[1806] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
| 3 | * Copyright (C) 2002-2009 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 |
|
---|
| 22 | using System;
|
---|
| 23 | using System.Collections.Generic;
|
---|
| 24 | using System.Linq;
|
---|
| 25 | using System.Text;
|
---|
| 26 | using HeuristicLab.Core;
|
---|
| 27 | using HeuristicLab.Data;
|
---|
| 28 | using HeuristicLab.DataAnalysis;
|
---|
[1848] | 29 | using System.Threading;
|
---|
[2415] | 30 | using SVM;
|
---|
[1806] | 31 |
|
---|
| 32 | namespace HeuristicLab.SupportVectorMachines {
|
---|
| 33 | public class SupportVectorCreator : OperatorBase {
|
---|
[1848] | 34 | private Thread trainingThread;
|
---|
| 35 | private object locker = new object();
|
---|
| 36 | private bool abortRequested = false;
|
---|
[1806] | 37 |
|
---|
| 38 | public SupportVectorCreator()
|
---|
| 39 | : base() {
|
---|
| 40 | //Dataset infos
|
---|
| 41 | AddVariableInfo(new VariableInfo("Dataset", "Dataset with all samples on which to apply the function", typeof(Dataset), VariableKind.In));
|
---|
[2440] | 42 | AddVariableInfo(new VariableInfo("TargetVariable", "Name of the target variable", typeof(StringData), VariableKind.In));
|
---|
[2421] | 43 | AddVariableInfo(new VariableInfo("InputVariables", "List of allowed input variable names", typeof(ItemList), VariableKind.In));
|
---|
[1806] | 44 | AddVariableInfo(new VariableInfo("SamplesStart", "Start index of samples in dataset to evaluate", typeof(IntData), VariableKind.In));
|
---|
| 45 | AddVariableInfo(new VariableInfo("SamplesEnd", "End index of samples in dataset to evaluate", typeof(IntData), VariableKind.In));
|
---|
[2349] | 46 | AddVariableInfo(new VariableInfo("MaxTimeOffset", "(optional) Maximal time offset for time-series prognosis", typeof(IntData), VariableKind.In));
|
---|
| 47 | AddVariableInfo(new VariableInfo("MinTimeOffset", "(optional) Minimal time offset for time-series prognosis", typeof(IntData), VariableKind.In));
|
---|
[1806] | 48 |
|
---|
| 49 | //SVM parameters
|
---|
| 50 | AddVariableInfo(new VariableInfo("SVMType", "String describing which SVM type is used. Valid inputs are: C_SVC, NU_SVC, ONE_CLASS, EPSILON_SVR, NU_SVR",
|
---|
| 51 | typeof(StringData), VariableKind.In));
|
---|
| 52 | AddVariableInfo(new VariableInfo("SVMKernelType", "String describing which SVM kernel is used. Valid inputs are: LINEAR, POLY, RBF, SIGMOID, PRECOMPUTED",
|
---|
| 53 | typeof(StringData), VariableKind.In));
|
---|
| 54 | AddVariableInfo(new VariableInfo("SVMCost", "Cost parameter (C) of C-SVC, epsilon-SVR and nu-SVR", typeof(DoubleData), VariableKind.In));
|
---|
| 55 | AddVariableInfo(new VariableInfo("SVMNu", "Nu parameter of nu-SVC, one-class SVM and nu-SVR", typeof(DoubleData), VariableKind.In));
|
---|
| 56 | AddVariableInfo(new VariableInfo("SVMGamma", "Gamma parameter in kernel function", typeof(DoubleData), VariableKind.In));
|
---|
[1837] | 57 | AddVariableInfo(new VariableInfo("SVMModel", "Represent the model learned by the SVM", typeof(SVMModel), VariableKind.New | VariableKind.Out));
|
---|
[1848] | 58 | }
|
---|
[1806] | 59 |
|
---|
[1848] | 60 | public override void Abort() {
|
---|
| 61 | abortRequested = true;
|
---|
| 62 | lock (locker) {
|
---|
| 63 | if (trainingThread != null && trainingThread.ThreadState == ThreadState.Running) {
|
---|
| 64 | trainingThread.Abort();
|
---|
| 65 | }
|
---|
| 66 | }
|
---|
[1806] | 67 | }
|
---|
| 68 |
|
---|
| 69 | public override IOperation Apply(IScope scope) {
|
---|
[1848] | 70 | abortRequested = false;
|
---|
[1806] | 71 | Dataset dataset = GetVariableValue<Dataset>("Dataset", scope, true);
|
---|
[2440] | 72 | string targetVariable = GetVariableValue<StringData>("TargetVariable", scope, true).Data;
|
---|
| 73 | ItemList inputVariables = GetVariableValue<ItemList>("InputVariables", scope, true);
|
---|
| 74 | var inputVariableNames = from x in inputVariables
|
---|
[2421] | 75 | select ((StringData)x).Data;
|
---|
[1806] | 76 | int start = GetVariableValue<IntData>("SamplesStart", scope, true).Data;
