1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 2002-2015 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 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>
|
---|
32 | /// Represents a neural network model for regression and classification
|
---|
33 | /// </summary>
|
---|
34 | [StorableClass]
|
---|
35 | [Item("NeuralNetworkModel", "Represents a neural network for regression and classification.")]
|
---|
36 | public sealed class NeuralNetworkModel : NamedItem, INeuralNetworkModel {
|
---|
37 |
|
---|
38 | private alglib.multilayerperceptron multiLayerPerceptron;
|
---|
39 | public alglib.multilayerperceptron MultiLayerPerceptron {
|
---|
40 | get { return multiLayerPerceptron; }
|
---|
41 | set {
|
---|
42 | if (value != multiLayerPerceptron) {
|
---|
43 | if (value == null) throw new ArgumentNullException();
|
---|
44 | multiLayerPerceptron = value;
|
---|
45 | OnChanged(EventArgs.Empty);
|
---|
46 | }
|
---|
47 | }
|
---|
48 | }
|
---|
49 |
|
---|
50 | [Storable]
|
---|
51 | private string targetVariable;
|
---|
52 | [Storable]
|
---|
53 | private string[] allowedInputVariables;
|
---|
54 | [Storable]
|
---|
55 | private double[] classValues;
|
---|
56 | [StorableConstructor]
|
---|
57 | private NeuralNetworkModel(bool deserializing)
|
---|
58 | : base(deserializing) {
|
---|
59 | if (deserializing)
|
---|
60 | multiLayerPerceptron = new alglib.multilayerperceptron();
|
---|
61 | }
|
---|
62 | private NeuralNetworkModel(NeuralNetworkModel original, Cloner cloner)
|
---|
63 | : base(original, cloner) {
|
---|
64 | multiLayerPerceptron = new alglib.multilayerperceptron();
|
---|
65 | multiLayerPerceptron.innerobj.chunks = (double[,])original.multiLayerPerceptron.innerobj.chunks.Clone();
|
---|
66 | multiLayerPerceptron.innerobj.columnmeans = (double[])original.multiLayerPerceptron.innerobj.columnmeans.Clone();
|
---|
67 | multiLayerPerceptron.innerobj.columnsigmas = (double[])original.multiLayerPerceptron.innerobj.columnsigmas.Clone();
|
---|
68 | multiLayerPerceptron.innerobj.derror = (double[])original.multiLayerPerceptron.innerobj.derror.Clone();
|
---|
69 | multiLayerPerceptron.innerobj.dfdnet = (double[])original.multiLayerPerceptron.innerobj.dfdnet.Clone();
|
---|
70 | multiLayerPerceptron.innerobj.neurons = (double[])original.multiLayerPerceptron.innerobj.neurons.Clone();
|
---|
71 | multiLayerPerceptron.innerobj.nwbuf = (double[])original.multiLayerPerceptron.innerobj.nwbuf.Clone();
|
---|
72 | multiLayerPerceptron.innerobj.structinfo = (int[])original.multiLayerPerceptron.innerobj.structinfo.Clone();
|
---|
73 | multiLayerPerceptron.innerobj.weights = (double[])original.multiLayerPerceptron.innerobj.weights.Clone();
|
---|
74 | multiLayerPerceptron.innerobj.x = (double[])original.multiLayerPerceptron.innerobj.x.Clone();
|
---|
75 | multiLayerPerceptron.innerobj.y = (double[])original.multiLayerPerceptron.innerobj.y.Clone();
|
---|
76 | targetVariable = original.targetVariable;
|
---|
77 | allowedInputVariables = (string[])original.allowedInputVariables.Clone();
|
---|
78 | if (original.classValues != null)
|
---|
79 | this.classValues = (double[])original.classValues.Clone();
|
---|
80 | }
|
---|
81 | public NeuralNetworkModel(alglib.multilayerperceptron multiLayerPerceptron, string targetVariable, IEnumerable<string> allowedInputVariables, double[] classValues = null)
|
---|
82 | : base() {
