredEvaluator.cs" /> 158 159 <Compile Include="Symbolic\SymbolicClassificationSolution.cs" /> 159 <Compile Include="Symbolic\SingleObjectiveSymbolicClassificationEvaluator.cs" /> 160 <Compile Include="Symbolic\SymbolicClassificationMeanSquaredErrorEvaluator.cs" /> 160 <Compile Include="Symbolic\Evaluators\SymbolicClassificationMeanSquaredErrorEvaluator.cs" /> 161 161 <Compile Include="Symbolic\SymbolicClassificationProblem.cs" /> 162 162 </ItemGroup> -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/Interfaces/ISymbolicClassificationAnalyzer.cs
r4366 r4391 20 20 #endregion 21 21 22 using HeuristicLab.Core; 23 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding; 24 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding.Interfaces; 25 using HeuristicLab.Optimization; 26 using HeuristicLab.Parameters; 27 28 // This interface is exactly the same as ISymbolicRegressionAnalyzer 29 // consider creating a base interface for both analyzer 30 22 using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic.Analyzers; 31 23 namespace HeuristicLab.Problems.DataAnalysis.Classification { 32 public interface ISymbolicClassificationAnalyzer : ISymbolicExpressionTreeAnalyzer { 33 ILookupParameter<ResultCollection> ResultsParameter { get; } 24 public interface ISymbolicClassificationAnalyzer : ISymbolicRegressionAnalyzer { 34 25 } 35 26 } -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/Interfaces/ISymbolicClassificationEvaluator.cs
r4366 r4391 20 20 #endregion 21 21 22 using System.Collections.Generic; 23 using HeuristicLab.Core; 24 using HeuristicLab.Data; 25 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding; 26 using HeuristicLab.Optimization; 27 using HeuristicLab.Problems.DataAnalysis.Symbolic; 22 using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic; 28 23 namespace HeuristicLab.Problems.DataAnalysis.Classification { 29 public interface ISymbolicClassificationEvaluator : ISingleObjectiveEvaluator { 30 ILookupParameter<ISymbolicExpressionTreeInterpreter> SymbolicExpressionTreeInterpreterParameter { get; } 31 ILookupParameter<SymbolicExpressionTree> SymbolicExpressionTreeParameter { get; } 32 ILookupParameter<ClassificationProblemData> RegressionProblemDataParameter { get; } 33 IValueLookupParameter<IntValue> SamplesStartParameter { get; } 34 IValueLookupParameter<IntValue> SamplesEndParameter { get; } 35 IValueLookupParameter<DoubleValue> UpperEstimationLimitParameter { get; } 36 IValueLookupParameter<DoubleValue> LowerEstimationLimitParameter { get; } 37 38 39 double Evaluate(ISymbolicExpressionTreeInterpreter interpreter, SymbolicExpressionTree tree, 40 double lowerEstimationLimit, double upperEstimationLimit, 41 Dataset dataset, string targetVariable, IEnumerable<double> sortedClassValues, IEnumerable<int> rows); 24 public interface ISymbolicClassificationEvaluator : ISymbolicRegressionEvaluator { 25 ClassificationProblemData ClassificationProblemData { get; } 42 26 } 43 27 } -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/Symbolic/Analyzer/ValidationBestSymbolicClassificationSolutionAnalyzer.cs
r4366 r4391 21 21 22 22 using System.Collections.Generic; 23 using System.Linq; 24 using HeuristicLab.Analysis; 23 25 using HeuristicLab.Core; 24 26 using HeuristicLab.Data; … … 29 31 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable; 30 32 using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic; 33 using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic.Analyzers; 31 34 using HeuristicLab.Problems.DataAnalysis.Symbolic; 32 35 … … 50 53 51 54 private const string ResultsParameterName = "Results"; 52 private const string BestValidationQualityParameterName = "BestValidationQuality"; 53 private const string BestValidationSolutionParameterName = "BestValidationSolution"; 55 private const string BestValidationQualityParameterName = "Best validation quality"; 56 private const string BestValidationSolutionParameterName = "Best validation