An implementation of Large Margin nearest neighbor metric learning technique. More...

#include <lmnn.hpp>

Public Member Functions
	LMNN (const arma::mat &dataset, const arma::Row< size_t > &labels, const size_t k, const MetricType metric=MetricType())
	Initialize the LMNN object, passing a dataset (distance metric is learned using this dataset) and labels. More...

template<typename... CallbackTypes>
void	LearnDistance (arma::mat &outputMatrix, CallbackTypes &&... callbacks)
	Perform Large Margin Nearest Neighbors metric learning. More...

const arma::mat &	Dataset () const
	Get the dataset reference.

const arma::Row< size_t > &	Labels () const
	Get the labels reference.

const double &	Regularization () const
	Access the regularization value.

double &	Regularization ()
	Modify the regularization value.

const size_t &	Range () const
	Access the range value.

size_t &	Range ()
	Modify the range value.

const size_t &	K () const
	Access the value of k.

size_t	K ()
	Modify the value of k.

const OptimizerType &	Optimizer () const
	Get the optimizer.

OptimizerType &	Optimizer ()

Detailed Description

template<typename MetricType = metric::SquaredEuclideanDistance, typename OptimizerType = ens::AMSGrad>
class mlpack::lmnn::LMNN< MetricType, OptimizerType >

An implementation of Large Margin nearest neighbor metric learning technique.

The method seeks to improve clustering & classification algorithms on a dataset by transforming the dataset representation in a more convenient form for them. It introduces the concept of target neighbors and impostors, focusing on the idea that the distance between impostors and the perimeters established by target neighbors should be large and that between target neighbors and data point should be small. It requires the knowledge of target neighbors beforehand. Moreover, target neighbors once initialized remain same.

For more details, see the following published paper:

@ARTICLE{weinberger09distance,
  author = {Weinberger, K.Q. and Saul, L.K.},
  title = {{Distance metric learning for large margin nearest neighbor
      classification}},
  journal = {The Journal of Machine Learning Research},
  year = {2009},
  volume = {10},
  pages = {207--244},
  publisher = {MIT Press}
}

Template Parameters

MetricType	The type of metric to use for computation.
OptimizerType	Optimizer to use for developing distance.

Constructor & Destructor Documentation

◆ LMNN()

template<typename MetricType , typename OptimizerType >

mlpack::lmnn::LMNN< MetricType, OptimizerType >::LMNN	(	const arma::mat &	dataset,
		const arma::Row< size_t > &	labels,
		const size_t	k,
		const MetricType	metric = `MetricType()`
	)

Initialize the LMNN object, passing a dataset (distance metric is learned using this dataset) and labels.

Takes in a reference to the dataset.

Initialization will copy both dataset and labels matrices to internal copies.

Parameters

dataset	Input dataset.
labels	Input dataset labels.
k	Number of targets to consider.
metric	Type of metric used for computation.

Copies the data, initializes all of the member variables and constraint object and generate constraints.

Member Function Documentation

◆ LearnDistance()

template<typename MetricType , typename OptimizerType >

template<typename... CallbackTypes>

void mlpack::lmnn::LMNN< MetricType, OptimizerType >::LearnDistance	(	arma::mat &	outputMatrix,
		CallbackTypes &&...	callbacks
	)

Perform Large Margin Nearest Neighbors metric learning.

The output distance matrix is written into the passed reference. If the LearnDistance() is called with an outputMatrix with correct dimensions, then that matrix will be used as the starting point for optimization.

Template Parameters

CallbackTypes Types of Callback functions.