This class implements K-Means clustering, using a variety of possible implementations of Lloyd's algorithm. More...

#include <kmeans.hpp>

Public Member Functions
	KMeans (const size_t maxIterations=1000, const MetricType metric=MetricType(), const InitialPartitionPolicy partitioner=InitialPartitionPolicy(), const EmptyClusterPolicy emptyClusterAction=EmptyClusterPolicy())
	Create a K-Means object and (optionally) set the parameters which K-Means will be run with. More...

void	Cluster (const MatType &data, const size_t clusters, arma::Row< size_t > &assignments, const bool initialGuess=false)
	Perform k-means clustering on the data, returning a list of cluster assignments. More...

void	Cluster (const MatType &data, size_t clusters, arma::mat &centroids, const bool initialGuess=false)
	Perform k-means clustering on the data, returning the centroids of each cluster in the centroids matrix. More...

void	Cluster (const MatType &data, const size_t clusters, arma::Row< size_t > &assignments, arma::mat &centroids, const bool initialAssignmentGuess=false, const bool initialCentroidGuess=false)
	Perform k-means clustering on the data, returning a list of cluster assignments and also the centroids of each cluster. More...

size_t	MaxIterations () const
	Get the maximum number of iterations.

size_t &	MaxIterations ()
	Set the maximum number of iterations.

const MetricType &	Metric () const
	Get the distance metric.

MetricType &	Metric ()
	Modify the distance metric.

const InitialPartitionPolicy &	Partitioner () const
	Get the initial partitioning policy.

InitialPartitionPolicy &	Partitioner ()
	Modify the initial partitioning policy.

const EmptyClusterPolicy &	EmptyClusterAction () const
	Get the empty cluster policy.

EmptyClusterPolicy &	EmptyClusterAction ()
	Modify the empty cluster policy.

template<typename Archive >
void	serialize (Archive &ar, const uint32_t version)
	Serialize the k-means object.

Detailed Description

template<typename MetricType = metric::EuclideanDistance, typename InitialPartitionPolicy = SampleInitialization, typename EmptyClusterPolicy = MaxVarianceNewCluster, template< class, class > class LloydStepType = NaiveKMeans, typename MatType = arma::mat>
class mlpack::kmeans::KMeans< MetricType, InitialPartitionPolicy, EmptyClusterPolicy, LloydStepType, MatType >

This class implements K-Means clustering, using a variety of possible implementations of Lloyd's algorithm.

Four template parameters can (optionally) be supplied: the distance metric to use, the policy for how to find the initial partition of the data, the actions to be taken when an empty cluster is encountered, and the implementation of a single Lloyd step to use.

A simple example of how to run K-Means clustering is shown below.

extern arma::mat data; // Dataset we want to run K-Means on.
arma::Row<size_t> assignments; // Cluster assignments.
arma::mat centroids; // Cluster centroids.
KMeans<> k; // Default options.
k.Cluster(data, 3, assignments, centroids); // 3 clusters.
// Cluster using the Manhattan distance, 100 iterations maximum, saving only
// the centroids.
KMeans<metric::ManhattanDistance> k(100);
k.Cluster(data, 6, centroids); // 6 clusters.

Template Parameters

MetricType	The distance metric to use for this KMeans; see metric::LMetric for an example.
InitialPartitionPolicy	Initial partitioning policy; must implement a default constructor and either 'void Cluster(const arma::mat&, const size_t, arma::Row<size_t>&)' or 'void Cluster(const arma::mat&, const size_t, arma::mat&)'.
EmptyClusterPolicy	Policy for what to do on an empty cluster; must implement a default constructor and 'void EmptyCluster(const arma::mat& data, const size_t emptyCluster, const arma::mat& oldCentroids, arma::mat& newCentroids, arma::Col<size_t>& counts, MetricType& metric, const size_t iteration)'.
LloydStepType	Implementation of single Lloyd step to use.

See also: RandomPartition, SampleInitialization, RefinedStart, AllowEmptyClusters, MaxVarianceNewCluster, NaiveKMeans, ElkanKMeans

Constructor & Destructor Documentation

◆ KMeans()

template<typename MetricType , typename InitialPartitionPolicy , typename EmptyClusterPolicy , template< class, class > class LloydStepType, typename MatType >

mlpack::kmeans::KMeans< MetricType, InitialPartitionPolicy, EmptyClusterPolicy, LloydStepType, MatType >::KMeans	(	const size_t	maxIterations = `1000`,
		const MetricType	metric = `MetricType()`,
		const InitialPartitionPolicy	partitioner = `InitialPartitionPolicy()`,
		const EmptyClusterPolicy	emptyClusterAction = `EmptyClusterPolicy()`
	)

Create a K-Means object and (optionally) set the parameters which K-Means will be run with.