solution"; 57 private const string BestSolutionGenerationParameterName = "Best solution generation"; 58 private const string BestSolutionInputvariableCountParameterName = "Variables used by best solution"; 59 private const string VariableFrequenciesParameterName = "VariableFrequencies"; 60 private const string VariableImpactsParameterName = "Variable Impacts"; 54 61 55 62 #region parameter properties … … 91 98 get { return (IValueLookupParameter<DoubleValue>)Parameters[LowerEstimationLimitParameterName]; } 92 99 } 100 public ILookupParameter<DataTable> VariableFrequenciesParameter { 101 get { return (ILookupParameter<DataTable>)Parameters[VariableFrequenciesParameterName]; } 102 } 93 103 94 104 public ILookupParameter<ResultCollection> ResultsParameter { … … 100 110 public ILookupParameter<SymbolicClassificationSolution> BestValidationSolutionParameter { 101 111 get { return (ILookupParameter<SymbolicClassificationSolution>)Parameters[BestValidationSolutionParameterName]; } 112 } 113 public ILookupParameter<IntValue> BestSolutionGenerationParameter { 114 get { return (ILookupParameter<IntValue>)Parameters[BestSolutionGenerationParameterName]; } 115 } 116 public ILookupParameter<DoubleMatrix> VariableImpactsParameter { 117 get { return (ILookupParameter<DoubleMatrix>)Parameters[VariableImpactsParameterName]; } 118 } 119 public ILookupParameter<IntValue> BestSolutionInputvariableCountParameter { 120 get { return (ILookupParameter<IntValue>)Parameters[BestSolutionInputvariableCountParameterName]; } 102 121 } 103 122 #endregion … … 140 159 get { return LowerEstimationLimitParameter.ActualValue; } 141 160 } 161 public DataTable VariableFrequencies { 162 get { return VariableFrequenciesParameter.ActualValue; } 163 } 142 164 143 165 public ResultCollection Results { … … 151 173 get { return BestValidationSolutionParameter.ActualValue; } 152 174 protected set { BestValidationSolutionParameter.ActualValue = value; } 175 } 176 public IntValue BestSolutionGeneration { 177 get { return BestSolutionGenerationParameter.ActualValue; } 178 protected set { BestSolutionGenerationParameter.ActualValue = value; } 179 } 180 public IntValue BestSolutionInputvariableCount { 181 get { return BestSolutionInputvariableCountParameter.ActualValue; } 182 protected set { BestSolutionInputvariableCountParameter.ActualValue = value; } 183 } 184 public DoubleMatrix VariableImpacts { 185 get { return VariableImpactsParameter.ActualValue; } 186 protected set { VariableImpactsParameter.ActualValue = value; } 153 187 } 154 188 #endregion … … 169 203 Parameters.Add(new ValueLookupParameter<DoubleValue>(UpperEstimationLimitParameterName, "The upper estimation limit that was set for the evaluation of the symbolic expression trees.")); 170 204 Parameters.Add(new ValueLookupParameter<DoubleValue>(LowerEstimationLimitParameterName, "The lower estimation limit that was set for the evaluation of the symbolic expression trees.")); 205 Parameters.Add(new LookupParameter<DataTable>(VariableFrequenciesParameterName, "The variable frequencies table to use for the calculation of variable impacts")); 171 206 172 207 Parameters.Add(new ValueLookupParameter<ResultCollection>(ResultsParameterName, "The results collection where the analysis values should be stored.")); 173 208 Parameters.Add(new LookupParameter<DoubleValue>(BestValidationQualityParameterName, "The validation quality of the best solution in the current run.")); 174 209 Parameters.Add(new LookupParameter<SymbolicClassificationSolution>(BestValidationSolutionParameterName, "The best solution on the validation data found in the current run.")); 210 Parameters.Add(new LookupParameter<IntValue>(BestSolutionGenerationParameterName, "The generation in which the best solution was found.")); 211 Parameters.Add(new LookupParameter<DoubleMatrix>(VariableImpactsParameterName, "The impacts of the input variables calculated during the run.")); 212 Parameters.Add(new