Skip to content
GitLab
Select Git revision
  • master default
Find file BlameHistoryPermalink
otbTrainGradientBoostedTree.hxx 5.26 KiB
1 
/*
2 
 * Copyright (C) 2005-2017 Centre National d'Etudes Spatiales (CNES)
3 
 *
4 
 * This file is part of Orfeo Toolbox
5 
 *
6 
 *     https://www.orfeo-toolbox.org/
7 
 *
8 
 * Licensed under the Apache License, Version 2.0 (the "License");
9 
 * you may not use this file except in compliance with the License.
10 
 * You may obtain a copy of the License at
11 
 *
12 
 *     http://www.apache.org/licenses/LICENSE-2.0
13 
 *
14 
 * Unless required by applicable law or agreed to in writing, software
15 
 * distributed under the License is distributed on an "AS IS" BASIS,
16 
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
17 
 * See the License for the specific language governing permissions and
18 
 * limitations under the License.
19 
 */
20
21 
#ifndef otbTrainGradientBoostedTree_txx
22 
#define otbTrainGradientBoostedTree_txx
23 
#include "otbLearningApplicationBase.h"
24 
#include "otbGradientBoostedTreeMachineLearningModel.h"
25
26 
namespace otb
27 
{
28 
namespace Wrapper
29 
{
30
31 
template <class TInputValue, class TOutputValue>
32 
void
33 
LearningApplicationBase<TInputValue,TOutputValue>
34 
::InitGradientBoostedTreeParams()
35 
{
36 
// disable GBTree model with OpenCV 3 (not implemented)
37 
#ifndef OTB_OPENCV_3
38 
  AddChoice("classifier.gbt", "Gradient Boosted Tree classifier");
39 
  SetParameterDescription(
40 
      "classifier.gbt",
41 
      "This group of parameters allows setting Gradient Boosted Tree classifier parameters. "
42 
      "See complete documentation here \\url{http://docs.opencv.org/modules/ml/doc/gradient_boosted_trees.html}.");
43
44 
  if (m_RegressionFlag)
45 
    {
46 
    AddParameter(ParameterType_Choice, "classifier.gbt.t", "Loss Function Type");
47 
    SetParameterDescription("classifier.gbt.t","Type of loss functionused for training.");
48 
    AddChoice("classifier.gbt.t.sqr","Squared Loss");
49 
    AddChoice("classifier.gbt.t.abs","Absolute Loss");
50 
    AddChoice("classifier.gbt.t.hub","Huber Loss");
51 
    }
52
53 
  //WeakCount
54 
  AddParameter(ParameterType_Int, "classifier.gbt.w", "Number of boosting algorithm iterations");
55 
  SetParameterInt("classifier.gbt.w",200);
56 
  SetParameterDescription(
57 
      "classifier.gbt.w",
58 
      "Number \"w\" of boosting algorithm iterations, with w*K being the total number of trees in "
59 
      "the GBT model, where K is the output number of classes.");
60
61 
  //Shrinkage
62 
  AddParameter(ParameterType_Float, "classifier.gbt.s", "Regularization parameter");
63 
  SetParameterFloat("classifier.gbt.s",0.01);
64 
  SetParameterDescription("classifier.gbt.s", "Regularization parameter.");
65
66 
  //SubSamplePortion
67 
  AddParameter(ParameterType_Float, "classifier.gbt.p",
68 
               "Portion of the whole training set used for each algorithm iteration");
69 
  SetParameterFloat("classifier.gbt.p",0.8);
70 
  SetParameterDescription(
7172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140
"classifier.gbt.p", "Portion of the whole training set used for each algorithm iteration. The subset is generated randomly."); //MaxDepth AddParameter(ParameterType_Int, "classifier.gbt.max", "Maximum depth of the tree"); SetParameterInt("classifier.gbt.max",3); SetParameterDescription( "classifier.gbt.max", "The training algorithm attempts to split each node while its depth is smaller than the maximum " "possible depth of the tree. The actual depth may be smaller if the other termination criteria are met, and/or " "if the tree is pruned."); //UseSurrogates : don't need to be exposed ! //AddParameter(ParameterType_Empty, "classifier.gbt.sur", "Surrogate splits will be built"); //SetParameterDescription("classifier.gbt.sur","These splits allow working with missing data and compute variable importance correctly."); #endif } template <class TInputValue, class TOutputValue> void LearningApplicationBase<TInputValue,TOutputValue> ::TrainGradientBoostedTree(typename ListSampleType::Pointer trainingListSample, typename TargetListSampleType::Pointer trainingLabeledListSample, std::string modelPath) { #ifdef OTB_OPENCV_3 (void) trainingListSample; (void) trainingLabeledListSample; (void) modelPath; #else typedef otb::GradientBoostedTreeMachineLearningModel<InputValueType, OutputValueType> GradientBoostedTreeType; typename GradientBoostedTreeType::Pointer classifier = GradientBoostedTreeType::New(); classifier->SetRegressionMode(this->m_RegressionFlag); classifier->SetInputListSample(trainingListSample); classifier->SetTargetListSample(trainingLabeledListSample); classifier->SetWeakCount(GetParameterInt("classifier.gbt.w")); classifier->SetShrinkage(GetParameterFloat("classifier.gbt.s")); classifier->SetSubSamplePortion(GetParameterFloat("classifier.gbt.p")); classifier->SetMaxDepth(GetParameterInt("classifier.gbt.max")); if (m_RegressionFlag) { switch (GetParameterInt("classifier.gbt.t")) { case 0: // SQUARED_LOSS classifier->SetLossFunctionType(CvGBTrees::SQUARED_LOSS); break; case 1: // ABSOLUTE_LOSS classifier->SetLossFunctionType(CvGBTrees::ABSOLUTE_LOSS); break; case 2: // HUBER_LOSS classifier->SetLossFunctionType(CvGBTrees::HUBER_LOSS); break; default: classifier->SetLossFunctionType(CvGBTrees::SQUARED_LOSS); break; } } else { classifier->SetLossFunctionType(CvGBTrees::DEVIANCE_LOSS); } classifier->Train(); classifier->Save(modelPath); #endif } } //end namespace wrapper } //end namespace otb
141142
#endif