[6240] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[9456] | 3 | * Copyright (C) 2002-2013 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[6240] | 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 HeuristicLab.Common;
|
---|
| 26 | using HeuristicLab.Core;
|
---|
| 27 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
| 28 | using HeuristicLab.Problems.DataAnalysis;
|
---|
| 29 |
|
---|
| 30 | namespace HeuristicLab.Algorithms.DataAnalysis {
|
---|
| 31 | /// <summary>
|
---|
[6241] | 32 | /// Represents a random forest model for regression and classification
|
---|
[6240] | 33 | /// </summary>
|
---|
| 34 | [StorableClass]
|
---|
[6241] | 35 | [Item("RandomForestModel", "Represents a random forest for regression and classification.")]
|
---|
| 36 | public sealed class RandomForestModel : NamedItem, IRandomForestModel {
|
---|
[6240] | 37 |
|
---|
| 38 | private alglib.decisionforest randomForest;
|
---|
| 39 | public alglib.decisionforest RandomForest {
|
---|
| 40 | get { return randomForest; }
|
---|
| 41 | set {
|
---|
| 42 | if (value != randomForest) {
|
---|
| 43 | if (value == null) throw new ArgumentNullException();
|
---|
| 44 | randomForest = value;
|
---|
| 45 | OnChanged(EventArgs.Empty);
|
---|
| 46 | }
|
---|
| 47 | }
|
---|
| 48 | }
|
---|
| 49 |
|
---|
| 50 | [Storable]
|
---|
| 51 | private string targetVariable;
|
---|
| 52 | [Storable]
|
---|
| 53 | private string[] allowedInputVariables;
|
---|
[6241] | 54 | [Storable]
|
---|
| 55 | private double[] classValues;
|
---|
[6240] | 56 | [StorableConstructor]
|
---|
[6241] | 57 | private RandomForestModel(bool deserializing)
|
---|
[6240] | 58 | : base(deserializing) {
|
---|
| 59 | if (deserializing)
|
---|
| 60 | randomForest = new alglib.decisionforest();
|
---|
| 61 | }
|
---|
[6241] | 62 | private RandomForestModel(RandomForestModel original, Cloner cloner)
|
---|
[6240] | 63 | : base(original, cloner) {
|
---|
| 64 | randomForest = new alglib.decisionforest();
|
---|
| 65 | randomForest.innerobj.bufsize = original.randomForest.innerobj.bufsize;
|
---|
| 66 | randomForest.innerobj.nclasses = original.randomForest.innerobj.nclasses;
|
---|
| 67 | randomForest.innerobj.ntrees = original.randomForest.innerobj.ntrees;
|
---|
| 68 | randomForest.innerobj.nvars = original.randomForest.innerobj.nvars;
|
---|
| 69 | randomForest.innerobj.trees = (double[])original.randomForest.innerobj.trees.Clone();
|
---|
| 70 | targetVariable = original.targetVariable;
|
---|
| 71 | allowedInputVariables = (string[])original.allowedInputVariables.Clone();
|
---|
[6241] | 72 | if (original.classValues != null)
|
---|
| 73 | this.classValues = (double[])original.classValues.Clone();
|
---|
[6240] | 74 | }
|
---|
[6241] | 75 | public RandomForestModel(alglib.decisionforest randomForest, string targetVariable, IEnumerable<string> allowedInputVariables, double[] classValues = null)
|
---|
[6240] | 76 | : base() {
|
---|
| 77 | this.name = ItemName;
|
---|
| 78 | this.description = ItemDescription;
|
---|
| 79 | this.randomForest = randomForest;
|
---|
| 80 | this.targetVariable = targetVariable;
|
---|
| 81 | this.allowedInputVariables = allowedInputVariables.ToArray();
|
---|
[6241] | 82 | if (classValues != null)
|
---|
| 83 | this.classValues = (double[])classValues.Clone();
|
---|
[6240] | 84 | }
|
---|
| 85 |
|
---|
| 86 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
[6241] | 87 | return new RandomForestModel(this, cloner);
|
---|
[6240] | 88 | }
|
---|
| 89 |
|
---|
| 90 | public IEnumerable<double> GetEstimatedValues(Dataset dataset, IEnumerable<int> rows) {
|
---|
| 91 | double[,] inputData = AlglibUtil.PrepareInputMatrix(dataset, allowedInputVariables, rows);
|
---|
| 92 |
|
---|
| 93 | int n = inputData.GetLength(0);
|
---|
| 94 | int columns = inputData.GetLength(1);
|
---|
| 95 | double[] x = new double[columns];
|
---|
| 96 | double[] y = new double[1];
|
---|
| 97 |
|
---|
| 98 | for (int row = 0; row < n; row++) {
|
---|
| 99 | for (int column = 0; column < columns; column++) {
|
---|
| 100 | x[column] = inputData[row, column];