LookupParameter<IntValue>(BestSolutionInputvariableCountParameterName, "The number of input variables used by the best solution.")); 213 175 214 } 176 215 … … 199 238 double quality = Evaluator.Evaluate(SymbolicExpressionTreeInterpreter, tree, 200 239 lowerEstimationLimit, upperEstimationLimit, ClassificationProblemData.Dataset, 201 targetVariable, ClassificationProblemData.SortedClassValues,rows);240 targetVariable, rows); 202 241 203 242 if ((Maximization.Value && quality > bestQuality) || … … 214 253 (!Maximization.Value && bestQuality < BestValidationQuality.Value); 215 254 if (newBest) { 255 double alpha, beta; 256 int trainingStart = ClassificationProblemData.TrainingSamplesStart.Value; 257 int trainingEnd = ClassificationProblemData.TrainingSamplesEnd.Value; 258 IEnumerable<int> trainingRows = Enumerable.Range(trainingStart, trainingEnd - trainingStart); 259 SymbolicRegressionScaledMeanSquaredErrorEvaluator.Calculate(SymbolicExpressionTreeInterpreter, bestTree, 260 lowerEstimationLimit, upperEstimationLimit, 261 ClassificationProblemData.Dataset, targetVariable, 262 trainingRows, out beta, out alpha); 263 264 // scale tree for solution 265 var scaledTree = SymbolicRegressionSolutionLinearScaler.Scale(bestTree, alpha, beta); 216 266 var model = new SymbolicRegressionModel((ISymbolicExpressionTreeInterpreter)SymbolicExpressionTreeInterpreter.Clone(), 217 bestTree);267 scaledTree); 218 268 219 269 if (BestValidationSolution == null) { … … 222 272 BestValidationSolution.Description = "Best solution on validation partition found over the whole run."; 223 273 BestValidationQuality = new DoubleValue(bestQuality); 274 BestSolutionGeneration = (IntValue)Generations.Clone(); 275 BestSolutionInputvariableCount = new IntValue(BestValidationSolution.Model.InputVariables.Count()); 276 224 277 Results.Add(new Result(BestValidationSolutionParameterName, BestValidationSolution)); 225 278 Results.Add(new Result(BestValidationQualityParameterName, BestValidationQuality)); 279 Results.Add(new Result(BestSolutionGenerationParameterName, BestSolutionGeneration)); 280 281 Results.Add(new Result(BestSolutionInputvariableCountParameterName, BestSolutionInputvariableCount)); 282 283 if (VariableFrequencies != null) { 284 VariableImpacts = CalculateVariableImpacts(VariableFrequencies); 285 Results.Add(new Result(VariableImpactsParameterName, VariableImpacts)); 286 } 226 287 227 288 } else { 228 289 BestValidationSolution.Model = model; 229 290 BestValidationQuality.Value = bestQuality; 291 BestSolutionGeneration.Value = Generations.Value; 292 BestSolutionInputvariableCount.Value = BestValidationSolution.Model.InputVariables.Count(); 293 294 if (VariableFrequencies != null) { 295 VariableImpacts = CalculateVariableImpacts(VariableFrequencies); 296 Results[VariableImpactsParameterName].Value = VariableImpacts; 297 } 230 298 } 231 299 } 232 300 return base.Apply(); 233 301 } 302 303 private static DoubleMatrix CalculateVariableImpacts(DataTable variableFrequencies) { 304 if (variableFrequencies != null) { 305 var impacts = new DoubleMatrix(variableFrequencies.Rows.Count, 1, new string[] { "Impact" }, variableFrequencies.Rows.Select(x => x.Name)); 306 impacts.SortableView = true; 307 int rowIndex = 0; 308 foreach (var dataRow in variableFrequencies.Rows) { 309 string variableName = dataRow.Name; 310 impacts[rowIndex++, 0] = dataRow.Values.Average(); 311 } 312 return impacts; 313 } else return new DoubleMatrix(1, 1); 314 } 234 315 } 235 316 } -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/Symbolic/Evaluators/SymbolicClassificationMeanSquaredErrorEvaluator.cs
