opennn::KMeans Class Reference

K-means clustering utility that partitions samples into the requested number of clusters. More...

#include <kmeans.h>

Public Member Functions
	KMeans (Index clusters=3, Index=100)
	Builds a K-means instance with the given cluster count and maximum number of iterations.
VectorI	calculate_outputs (const MatrixR &)
	Assigns each row of the input matrix to its nearest cluster.
VectorR	elbow_method (const MatrixR &, Index=10)
	Runs the elbow method on the supplied data over a range of cluster counts.
Index	find_optimal_clusters (const VectorR &) const
	Returns the cluster count located at the elbow of the supplied distortion curve.
VectorI	get_cluster_labels () const
	Returns the cluster label assigned to each fitted sample.
MatrixR	get_cluster_centers () const
	Returns the centroid of each cluster as rows of the returned matrix.
Index	get_clusters_number () const
	Returns the number of clusters configured for the algorithm.
void	fit (const MatrixR &)
	Fits the K-means model on the supplied data matrix.
void	set_cluster_number (const Index)
	Sets the desired number of clusters.
void	set_centers_random (const MatrixR &)
	Initializes cluster centres by sampling at random from the supplied data.

Detailed Description

K-means clustering utility that partitions samples into the requested number of clusters.

Constructor & Destructor Documentation

KMeans()

opennn::KMeans::KMeans	(	Index	clusters = 3,
		Index	= 100 )

Builds a K-means instance with the given cluster count and maximum number of iterations.

Member Function Documentation

calculate_outputs()

VectorI opennn::KMeans::calculate_outputs

(

const MatrixR &

)

Assigns each row of the input matrix to its nearest cluster.

Returns

Vector with the cluster index for every row.

elbow_method()

VectorR opennn::KMeans::elbow_method	(	const MatrixR &	,
		Index	= 10 )

Runs the elbow method on the supplied data over a range of cluster counts.

Returns

Vector with the within-cluster distortion for each tested cluster count.

find_optimal_clusters()

Index opennn::KMeans::find_optimal_clusters

(

const VectorR &

)

const

Returns the cluster count located at the elbow of the supplied distortion curve.

fit()

void opennn::KMeans::fit

(

const MatrixR &

)

Fits the K-means model on the supplied data matrix.

get_cluster_centers()

MatrixR opennn::KMeans::get_cluster_centers

(

)

const

Returns the centroid of each cluster as rows of the returned matrix.

get_cluster_labels()

VectorI opennn::KMeans::get_cluster_labels

(

)

const

Returns the cluster label assigned to each fitted sample.

get_clusters_number()

Index opennn::KMeans::get_clusters_number

(

)

const

Returns the number of clusters configured for the algorithm.

set_centers_random()

void opennn::KMeans::set_centers_random

(

const MatrixR &

)

Initializes cluster centres by sampling at random from the supplied data.

set_cluster_number()

void opennn::KMeans::set_cluster_number

(

const Index

)

Sets the desired number of clusters.