Multiclass GentleAdaboosting

A fast Gentle Adaboost classifier with two kind of weaklearners

Description

A fast Gentle Adaboost classifier with two different weak-learners: i) decision stump and ii) perceptron. Multiclass is performed with the one-against-all strategy.

Usage
  ------

  model = gentleboost_model(X , y , [T] , [options]);

  
  Inputs
  -------

  X Features matrix (d x N)
  y Labels (1 x N). If y represent binary labels vector then y_i={-1,1}, i=1,...,N
  T Number of weak learners (default T = 100)
  options
              weaklearner Choice of weaklearner 0 : Decision Stump, 1 : Perceptron (default weaklearner = 0)

epsi Epsilon constant in the sigmoid function used in the perceptron (default epsi = 1)
              lambda Regularization parameter for the perceptron's weights update (default lambda = 1e-3)
max_ite Maximum number of iterations of the perceptron algorithm (default max_ite = 100)

  Outputs
  -------
  
  model Structure of model ouput
  
featureIdx Features index of the T best weaklearners (T x m) where m is the number of class.
For binary classification m is force to 1.
th Theta (T x m)
b Bias parameter (T x m)
a Decision Stump extra parameter (T x m)

Please run mexme_gentleboost to compile mex-files on your plateform.

Please run test_gentleboost_model to run the demo.

N.B. Last build of libsvm is also included and slightly modified to supress verbose.

• MATLAB release: MATLAB 7.5 (R2007b)
• Other requirements: A C/C++ compiler
• Published M Files: test_gentleboost_model

Other Files

auroc.m,
basicroc.m,
gentleboost_model.c,
gentleboost_predict.c,
html/test_gentleboost_model.png,
html/test_gentleboost_model_01.png,
html/test_gentleboost_model_02.png,
html/test_gentleboost_model_03.png,
html/test_gentleboost_model_04.png,
ini_proto.m,
iris.mat,
license.txt,
mexme_gentleboost.m,
NN_predict.c,
sampling.m,
samplingset.m,
srng_model.c,
svm.cpp,
svm.h,
svm_model_matlab.c,
svm_model_matlab.h,
svmpredict.c,
svmtrain.c,
test_gentleboost_model.m,
wine.mat