r4366 r4391 20 20 #endregion 21 21 22 using System;23 using System.Collections.Generic;24 using System.Linq;25 using HeuristicLab.Common;26 22 using HeuristicLab.Core; 27 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;28 23 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable; 29 using HeuristicLab.Problems.DataAnalysis.Evaluators; 30 using HeuristicLab.Problems.DataAnalysis.Symbolic; 24 using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic; 31 25 32 26 namespace HeuristicLab.Problems.DataAnalysis.Classification { 33 27 [Item("SymbolicClassificationMeanSquaredErrorEvaluator", "Calculates the mean squared error of a symbolic classification solution.")] 34 28 [StorableClass] 35 public class SymbolicClassifacitionMeanSquaredErrorEvaluator : SingleObjectiveSymbolicClassificationEvaluator { 29 public class SymbolicClassifacitionMeanSquaredErrorEvaluator : SymbolicRegressionMeanSquaredErrorEvaluator, ISymbolicClassificationEvaluator { 30 public ClassificationProblemData ClassificationProblemData { 31 get { return (ClassificationProblemData)RegressionProblemData; } 32 } 36 33 37 34 public SymbolicClassifacitionMeanSquaredErrorEvaluator() 38 35 : base() { 39 36 } 40 41 public override double Evaluate(ISymbolicExpressionTreeInterpreter interpreter, SymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, Dataset dataset, string targetVariable, IEnumerable<double> sortedClassValues, IEnumerable<int> rows) {42 double mse = Calculate(interpreter, solution, lowerEstimationLimit, upperEstimationLimit, dataset, targetVariable, sortedClassValues, rows);43 return mse;44 }45 46 public static double Calculate(ISymbolicExpressionTreeInterpreter interpreter, SymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, Dataset dataset, string targetVariable, IEnumerable<double> sortedClassValues, IEnumerable<int> rows) {47 IEnumerable<double> estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, dataset, rows);48 IEnumerable<double> originalValues = dataset.GetEnumeratedVariableValues(targetVariable, rows);49 IEnumerator<double> originalEnumerator = originalValues.GetEnumerator();50 IEnumerator<double> estimatedEnumerator = estimatedValues.GetEnumerator();51 OnlineMeanSquaredErrorEvaluator mseEvaluator = new OnlineMeanSquaredErrorEvaluator();52 53 54 double firstClassValue = sortedClassValues.First();55 double lastClassValue = sortedClassValues.Last();56 while (originalEnumerator.MoveNext() && estimatedEnumerator.MoveNext()) {57 double estimated = estimatedEnumerator.Current;58 double original = originalEnumerator.Current;59 if (double.IsNaN(estimated))60 estimated = upperEstimationLimit;61 else if (estimated < original && original.IsAlmost(firstClassValue))62 estimated = original;63 else if (estimated > original && original.IsAlmost(lastClassValue))64 estimated = original;65 else66 estimated = Math.Min(upperEstimationLimit, Math.Max(lowerEstimationLimit, estimated));67 mseEvaluator.Add(original, estimated);68 }69 70 if (estimatedEnumerator.MoveNext() || originalEnumerator.MoveNext()) {71 throw new ArgumentException("Number of elements in original and estimated enumeration doesn't match.");72 } else {73 return mseEvaluator.MeanSquaredError;74 }75 }76 37 } 77 38 } -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/Symbolic/SymbolicClassificationProblem.cs
r4366 r4391 33 33 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable; 34 34 using HeuristicLab.PluginInfrastructure; 35 using HeuristicLab.Problems.DataAnalysis.Regression.Symbolic.Analyzers; 35 36 using HeuristicLab.Problems.DataAnalysis.Symbolic; 36 37 using HeuristicLab.Problems.DataAnalysis.Symbolic.Symbols; … … 228 229 Operators.AddRange(ApplicationManager.Manager.GetInstances<ISymbolicExpressionTreeOperator>().OfType<IOperator>()); 229 230 Operators.Add(new MinAverageMaxSymbolicExpressionTreeSizeAnalyzer()); 231 Operators.Add(new SymbolicRegressionVariableFrequencyAnalyzer()); 230 232 Operators.Add(new ValidationBestSymbolicClassificationSolutionAnalyzer()); 231 233 } … … 294 296 295 297 private void ParameterizeAnalyzers() { 296 foreach (ISymbolic ExpressionTreeAnalyzer analyzer in Operators.OfType<ISymbolicExpressionTreeAnalyzer>()) {298 foreach (ISymbolicRegressionAnalyzer analyzer in Operators.OfType<ISymbolicRegressionAnalyzer>()) { 297 299 analyzer.SymbolicExpressionTreeParameter.ActualName = SolutionCreator.SymbolicExpressionTreeParameter.ActualName; 298 300 var bestValidationSolutionAnalyzer = analyzer as ValidationBestSymbolicClassificationSolutionAnalyzer; … … 306 308 bestValidationSolutionAnalyzer.ValidationSamplesEndParameter.Value = ValidationSamplesEnd; 307 309 } 310 var varFreqAnalyzer = analyzer as SymbolicRegressionVariableFrequencyAnalyzer; 311 if (varFreqAnalyzer != null) { 312 varFreqAnalyzer.ProblemDataParameter.ActualName = ClassificationProblemDataParameter.Name; 313 } 308 314 } 309 315 } -