Construct the K-Means object.

Parameters

maxIterations	Maximum number of iterations allowed before giving up (0 is valid, but the algorithm may never terminate).
metric	Optional MetricType object; for when the metric has state it needs to store.
partitioner	Optional InitialPartitionPolicy object; for when a specially initialized partitioning policy is required.
emptyClusterAction	Optional EmptyClusterPolicy object; for when a specially initialized empty cluster policy is required.

Member Function Documentation

◆ Cluster() [1/3]

template<typename MetricType , typename InitialPartitionPolicy , typename EmptyClusterPolicy , template< class, class > class LloydStepType, typename MatType >

void mlpack::kmeans::KMeans< MetricType, InitialPartitionPolicy, EmptyClusterPolicy, LloydStepType, MatType >::Cluster	(	const MatType &	data,
		const size_t	clusters,
		arma::Row< size_t > &	assignments,
		const bool	initialGuess = `false`
	)

inline

Perform k-means clustering on the data, returning a list of cluster assignments.

Optionally, the vector of assignments can be set to an initial guess of the cluster assignments; to do this, set initialGuess to true.

Template Parameters

MatType Type of matrix (arma::mat or arma::sp_mat).

Parameters

data	Dataset to cluster.
clusters	Number of clusters to compute.
assignments	Vector to store cluster assignments in.
initialGuess	If true, then it is assumed that assignments has a list of initial cluster assignments.

This just forward to the other function, which returns the centroids too. If this is properly inlined, there shouldn't be any performance penalty whatsoever.

◆ Cluster() [2/3]

template<typename MetricType , typename InitialPartitionPolicy , typename EmptyClusterPolicy , template< class, class > class LloydStepType, typename MatType >

void mlpack::kmeans::KMeans< MetricType, InitialPartitionPolicy, EmptyClusterPolicy, LloydStepType, MatType >::Cluster	(	const MatType &	data,
		size_t	clusters,
		arma::mat &	centroids,
		const bool	initialGuess = `false`
	)

Perform k-means clustering on the data, returning the centroids of each cluster in the centroids matrix.

Perform k-means clustering on the data, returning a list of cluster assignments and the centroids of each cluster.

Optionally, the initial centroids can be specified by filling the centroids matrix with the initial centroids and specifying initialGuess = true.

Template Parameters

MatType Type of matrix (arma::mat or arma::sp_mat).

Parameters

data	Dataset to cluster.
clusters	Number of clusters to compute.
centroids	Matrix in which centroids are stored.
initialGuess	If true, then it is assumed that centroids contains the initial cluster centroids.

◆ Cluster() [3/3]

template<typename MetricType , typename InitialPartitionPolicy , typename EmptyClusterPolicy , template< class, class > class LloydStepType, typename MatType >

void mlpack::kmeans::KMeans< MetricType, InitialPartitionPolicy, EmptyClusterPolicy, LloydStepType, MatType >::Cluster	(	const MatType &	data,
		const size_t	clusters,
		arma::Row< size_t > &	assignments,
		arma::mat &	centroids,
		const bool	initialAssignmentGuess = `false`,
		const bool	initialCentroidGuess = `false`
	)

Perform k-means clustering on the data, returning a list of cluster assignments and also the centroids of each cluster.

Perform k-means clustering on the data, returning a list of cluster assignments and the centroids of each cluster.

Optionally, the vector of assignments can be set to an initial guess of the cluster assignments; to do this, set initialAssignmentGuess to true. Another way to set initial cluster guesses is to fill the centroids matrix with the centroid guesses, and then set initialCentroidGuess to true. initialAssignmentGuess supersedes initialCentroidGuess, so if both are set to true, the assignments vector is used.

Template Parameters

MatType Type of matrix (arma::mat or arma::sp_mat).

Parameters

data	Dataset to cluster.
clusters	Number of clusters to compute.
assignments	Vector to store cluster assignments in.
centroids	Matrix in which centroids are stored.
initialAssignmentGuess	If true, then it is assumed that assignments has a list of initial cluster assignments.
initialCentroidGuess	If true, then it is assumed that centroids contains the initial centroids of each cluster.

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

src/mlpack/methods/kmeans/kmeans.hpp
src/mlpack/methods/kmeans/kmeans_impl.hpp

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ KMeans()

Member Function Documentation

◆ Cluster() [1/3]

◆ Cluster() [2/3]

◆ Cluster() [3/3]