[12332] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
| 3 | * Copyright (C) 2002-2015 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
| 4 | * and the BEACON Center for the Study of Evolution in Action.
|
---|
| 5 | *
|
---|
| 6 | * This file is part of HeuristicLab.
|
---|
| 7 | *
|
---|
| 8 | * HeuristicLab is free software: you can redistribute it and/or modify
|
---|
| 9 | * it under the terms of the GNU General Public License as published by
|
---|
| 10 | * the Free Software Foundation, either version 3 of the License, or
|
---|
| 11 | * (at your option) any later version.
|
---|
| 12 | *
|
---|
| 13 | * HeuristicLab is distributed in the hope that it will be useful,
|
---|
| 14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 16 | * GNU General Public License for more details.
|
---|
| 17 | *
|
---|
| 18 | * You should have received a copy of the GNU General Public License
|
---|
| 19 | * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
|
---|
| 20 | */
|
---|
| 21 | #endregion
|
---|
| 22 |
|
---|
| 23 | using System;
|
---|
| 24 | using System.Collections.Generic;
|
---|
| 25 | using System.Diagnostics.Contracts;
|
---|
| 26 | using System.Linq;
|
---|
| 27 | using HeuristicLab.Problems.DataAnalysis;
|
---|
| 28 | using HeuristicLab.Random;
|
---|
| 29 |
|
---|
| 30 | namespace HeuristicLab.Algorithms.DataAnalysis {
|
---|
| 31 | public static class GradientBoostedTreesAlgorithmStatic {
|
---|
| 32 | #region static API
|
---|
| 33 |
|
---|
| 34 | public interface IGbmState {
|
---|
| 35 | IRegressionModel GetModel();
|
---|
| 36 | double GetTrainLoss();
|
---|
| 37 | double GetTestLoss();
|
---|
| 38 | IEnumerable<KeyValuePair<string, double>> GetVariableRelevance();
|
---|
| 39 | }
|
---|
| 40 |
|
---|
| 41 | // created through factory method
|
---|
[12590] | 42 | // GbmState details are private API users can only use methods from IGbmState
|
---|
[12332] | 43 | private class GbmState : IGbmState {
|
---|
[12698] | 44 | internal IRegressionProblemData problemData { get; private set; }
|
---|
| 45 | internal ILossFunction lossFunction { get; private set; }
|
---|
| 46 | internal int maxSize { get; private set; }
|
---|
| 47 | internal double nu { get; private set; }
|
---|
| 48 | internal double r { get; private set; }
|
---|
| 49 | internal double m { get; private set; }
|
---|
[12710] | 50 | internal int[] trainingRows { get; private set; }
|
---|
| 51 | internal int[] testRows { get; private set; }
|
---|
[12698] | 52 | internal RegressionTreeBuilder treeBuilder { get; private set; }
|
---|
[12332] | 53 |
|
---|
[14029] | 54 | private readonly uint randSeed;
|
---|
[12698] | 55 | private MersenneTwister random { get; set; }
|
---|
[12332] | 56 |
|
---|
| 57 | // array members (allocate only once)
|
---|
| 58 | internal double[] pred;
|
---|
| 59 | internal double[] predTest;
|
---|
| 60 | internal double[] y;
|
---|
| 61 | internal int[] activeIdx;
|
---|
[12597] | 62 | internal double[] pseudoRes;
|
---|
[12332] | 63 |
|
---|
[12698] | 64 | private readonly IList<IRegressionModel> models;
|
---|
| 65 | private readonly IList<double> weights;
|
---|
[12332] | 66 |
|
---|
[12632] | 67 | public GbmState(IRegressionProblemData problemData, ILossFunction lossFunction, uint randSeed, int maxSize, double r, double m, double nu) {
|
---|
| 68 | // default settings for MaxSize, Nu and R
|
---|
| 69 | this.maxSize = maxSize;
|
---|
[12332] | 70 | this.nu = nu;
|
---|