branches/HeuristicLab.Classification/HeuristicLab.Problems.DataAnalysis.Classification/3.3/Symbolic/SymbolicClassificationSolution.cs
r4366 r4391 31 31 namespace HeuristicLab.Problems.DataAnalysis.Classification { 32 32 /// <summary> 33 /// Represents a solution for a symbolic regression problem which can be visualized in the GUI.33 /// Represents a solution for a symbolic classification problem which can be visualized in the GUI. 34 34 /// </summary> 35 35 [Item("SymbolicClassificationSolution", "Represents a solution for a symbolic classification problem which can be visualized in the GUI.")] 36 36 [StorableClass] 37 public sealedclass SymbolicClassificationSolution : DataAnalysisSolution, IClassificationSolution {37 public class SymbolicClassificationSolution : DataAnalysisSolution, IClassificationSolution { 38 38 private SymbolicClassificationSolution() : base() { } 39 39 public SymbolicClassificationSolution(ClassificationProblemData problemData, SymbolicRegressionModel model, double lowerEstimationLimit, double upperEstimationLimit) … … 46 46 } 47 47 48 public new SymbolicRegressionModel Model { 49 get { return (SymbolicRegressionModel)base.Model; } 50 set { base.Model = value; } 51 } 52 48 53 public new ClassificationProblemData ProblemData { 49 54 get { return (ClassificationProblemData)base.ProblemData; } 50 55 set { base.ProblemData = value; } 51 }52 53 public new SymbolicRegressionModel Model {54 get { return (SymbolicRegressionModel)base.Model; }55 set { base.Model = value; }56 56 } 57 57 … … 65 65 } 66 66 67 protected List<double> estimatedValues; 68 public override IEnumerable<double> EstimatedValues { 69 get { 70 if (estimatedValues == null) RecalculateEstimatedValues(); 71 return estimatedValues.AsEnumerable(); 72 } 73 } 74 75 public override IEnumerable<double> EstimatedTrainingValues { 76 get { 77 if (estimatedValues == null) RecalculateEstimatedValues(); 78 int start = ProblemData.TrainingSamplesStart.Value; 79 int n = ProblemData.TrainingSamplesEnd.Value - start; 80 return estimatedValues.Skip(start).Take(n).ToList(); 81 } 82 } 83 84 public override IEnumerable<double> EstimatedTestValues { 85 get { 86 if (estimatedValues == null) RecalculateEstimatedValues(); 87 int start = ProblemData.TestSamplesStart.Value; 88 int n = ProblemData.TestSamplesEnd.Value - start; 89 return estimatedValues.Skip(start).Take(n).ToList(); 90 } 91 } 92 67 93 private void RecalculateClassIntermediates() { 68 int slices = 1000; 94 int slices = 100; 95 96 List<int> classInstances = (from classValue in ProblemData.Dataset.GetVariableValues(ProblemData.TargetVariable.Value) 97 group classValue by classValue into grouping 98 select grouping.Count()).ToList(); 69 99 70 100 List<KeyValuePair<double, double>> estimatedTargetValues = 71 (from row in Enumerable.Range(ProblemData.TrainingSamplesStart.Value, ProblemData.TrainingSamplesEnd.Value - ProblemData.TrainingSamplesStart.Value) 72 select new KeyValuePair<double, double>( 73 estimatedValues[row], 74 ProblemData.Dataset[ProblemData.TargetVariable.Value, row])).ToList(); 75 76 List<double> originalClasses = ProblemData.Dataset.GetVariableValues(ProblemData.TargetVariable.Value).Distinct().OrderBy(x => x).ToList(); 77 int numberOfClasses = originalClasses.Count; 78 79 double[] thresholds = new double[numberOfClasses + 1]; 101 (from row in