|
---|
| 101 | }
|
---|
| 102 | alglib.dfprocess(randomForest, x, ref y);
|
---|
| 103 | yield return y[0];
|
---|
| 104 | }
|
---|
| 105 | }
|
---|
| 106 |
|
---|
[6241] | 107 | public IEnumerable<double> GetEstimatedClassValues(Dataset dataset, IEnumerable<int> rows) {
|
---|
| 108 | double[,] inputData = AlglibUtil.PrepareInputMatrix(dataset, allowedInputVariables, rows);
|
---|
| 109 |
|
---|
| 110 | int n = inputData.GetLength(0);
|
---|
| 111 | int columns = inputData.GetLength(1);
|
---|
| 112 | double[] x = new double[columns];
|
---|
| 113 | double[] y = new double[randomForest.innerobj.nclasses];
|
---|
| 114 |
|
---|
| 115 | for (int row = 0; row < n; row++) {
|
---|
| 116 | for (int column = 0; column < columns; column++) {
|
---|
| 117 | x[column] = inputData[row, column];
|
---|
| 118 | }
|
---|
| 119 | alglib.dfprocess(randomForest, x, ref y);
|
---|
| 120 | // find class for with the largest probability value
|
---|
| 121 | int maxProbClassIndex = 0;
|
---|
| 122 | double maxProb = y[0];
|
---|
| 123 | for (int i = 1; i < y.Length; i++) {
|
---|
| 124 | if (maxProb < y[i]) {
|
---|
| 125 | maxProb = y[i];
|
---|
| 126 | maxProbClassIndex = i;
|
---|
| 127 | }
|
---|
| 128 | }
|
---|
| 129 | yield return classValues[maxProbClassIndex];
|
---|
| 130 | }
|
---|
| 131 | }
|
---|
| 132 |
|
---|
[6603] | 133 | public IRandomForestRegressionSolution CreateRegressionSolution(IRegressionProblemData problemData) {
|
---|
[8528] | 134 | return new RandomForestRegressionSolution(new RegressionProblemData(problemData), this);
|
---|
[6603] | 135 | }
|
---|
| 136 | IRegressionSolution IRegressionModel.CreateRegressionSolution(IRegressionProblemData problemData) {
|
---|
| 137 | return CreateRegressionSolution(problemData);
|
---|
| 138 | }
|
---|
[6604] | 139 | public IRandomForestClassificationSolution CreateClassificationSolution(IClassificationProblemData problemData) {
|
---|
[8528] | 140 | return new RandomForestClassificationSolution(new ClassificationProblemData(problemData), this);
|
---|
[6604] | 141 | }
|
---|
| 142 | IClassificationSolution IClassificationModel.CreateClassificationSolution(IClassificationProblemData problemData) {
|
---|
| 143 | return CreateClassificationSolution(problemData);
|
---|
| 144 | }
|
---|
[6603] | 145 |
|
---|
[6240] | 146 | #region events
|
---|
| 147 | public event EventHandler Changed;
|
---|
| 148 | private void OnChanged(EventArgs e) {
|
---|
| 149 | var handlers = Changed;
|
---|
| 150 | if (handlers != null)
|
---|
| 151 | handlers(this, e);
|
---|
| 152 | }
|
---|
| 153 | #endregion
|
---|
| 154 |
|
---|
| 155 | #region persistence
|
---|
| 156 | [Storable]
|
---|
| 157 | private int RandomForestBufSize {
|
---|
| 158 | get {
|
---|
| 159 | return randomForest.innerobj.bufsize;
|
---|
| 160 | }
|
---|
| 161 | set {
|
---|
| 162 | randomForest.innerobj.bufsize = value;
|
---|
| 163 | }
|
---|
| 164 | }
|
---|
| 165 | [Storable]
|
---|
| 166 | private int RandomForestNClasses {
|
---|
| 167 | get {
|
---|
| 168 | return randomForest.innerobj.nclasses;
|
---|
| 169 | }
|
---|
| 170 | set {
|
---|
| 171 | randomForest.innerobj.nclasses = value;
|
---|
| 172 | }
|
---|
| 173 | }
|
---|
| 174 | [Storable]
|
---|
| 175 | private int RandomForestNTrees {
|
---|
| 176 | get {
|
---|
| 177 | return randomForest.innerobj.ntrees;
|
---|
| 178 | }
|
---|
| 179 | set {
|
---|
| 180 | randomForest.innerobj.ntrees = value;
|
---|
| 181 | }
|
---|
| 182 | }
|
---|
| 183 | [Storable]
|
---|
| 184 | private int RandomForestNVars {
|
---|
| 185 | get {
|
---|
| 186 | return randomForest.innerobj.nvars;
|
---|
| 187 | }
|
---|
| 188 | set {
|
---|
| 189 | randomForest.innerobj.nvars = value;
|
---|
| 190 | }
|
---|
| 191 | }
|
---|
| 192 | [Storable]
|
---|
| 193 | private double[] RandomForestTrees {
|
---|
| 194 | get {
|
---|
| 195 | return randomForest.innerobj.trees;
|
---|
| 196 | }
|
---|
| 197 | set {
|
---|
| 198 | randomForest.innerobj.trees = value;
|
---|
| 199 | }
|
---|
| 200 | }
|
---|
| 201 | #endregion
|
---|
| 202 | }
|
---|
| 203 | }
|
---|