| 71 | this.r = r;
|
---|
| 72 | this.m = m;
|
---|
| 73 |
|
---|
[14029] | 74 | this.randSeed = randSeed;
|
---|
[12332] | 75 | random = new MersenneTwister(randSeed);
|
---|
| 76 | this.problemData = problemData;
|
---|
[12710] | 77 | this.trainingRows = problemData.TrainingIndices.ToArray();
|
---|
| 78 | this.testRows = problemData.TestIndices.ToArray();
|
---|
[12332] | 79 | this.lossFunction = lossFunction;
|
---|
| 80 |
|
---|
[12710] | 81 | int nRows = trainingRows.Length;
|
---|
[12332] | 82 |
|
---|
[12710] | 83 | y = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, trainingRows).ToArray();
|
---|
[12332] | 84 |
|
---|
| 85 | treeBuilder = new RegressionTreeBuilder(problemData, random);
|
---|
| 86 |
|
---|
[12371] | 87 | activeIdx = Enumerable.Range(0, nRows).ToArray();
|
---|
[12332] | 88 |
|
---|
[12697] | 89 | var zeros = Enumerable.Repeat(0.0, nRows).ToArray();
|
---|
| 90 | double f0 = lossFunction.LineSearch(y, zeros, activeIdx, 0, nRows - 1); // initial constant value (mean for squared errors)
|
---|
[12332] | 91 | pred = Enumerable.Repeat(f0, nRows).ToArray();
|
---|
[12710] | 92 | predTest = Enumerable.Repeat(f0, testRows.Length).ToArray();
|
---|
[12597] | 93 | pseudoRes = new double[nRows];
|
---|
[12332] | 94 |
|
---|
| 95 | models = new List<IRegressionModel>();
|
---|
| 96 | weights = new List<double>();
|
---|
| 97 | // add constant model
|
---|
[14029] | 98 | models.Add(new ConstantModel(f0, problemData.TargetVariable));
|
---|
[12332] | 99 | weights.Add(1.0);
|
---|
| 100 | }
|
---|
| 101 |
|
---|
| 102 | public IRegressionModel GetModel() {
|
---|
[14029] | 103 | #pragma warning disable 618
|
---|
| 104 | var model = new GradientBoostedTreesModel(models, weights);
|
---|
| 105 | #pragma warning restore 618
|
---|
| 106 | // we don't know the number of iterations here but the number of weights is equal
|
---|
| 107 | // to the number of iterations + 1 (for the constant model)
|
---|
| 108 | // wrap the actual model in a surrogate that enables persistence and lazy recalculation of the model if necessary
|
---|
| 109 | return new GradientBoostedTreesModelSurrogate(problemData, randSeed, lossFunction, weights.Count - 1, maxSize, r, m, nu, model);
|
---|
[12332] | 110 | }
|
---|
| 111 | public IEnumerable<KeyValuePair<string, double>> GetVariableRelevance() {
|
---|
| 112 | return treeBuilder.GetVariableRelevance();
|
---|
| 113 | }
|
---|
| 114 |
|
---|
| 115 | public double GetTrainLoss() {
|
---|
| 116 | int nRows = y.Length;
|
---|
[12696] | 117 | return lossFunction.GetLoss(y, pred) / nRows;
|
---|
[12332] | 118 | }
|
---|
| 119 | public double GetTestLoss() {
|
---|
[12710] | 120 | var yTest = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, testRows);
|
---|
| 121 | var nRows = testRows.Length;
|
---|
[12696] | 122 | return lossFunction.GetLoss(yTest, predTest) / nRows;
|
---|
[12332] | 123 | }
|
---|
[12698] | 124 |
|
---|
| 125 | internal void AddModel(IRegressionModel m, double weight) {
|
---|
| 126 | models.Add(m);
|
---|
| 127 | weights.Add(weight);
|
---|
| 128 | }
|
---|
[12332] | 129 | }
|
---|
| 130 |
|
---|
| 131 | // simple interface
|
---|
[14029] | 132 | public static GradientBoostedTreesSolution TrainGbm(IRegressionProblemData problemData, ILossFunction lossFunction, int maxSize, double nu, double r, double m, int maxIterations, uint randSeed = 31415) {
|
---|
[12332] | 133 | Contract.Assert(r > 0);
|
---|
| 134 | Contract.Assert(r <= 1.0);
|
---|
