mlpack::adaboost::AdaBoost< WeakLearnerType, MatType > Class Template Reference

The AdaBoost class. More...

#include <adaboost.hpp>

Public Member Functions
	AdaBoost (const MatType &data, const arma::Row< size_t > &labels, const size_t numClasses, const WeakLearnerType &other, const size_t iterations=100, const double tolerance=1e-6)
	Constructor. More...
	AdaBoost (const double tolerance=1e-6)
	Create the AdaBoost object without training. More...
double	Tolerance () const
	Get the tolerance for stopping the optimization during training.
double &	Tolerance ()
	Modify the tolerance for stopping the optimization during training.
size_t	NumClasses () const
	Get the number of classes this model is trained on.
size_t	WeakLearners () const
	Get the number of weak learners in the model.
double	Alpha (const size_t i) const
	Get the weights for the given weak learner.
double &	Alpha (const size_t i)
	Modify the weight for the given weak learner (be careful!).
const WeakLearnerType &	WeakLearner (const size_t i) const
	Get the given weak learner.
WeakLearnerType &	WeakLearner (const size_t i)
	Modify the given weak learner (be careful!).
double	Train (const MatType &data, const arma::Row< size_t > &labels, const size_t numClasses, const WeakLearnerType &learner, const size_t iterations=100, const double tolerance=1e-6)
	Train AdaBoost on the given dataset. More...
void	Classify (const MatType &test, arma::Row< size_t > &predictedLabels, arma::mat &probabilities)
	Classify the given test points. More...
void	Classify (const MatType &test, arma::Row< size_t > &predictedLabels)
	Classify the given test points. More...
template<typename Archive >
void	serialize (Archive &ar, const uint32_t)
	Serialize the AdaBoost model.

Detailed Description

template<typename WeakLearnerType = mlpack::perceptron::Perceptron<>, typename MatType = arma::mat>
class mlpack::adaboost::AdaBoost< WeakLearnerType, MatType >

The AdaBoost class.

AdaBoost is a boosting algorithm, meaning that it combines an ensemble of weak learners to produce a strong learner. For more information on AdaBoost, see the following paper:

@article{schapire1999improved,
  author = {Schapire, Robert E. and Singer, Yoram},
  title = {Improved Boosting Algorithms Using Confidence-rated Predictions},
  journal = {Machine Learning},
  volume = {37},
  number = {3},
  month = dec,
  year = {1999},
  issn = {0885-6125},
  pages = {297--336},
}

This class is general, and can be used with any type of weak learner, so long as the learner implements the following functions:

// A boosting constructor, which learns using the training parameters of the
// given other WeakLearner, but uses the given instance weights for training.
WeakLearner(WeakLearner& other,
            const MatType& data,
            const arma::Row<size_t>& labels,
            const arma::rowvec& weights);

// Given the test points, classify them and output predictions into
// predictedLabels.
void Classify(const MatType& data, arma::Row<size_t>& predictedLabels);

For more information on and examples of weak learners, see perceptron::Perceptron<> and tree::ID3DecisionStump.

Template Parameters

MatType	Data matrix type (i.e. arma::mat or arma::sp_mat).
WeakLearnerType	Type of weak learner to use.

Constructor & Destructor Documentation

AdaBoost() [1/2]

template<typename WeakLearnerType, typename MatType>

mlpack::adaboost::AdaBoost< WeakLearnerType, MatType >::AdaBoost	(	const MatType &	data,
		const arma::Row< size_t > &	labels,
		const size_t	numClasses,
		const WeakLearnerType &	other,
		const size_t	iterations = 100,
		const double	tol = 1e-6
	)

Constructor.

This runs the AdaBoost.MH algorithm to provide a trained boosting model. This constructor takes an already-initialized weak learner; all other weak learners will learn with the same parameters as the given weak learner.

Parameters

data	Input data.
labels	Corresponding labels.
numClasses	The number of classes.
iterations	Number of boosting rounds.
tolerance	The tolerance for change in values of rt.
other	Weak learner that has already been initialized.

Currently runs the AdaBoost.MH algorithm.

Parameters

data	Input data
labels	Corresponding labels
iterations	Number of boosting rounds
tol	Tolerance for termination of Adaboost.MH.
other	Weak Learner, which has been initialized already.

AdaBoost() [2/2]

template<typename WeakLearnerType, typename MatType>

mlpack::adaboost::AdaBoost< WeakLearnerType, MatType >::AdaBoost

(

const double

tolerance = 1e-6

)

Create the AdaBoost object without training.

Be sure to call Train() before calling Classify()!

Member Function Documentation

Classify() [1/2]

template<typename WeakLearnerType , typename MatType>

void mlpack::adaboost::AdaBoost< WeakLearnerType, MatType >::Classify	(	const MatType &	test,
		arma::Row< size_t > &	predictedLabels,
		arma::mat &	probabilities
	)

Classify the given test points.

Parameters

test	Testing data.
predictedLabels	Vector in which the predicted labels of the test set will be stored.
probabilities	matrix to store the predicted class probabilities for each point in the test set.

Classify() [2/2]

template<typename WeakLearnerType , typename MatType>

void mlpack::adaboost::AdaBoost< WeakLearnerType, MatType >::Classify	(	const MatType &	test,
		arma::Row< size_t > &	predictedLabels
	)

Classify the given test points.

Parameters

test	Testing data.
predictedLabels	Vector in which the predicted labels of the test set will be stored.

Train()

template<typename WeakLearnerType, typename MatType>

double mlpack::adaboost::AdaBoost< WeakLearnerType, MatType >::Train	(	const MatType &	data,
		const arma::Row< size_t > &	labels,
		const size_t	numClasses,
		const WeakLearnerType &	learner,
		const size_t	iterations = 100,
		const double	tolerance = 1e-6
	)

Train AdaBoost on the given dataset.

This method takes an initialized WeakLearnerType; the parameters for this weak learner will be used to train each of the weak learners during AdaBoost training. Note that this will completely overwrite any model that has already been trained with this object.

Parameters

data	Dataset to train on.
labels	Labels for each point in the dataset.
numClasses	The number of classes.
learner	Learner to use for training.
iterations	Number of boosting rounds.
tolerance	The tolerance for change in values of rt.

Returns

The upper bound for training error.

---

The documentation for this class was generated from the following files:

• src/mlpack/methods/adaboost/adaboost.hpp
• src/mlpack/methods/adaboost/adaboost_impl.hpp