|
---|
| 77 | int end = GetVariableValue<IntData>("SamplesEnd", scope, true).Data;
|
---|
[2349] | 78 | IntData maxTimeOffsetData = GetVariableValue<IntData>("MaxTimeOffset", scope, true, false);
|
---|
| 79 | int maxTimeOffset = maxTimeOffsetData == null ? 0 : maxTimeOffsetData.Data;
|
---|
| 80 | IntData minTimeOffsetData = GetVariableValue<IntData>("MinTimeOffset", scope, true, false);
|
---|
| 81 | int minTimeOffset = minTimeOffsetData == null ? 0 : minTimeOffsetData.Data;
|
---|
[1806] | 82 | string svmType = GetVariableValue<StringData>("SVMType", scope, true).Data;
|
---|
| 83 | string svmKernelType = GetVariableValue<StringData>("SVMKernelType", scope, true).Data;
|
---|
| 84 |
|
---|
[2550] | 85 | double cost = GetVariableValue<DoubleData>("SVMCost", scope, true).Data;
|
---|
| 86 | double nu = GetVariableValue<DoubleData>("SVMNu", scope, true).Data;
|
---|
| 87 | double gamma = GetVariableValue<DoubleData>("SVMGamma", scope, true).Data;
|
---|
[1806] | 88 |
|
---|
[2550] | 89 | SVMModel modelData = null;
|
---|
[1848] | 90 | lock (locker) {
|
---|
| 91 | if (!abortRequested) {
|
---|
| 92 | trainingThread = new Thread(() => {
|
---|
[2550] | 93 | modelData = TrainModel(dataset, targetVariable, inputVariableNames,
|
---|
| 94 | start, end, minTimeOffset, maxTimeOffset,
|
---|
| 95 | svmType, svmKernelType,
|
---|
| 96 | cost, nu, gamma);
|
---|
[1848] | 97 | });
|
---|
| 98 | trainingThread.Start();
|
---|
| 99 | }
|
---|
| 100 | }
|
---|
[2163] | 101 | if (!abortRequested) {
|
---|
| 102 | trainingThread.Join();
|
---|
| 103 | trainingThread = null;
|
---|
| 104 | }
|
---|
[2550] | 105 |
|
---|
| 106 |
|
---|
| 107 | if (!abortRequested) {
|
---|
| 108 | //persist variables in scope
|
---|
| 109 | scope.AddVariable(new Variable(scope.TranslateName("SVMModel"), modelData));
|
---|
| 110 | return null;
|
---|
| 111 | } else {
|
---|
| 112 | return new AtomicOperation(this, scope);
|
---|
| 113 | }
|
---|
| 114 | }
|
---|
| 115 |
|
---|
| 116 | public static SVMModel TrainRegressionModel(
|
---|
| 117 | Dataset dataset, string targetVariable, IEnumerable<string> inputVariables,
|
---|
| 118 | int start, int end,
|
---|
| 119 | double cost, double nu, double gamma) {
|
---|
| 120 | return TrainModel(dataset, targetVariable, inputVariables, start, end, 0, 0, "NU_SVR", "RBF", cost, nu, gamma);
|
---|
| 121 | }
|
---|
| 122 |
|
---|
| 123 | public static SVMModel TrainModel(
|
---|
| 124 | Dataset dataset, string targetVariable, IEnumerable<string> inputVariables,
|
---|
| 125 | int start, int end,
|
---|
| 126 | int minTimeOffset, int maxTimeOffset,
|
---|
| 127 | string svmType, string kernelType,
|
---|
| 128 | double cost, double nu, double gamma) {
|
---|
| 129 | int targetVariableIndex = dataset.GetVariableIndex(targetVariable);
|
---|
| 130 |
|
---|
| 131 | //extract SVM parameters from scope and set them
|
---|
| 132 | SVM.Parameter parameter = new SVM.Parameter();
|
---|
| 133 | parameter.SvmType = (SVM.SvmType)Enum.Parse(typeof(SVM.SvmType), svmType, true);
|
---|
| 134 | parameter.KernelType = (SVM.KernelType)Enum.Parse(typeof(SVM.KernelType), kernelType, true);
|
---|
| 135 | parameter.C = cost;
|
---|
| 136 | parameter.Nu = nu;
|
---|
| 137 | parameter.Gamma = gamma;
|
---|
| 138 | parameter.CacheSize = 500;
|
---|
| 139 | parameter.Probability = false;
|
---|
| 140 |
|
---|
| 141 |
|
---|
| 142 | SVM.Problem problem = SVMHelper.CreateSVMProblem(dataset, targetVariableIndex, inputVariables, start, end, minTimeOffset, maxTimeOffset);
|
---|
| 143 | SVM.RangeTransform rangeTransform = SVM.RangeTransform.Compute(problem);
|
---|
| 144 | SVM.Problem scaledProblem = rangeTransform.Scale(problem);
|
---|
| 145 | var model = new SVMModel();
|
---|
| 146 |
|
---|
| 147 | model.Model = SVM.Training.Train(scaledProblem, parameter);
|
---|
| 148 | model.RangeTransform = rangeTransform;
|
---|
| 149 |
|
---|
[1848] | 150 | return model;
|
---|
| 151 | }
|
---|
[1806] | 152 | }
|
---|
| 153 | }
|
---|