[15748] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
| 3 | * Copyright (C) 2002-2018 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.Encodings.SymbolicExpressionTreeEncoding;
|
---|
| 28 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
| 29 | using HeuristicLab.Problems.DataAnalysis;
|
---|
| 30 | using HeuristicLab.Problems.DataAnalysis.Symbolic;
|
---|
| 31 |
|
---|
| 32 | namespace HeuristicLab.Algorithms.DataAnalysis {
|
---|
| 33 | [StorableClass]
|
---|
| 34 | [Item("BayesianNonlinearRegressionModel", "")]
|
---|
| 35 | public sealed class BayesianNonlinearRegressionModel : RegressionModel, IConfidenceRegressionModel {
|
---|
| 36 | private const int SAMPLE_SIZE = 100;
|
---|
| 37 |
|
---|
| 38 |
|
---|
| 39 | [Storable]
|
---|
| 40 | public ISymbolicExpressionTree Tree {
|
---|
| 41 | get; private set;
|
---|
| 42 | }
|
---|
| 43 |
|
---|
| 44 | private double[][] parameterEmpiricalDistribution;
|
---|
| 45 | public IEnumerable<double[]> ParameterEmpiricalDistribution {
|
---|
| 46 | get { return parameterEmpiricalDistribution; }
|
---|
| 47 | }
|
---|
| 48 |
|
---|
| 49 | public ISymbolicDataAnalysisExpressionTreeInterpreter Interpreter {
|
---|
| 50 | get; private set;
|
---|
| 51 | }
|
---|
| 52 |
|
---|
| 53 | public override IEnumerable<string> VariablesUsedForPrediction {
|
---|
| 54 | get { return allowedInputVariables; }
|
---|
| 55 | }
|
---|
| 56 |
|
---|
| 57 | [Storable]
|
---|
| 58 | private string[] allowedInputVariables;
|
---|
| 59 |
|
---|
| 60 | [StorableConstructor]
|
---|
| 61 | private BayesianNonlinearRegressionModel(bool deserializing)
|
---|
| 62 | : base(deserializing) {
|
---|
| 63 | }
|
---|
| 64 | private BayesianNonlinearRegressionModel(BayesianNonlinearRegressionModel original, Cloner cloner)
|
---|
| 65 | : base(original, cloner) {
|
---|
| 66 | allowedInputVariables = (string[])original.allowedInputVariables.Clone();
|
---|
| 67 | this.Tree = cloner.Clone(original.Tree);
|
---|
| 68 | this.parameterEmpiricalDistribution = original.parameterEmpiricalDistribution;
|
---|
| 69 | this.Interpreter = cloner.Clone(original.Interpreter);
|
---|
| 70 | }
|
---|
| 71 | public BayesianNonlinearRegressionModel(
|
---|
| 72 | ISymbolicExpressionTree tree,
|
---|
| 73 | double[][] parameterEmpiricalDistribution,
|
---|
| 74 | ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
|
---|
| 75 | string targetVariable,
|
---|
| 76 | IEnumerable<string> allowedInputVariables
|
---|
| 77 | )
|
---|
| 78 | : base(targetVariable) {
|
---|
| 79 | this.name = ItemName;
|
---|
| 80 | this.description = ItemDescription;
|
---|
| 81 | this.Tree = tree;
|
---|
| 82 | this.parameterEmpiricalDistribution = (double[][])parameterEmpiricalDistribution.Clone();
|
---|
| 83 | this.Interpreter = interpreter;
|
---|
| 84 | this.allowedInputVariables = allowedInputVariables.ToArray();
|
---|
| 85 | }
|
---|
| 86 |
|
---|
| 87 | [StorableHook(HookType.AfterDeserialization)]
|
---|
| 88 | private void AfterDeserialization() {
|
---|
| 89 | }
|
---|
| 90 |
|
---|
| 91 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
| 92 | return new BayesianNonlinearRegressionModel(this, cloner);
|
---|
| 93 | }
|
---|
| 94 |
|
---|
| 95 |
|
---|
| 96 | public override IEnumerable<double> GetEstimatedValues(IDataset dataset, IEnumerable<int> rows) {
|
---|
| 97 | var y = Sample(dataset, rows);
|
---|
| 98 | return y.Select(yi => yi.Average());
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | public IEnumerable<double> GetEstimatedVariances(IDataset dataset, IEnumerable<int> rows) {
|
---|
| 102 | var y = Sample(dataset, rows);
|
---|
| 103 | return y.Select(yi => yi.VariancePop());
|
---|
| 104 | }
|
---|
| 105 |
|
---|
| 106 | private IList<double>[] Sample(IDataset dataset, IEnumerable<int> rows) {
|
---|
| 107 | List<double>[] y = rows.Select(_ => new List<double>(SAMPLE_SIZE)).ToArray();
|
---|
| 108 | var rand = new System.Random(1234); // TODO;
|
---|
| 109 | for (int s = 0; s < SAMPLE_SIZE; s++) {
|
---|
| 110 | var paramIdx = rand.Next(parameterEmpiricalDistribution.Length);
|
---|
| 111 | UpdateConstants(Tree, parameterEmpiricalDistribution[paramIdx]);
|
---|
| 112 | int predRow = 0;
|
---|
| 113 | foreach (var pred in Interpreter.GetSymbolicExpressionTreeValues(Tree, dataset, rows)) {
|
---|
| 114 | y[predRow].Add(pred);
|
---|
| 115 | predRow++;
|
---|
| 116 | }
|
---|
| 117 | }
|
---|
| 118 | return y;
|
---|
| 119 | }
|
---|
| 120 |
|
---|
| 121 | public override IRegressionSolution CreateRegressionSolution(IRegressionProblemData problemData) {
|
---|
| 122 | return new ConfidenceRegressionSolution(this, new RegressionProblemData(problemData));
|
---|
| 123 | }
|
---|
| 124 |
|
---|
| 125 |
|
---|
| 126 | #region taken from ConstantOptEval TODO
|
---|
| 127 | private static void UpdateConstants(ISymbolicExpressionTree tree, double[] constants) {
|
---|
| 128 | int i = 0;
|
---|
| 129 | foreach (var node in tree.Root.IterateNodesPrefix().OfType<SymbolicExpressionTreeTerminalNode>()) {
|
---|
| 130 | ConstantTreeNode constantTreeNode = node as ConstantTreeNode;
|
---|
| 131 | VariableTreeNodeBase variableTreeNodeBase = node as VariableTreeNodeBase;
|
---|
| 132 | FactorVariableTreeNode factorVarTreeNode = node as FactorVariableTreeNode;
|
---|
| 133 | if (constantTreeNode != null)
|
---|
| 134 | constantTreeNode.Value = constants[i++];
|
---|
| 135 | // else if (updateVariableWeights && variableTreeNodeBase != null)
|
---|
| 136 | // variableTreeNodeBase.Weight = constants[i++];
|
---|
| 137 | else if (factorVarTreeNode != null) {
|
---|
| 138 | for (int j = 0; j < factorVarTreeNode.Weights.Length; j++)
|
---|
| 139 | factorVarTreeNode.Weights[j] = constants[i++];
|
---|
| 140 | }
|
---|
| 141 | }
|
---|
| 142 | }
|
---|
| 143 | #endregion
|
---|
| 144 |
|
---|
| 145 | }
|
---|
| 146 | }
|
---|