| 135 | Contract.Assert(nu > 0);
|
---|
| 136 | Contract.Assert(nu <= 1.0);
|
---|
| 137 |
|
---|
[12698] | 138 | var state = (GbmState)CreateGbmState(problemData, lossFunction, randSeed, maxSize, r, m, nu);
|
---|
[12332] | 139 |
|
---|
| 140 | for (int iter = 0; iter < maxIterations; iter++) {
|
---|
| 141 | MakeStep(state);
|
---|
| 142 | }
|
---|
| 143 |
|
---|
[12698] | 144 | var model = state.GetModel();
|
---|
[14029] | 145 | return new GradientBoostedTreesSolution(model, (IRegressionProblemData)problemData.Clone());
|
---|
[12332] | 146 | }
|
---|
| 147 |
|
---|
| 148 | // for custom stepping & termination
|
---|
[12632] | 149 | public static IGbmState CreateGbmState(IRegressionProblemData problemData, ILossFunction lossFunction, uint randSeed, int maxSize = 3, double r = 0.66, double m = 0.5, double nu = 0.01) {
|
---|
| 150 | return new GbmState(problemData, lossFunction, randSeed, maxSize, r, m, nu);
|
---|
[12332] | 151 | }
|
---|
| 152 |
|
---|
[12698] | 153 | // use default settings for maxSize, nu, r from state
|
---|
[12332] | 154 | public static void MakeStep(IGbmState state) {
|
---|
| 155 | var gbmState = state as GbmState;
|
---|
| 156 | if (gbmState == null) throw new ArgumentException("state");
|
---|
| 157 |
|
---|
[12632] | 158 | MakeStep(gbmState, gbmState.maxSize, gbmState.nu, gbmState.r, gbmState.m);
|
---|
[12332] | 159 | }
|
---|
| 160 |
|
---|
[12698] | 161 | // allow dynamic adaptation of maxSize, nu and r (even though this is not used)
|
---|
[12632] | 162 | public static void MakeStep(IGbmState state, int maxSize, double nu, double r, double m) {
|
---|
[12332] | 163 | var gbmState = state as GbmState;
|
---|
| 164 | if (gbmState == null) throw new ArgumentException("state");
|
---|
| 165 |
|
---|
| 166 | var problemData = gbmState.problemData;
|
---|
| 167 | var lossFunction = gbmState.lossFunction;
|
---|
| 168 | var yPred = gbmState.pred;
|
---|
| 169 | var yPredTest = gbmState.predTest;
|
---|
| 170 | var treeBuilder = gbmState.treeBuilder;
|
---|
| 171 | var y = gbmState.y;
|
---|
| 172 | var activeIdx = gbmState.activeIdx;
|
---|
[12597] | 173 | var pseudoRes = gbmState.pseudoRes;
|
---|
[12710] | 174 | var trainingRows = gbmState.trainingRows;
|
---|
| 175 | var testRows = gbmState.testRows;
|
---|
[12332] | 176 |
|
---|
[12698] | 177 | // copy output of gradient function to pre-allocated rim array (pseudo-residual per row and model)
|
---|
[12332] | 178 | int rimIdx = 0;
|
---|
[12696] | 179 | foreach (var g in lossFunction.GetLossGradient(y, yPred)) {
|
---|
[12597] | 180 | pseudoRes[rimIdx++] = g;
|
---|
[12332] | 181 | }
|
---|
| 182 |
|
---|
[12697] | 183 | var tree = treeBuilder.CreateRegressionTreeForGradientBoosting(pseudoRes, yPred, maxSize, activeIdx, lossFunction, r, m);
|
---|
[12332] | 184 |
|
---|
| 185 | int i = 0;
|
---|
[12710] | 186 | foreach (var pred in tree.GetEstimatedValues(problemData.Dataset, trainingRows)) {
|
---|
[12332] | 187 | yPred[i] = yPred[i] + nu * pred;
|
---|
| 188 | i++;
|
---|
| 189 | }
|
---|
| 190 | // update predictions for validation set
|
---|
| 191 | i = 0;
|
---|
[12710] | 192 | foreach (var pred in tree.GetEstimatedValues(problemData.Dataset, testRows)) {
|
---|
[12332] | 193 | yPredTest[i] = yPredTest[i] + nu * pred;
|
---|
| 194 | i++;
|
---|
| 195 | }
|
---|
| 196 |
|
---|
[12698] | 197 | gbmState.AddModel(tree, nu);
|
---|
[12332] | 198 | }
|
---|
| 199 | #endregion
|
---|
| 200 | }
|
---|
| 201 | }
|
---|