Enumerable.Range(ProblemData.TrainingSamplesStart.Value, ProblemData.TrainingSamplesEnd.Value - ProblemData.TrainingSamplesStart.Value) 102 select new KeyValuePair<double, double>( 103 estimatedValues[row], 104 ProblemData.Dataset[ProblemData.TargetVariable.Value, row])).ToList(); 105 106 List<double> originalClasses = ProblemData.SortedClassValues.ToList(); 107 double[] thresholds = new double[ProblemData.NumberOfClasses + 1]; 80 108 thresholds[0] = double.NegativeInfinity; 81 109 thresholds[thresholds.Length - 1] = double.PositiveInfinity; 82 83 110 84 111 for (int i = 1; i < thresholds.Length - 1; i++) { … … 88 115 89 116 double bestThreshold = double.NaN; 90 double best Quality = double.NegativeInfinity;117 double bestClassificationScore = double.PositiveInfinity; 91 118 92 119 while (actualThreshold < originalClasses[i]) { 93 int truePosivites = 0; 94 int falsePosivites = 0; 95 int trueNegatives = 0; 96 int falseNegatives = 0; 120 double classificationScore = 0.0; 97 121 98 122 foreach (KeyValuePair<double, double> estimatedTarget in estimatedTargetValues) { … … 100 124 if (estimatedTarget.Value.IsAlmost(originalClasses[i - 1])) { 101 125 if (estimatedTarget.Key > lowerThreshold && estimatedTarget.Key < actualThreshold) 102 truePosivites++; 126 //true positive 127 classificationScore += ProblemData.MisclassificationMatrix[i - 1, i - 1] / classInstances[i - 1]; 103 128 else 104 falseNegatives++; 129 //false negative 130 classificationScore += ProblemData.MisclassificationMatrix[i, i - 1] / classInstances[i - 1]; 105 131 } 106 132 //all negatives 107 133 else { 108 134 if (estimatedTarget.Key > lowerThreshold && estimatedTarget.Key < actualThreshold) 109 falsePosivites++;135 classificationScore += ProblemData.MisclassificationMatrix[i - 1, i] / classInstances[i]; 110 136 else 111 trueNegatives++; 137 //true negative, consider only upper class 138 classificationScore += ProblemData.MisclassificationMatrix[i, i] / classInstances[i]; 112 139 } 113 140 } 114 115 //mkommend 30.08.2010 116 //matthews correlation coefficient taken from http://en.wikipedia.org/wiki/Matthews_correlation_coefficient 117 //MCC = [(TP * FP) - (FP * FN)] / sqrt((TP + FP) * (TP + FN) * (TN + FP) * (TN + FN)) 118 double dividend = truePosivites * falsePosivites - falsePosivites * falseNegatives; 119 double divisor = Math.Sqrt((truePosivites + falsePosivites) * (truePosivites + falsePosivites) * 120 (trueNegatives + falsePosivites) * (trueNegatives + falseNegatives)); 121 if (divisor == 0) 122 divisor = 1; 123 124 double mcc = dividend / divisor; 125 126 if (bestQuality < mcc) { 127 bestQuality = mcc; 141 if (classificationScore < bestClassificationScore) { 142 bestClassificationScore = classificationScore; 128 143 bestThreshold = actualThreshold; 129 144 } … … 149 164 actualThresholds = new List<double>(value); 150 165 OnThresholdsChanged(); 151 }152 }153 154 private List<double> estimatedValues;155 public override IEnumerable<double> EstimatedValues {156 get {157 if (estimatedValues == null) RecalculateEstimatedValues();158 return estimatedValues.AsEnumerable();159 166 } 160 167 } … … 172 179 } 173 180 174 public override IEnumerable<double> EstimatedTrainingValues {175 get {176 if (estimatedValues == null) RecalculateEstimatedValues();177 int start = ProblemData.TrainingSamplesStart.Value;178 int n = ProblemData.TrainingSamplesEnd.Value - start;179 return estimatedValues.Skip(start).Take(n).ToList();180 }181 }182 181 public IEnumerable<double> EstimatedTrainingClassValues { 183 182 get { … … 188 187 } 189 188 190 public override IEnumerable<double> EstimatedTestValues {191 get {192 if (estimatedValues == null) RecalculateEstimatedValues();193 int start = ProblemData.TestSamplesStart.Value;194 int n = ProblemData.TestSamplesEnd.Value - start;195 return estimatedValues.Skip(start).Take(n).ToList();196 }197 }198 189 public IEnumerable<double> EstimatedTestClassValues { 199 190 get {
Note: See TracChangeset
for help on using the changeset viewer.