|
---|
83 | this.name = ItemName;
|
---|
84 | this.description = ItemDescription;
|
---|
85 | this.multiLayerPerceptron = multiLayerPerceptron;
|
---|
86 | this.targetVariable = targetVariable;
|
---|
87 | this.allowedInputVariables = allowedInputVariables.ToArray();
|
---|
88 | if (classValues != null)
|
---|
89 | this.classValues = (double[])classValues.Clone();
|
---|
90 | }
|
---|
91 |
|
---|
92 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
93 | return new NeuralNetworkModel(this, cloner);
|
---|
94 | }
|
---|
95 |
|
---|
96 | public IEnumerable<double> GetEstimatedValues(Dataset dataset, IEnumerable<int> rows) {
|
---|
97 | double[,] inputData = AlglibUtil.PrepareInputMatrix(dataset, allowedInputVariables, rows);
|
---|
98 |
|
---|
99 | int n = inputData.GetLength(0);
|
---|
100 | int columns = inputData.GetLength(1);
|
---|
101 | double[] x = new double[columns];
|
---|
102 | double[] y = new double[1];
|
---|
103 |
|
---|
104 | for (int row = 0; row < n; row++) {
|
---|
105 | for (int column = 0; column < columns; column++) {
|
---|
106 | x[column] = inputData[row, column];
|
---|
107 | }
|
---|
108 | alglib.mlpprocess(multiLayerPerceptron, x, ref y);
|
---|
109 | yield return y[0];
|
---|
110 | }
|
---|
111 | }
|
---|
112 |
|
---|
113 | public IEnumerable<double> GetEstimatedClassValues(Dataset dataset, IEnumerable<int> rows) {
|
---|
114 | double[,] inputData = AlglibUtil.PrepareInputMatrix(dataset, allowedInputVariables, rows);
|
---|
115 |
|
---|
116 | int n = inputData.GetLength(0);
|
---|
117 | int columns = inputData.GetLength(1);
|
---|
118 | double[] x = new double[columns];
|
---|
119 | double[] y = new double[classValues.Length];
|
---|
120 |
|
---|
121 | for (int row = 0; row < n; row++) {
|
---|
122 | for (int column = 0; column < columns; column++) {
|
---|
123 | x[column] = inputData[row, column];
|
---|
124 | }
|
---|
125 | alglib.mlpprocess(multiLayerPerceptron, x, ref y);
|
---|
126 | // find class for with the largest probability value
|
---|
127 | int maxProbClassIndex = 0;
|
---|
128 | double maxProb = y[0];
|
---|
129 | for (int i = 1; i < y.Length; i++) {
|
---|
130 | if (maxProb < y[i]) {
|
---|
131 | maxProb = y[i];
|
---|
132 | maxProbClassIndex = i;
|
---|
133 | }
|
---|
134 | }
|
---|
135 | yield return classValues[maxProbClassIndex];
|
---|
136 | }
|
---|
137 | }
|
---|
138 |
|
---|
139 | public INeuralNetworkRegressionSolution CreateRegressionSolution(IRegressionProblemData problemData) {
|
---|
140 | return new NeuralNetworkRegressionSolution(new RegressionProblemData(problemData), this);
|
---|
141 | }
|
---|
142 | IRegressionSolution IRegressionModel.CreateRegressionSolution(IRegressionProblemData problemData) {
|
---|
143 | return CreateRegressionSolution(problemData);
|
---|
144 | }
|
---|
145 | public INeuralNetworkClassificationSolution CreateClassificationSolution(IClassificationProblemData problemData) {
|
---|
146 | return new NeuralNetworkClassificationSolution(new ClassificationProblemData(problemData), this);
|
---|
147 | }
|
---|
148 | IClassificationSolution IClassificationModel.CreateClassificationSolution(IClassificationProblemData problemData) {
|
---|
149 | return CreateClassificationSolution(problemData);
|
---|
150 | }
|
---|
151 |
|
---|
152 | #region events
|
---|
153 | public event EventHandler Changed;
|
---|
154 | private void OnChanged(EventArgs e) {
|
---|
155 | var handlers = Changed;
|
---|
156 | if (handlers != null)
|
---|
157 | handlers(this, e);
|
---|
158 | }
|
---|
159 | #endregion
|
---|
160 |
|
---|
161 | #region persistence
|
---|
162 | [Storable]
|
---|
163 | private double[,] MultiLayerPerceptronChunks {
|
---|
164 | get {
|
---|
165 | return multiLayerPerceptron.innerobj.chunks;
|
---|
166 | }
|
---|
167 | set {
|
---|
168 | multiLayerPerceptron.innerobj.chunks = value;
|
---|
169 | }
|
---|
170 | }
|
---|
171 | [Storable]
|
---|
172 | private double[] MultiLayerPerceptronColumnMeans {
|
---|
173 | get {
|
---|
174 | return multiLayerPerceptron.innerobj.columnmeans;
|
---|
175 | }
|
---|
176 | set {
|
---|
177 | multiLayerPerceptron.innerobj.columnmeans = value;
|
---|
178 | }
|
---|
179 | }
|
---|
180 | [Storable]
|
---|
181 | private double[] MultiLayerPerceptronColumnSigmas {
|
---|
182 | get {
|
---|
183 | return multiLayerPerceptron.innerobj.columnsigmas;
|
---|
184 | }
|
---|
185 | set {
|
---|
186 | multiLayerPerceptron.innerobj.columnsigmas = value;
|
---|
187 | }
|
---|
188 | }
|
---|
189 | [Storable]
|
---|
190 | private double[] MultiLayerPerceptronDError {
|
---|
191 | get {
|
---|
192 | return multiLayerPerceptron.innerobj.derror;
|
---|
193 | }
|
---|
194 | set {
|
---|
195 | multiLayerPerceptron.innerobj.derror = value;
|
---|
196 | }
|
---|
197 | }
|
---|
198 | [Storable]
|
---|
199 | private double[] MultiLayerPerceptronDfdnet {
|
---|
200 | get {
|
---|
201 | return multiLayerPerceptron.innerobj.dfdnet;
|
---|
202 | }
|
---|
203 | set {
|
---|
204 | multiLayerPerceptron.innerobj.dfdnet = value;
|
---|
205 | }
|
---|
206 | }
|
---|
207 | [Storable]
|
---|
208 | private double[] MultiLayerPerceptronNeurons {
|
---|
209 | get {
|
---|
210 | return multiLayerPerceptron.innerobj.neurons;
|
---|
211 | }
|
---|
212 | set {
|
---|
213 | multiLayerPerceptron.innerobj.neurons = value;
|
---|
214 | }
|
---|
215 | }
|
---|
216 | [Storable]
|
---|
217 | private double[] MultiLayerPerceptronNwbuf {
|
---|
218 | get {
|
---|
219 | return multiLayerPerceptron.innerobj.nwbuf;
|
---|
220 | }
|
---|
221 | set {
|
---|
222 | multiLayerPerceptron.innerobj.nwbuf = value;
|
---|
223 | }
|
---|
224 | }
|
---|
225 | [Storable]
|
---|
226 | private int[] MultiLayerPerceptronStuctinfo {
|
---|
227 | get {
|
---|
228 | return multiLayerPerceptron.innerobj.structinfo;
|
---|
229 | }
|
---|
230 | set {
|
---|
231 | multiLayerPerceptron.innerobj.structinfo = value;
|
---|
232 | }
|
---|
233 | }
|
---|
234 | [Storable]
|
---|
235 | private double[] MultiLayerPerceptronWeights {
|
---|
236 | get {
|
---|
237 | return multiLayerPerceptron.innerobj.weights;
|
---|
238 | }
|
---|
239 | set {
|
---|
240 | multiLayerPerceptron.innerobj.weights = value;
|
---|
241 | }
|
---|
242 | }
|
---|
243 | [Storable]
|
---|
244 | private double[] MultiLayerPerceptronX {
|
---|
245 | get {
|
---|
246 | return multiLayerPerceptron.innerobj.x;
|
---|
247 | }
|
---|
248 | set {
|
---|
249 | multiLayerPerceptron.innerobj.x = value;
|
---|
250 | }
|
---|
251 | }
|
---|
252 | [Storable]
|
---|
253 | private double[] MultiLayerPerceptronY {
|
---|
254 | get {
|
---|
255 | return multiLayerPerceptron.innerobj.y;
|
---|
256 | }
|
---|
257 | set {
|
---|
258 | multiLayerPerceptron.innerobj.y = value;
|
---|
259 | }
|
---|
260 | }
|
---|
261 | #endregion
|
---|
262 | }
|
---|
263 | }
|
---|