UTILSLIB::KMeans Class Reference

K-Means Clustering. More...

#include <kmeans.h>

Public Types
typedef QSharedPointer< KMeans >	SPtr
typedef QSharedPointer< const KMeans >	ConstSPtr

Public Member Functions
	KMeans (QString distance=QString("sqeuclidean"), QString start=QString("sample"), qint32 replicates=1, QString emptyact=QString("error"), bool online=true, qint32 maxit=100)
	KMeans (KMeansDistance distance, KMeansStart start=KMeansStart::Sample, qint32 replicates=1, KMeansEmptyAction emptyact=KMeansEmptyAction::Error, bool online=true, qint32 maxit=100)
bool	calculate (const Eigen::MatrixXd &X, qint32 kClusters, Eigen::VectorXi &idx, Eigen::MatrixXd &C, Eigen::VectorXd &sumD, Eigen::MatrixXd &D)

Detailed Description

K-Means Clustering.

K-Means Clustering

Definition at line 108 of file kmeans.h.

Member Typedef Documentation

ConstSPtr

typedef QSharedPointer<const KMeans> UTILSLIB::KMeans::ConstSPtr

Const shared pointer type for KMeans.

Definition at line 112 of file kmeans.h.

SPtr

typedef QSharedPointer<KMeans> UTILSLIB::KMeans::SPtr

Shared pointer type for KMeans.

Definition at line 111 of file kmeans.h.

Constructor & Destructor Documentation

KMeans() [1/2]

KMeans::KMeans ( QString distance = QString("sqeuclidean"), QString start = QString("sample"), qint32 replicates = 1, QString emptyact = QString("error"), bool online = true, qint32 maxit = 100 )

explicit

Constructs a KMeans algorithm object.

Parameters

[in]	distance	(optional) K-Means distance measure: "sqeuclidean" (default), "cityblock" , "cosine", "correlation", "hamming".
[in]	start	(optional) Cluster initialization: "sample" (default), "uniform", "cluster".
[in]	replicates	(optional) Number of K-Means replicates, which are generated. Best is returned.
[in]	emptyact	(optional) What happens if a cluster goes empty: "error" (default), "drop", "singleton".
[in]	online	(optional) If centroids should be updated during iterations: true (default), false.
[in]	maxit	(optional) Maximal number of iterations per replicate; 100 by default.

Definition at line 65 of file kmeans.cpp.

KMeans() [2/2]

KMeans::KMeans ( KMeansDistance distance, KMeansStart start = KMeansStart::Sample, qint32 replicates = 1, KMeansEmptyAction emptyact = KMeansEmptyAction::Error, bool online = true, qint32 maxit = 100 )

explicit

Constructs a KMeans algorithm object using enum parameters (preferred for performance).

Parameters

[in]	distance	Distance metric.
[in]	start	Cluster initialization strategy.
[in]	replicates	Number of K-Means replicates. Best is returned.
[in]	emptyact	What happens if a cluster goes empty.
[in]	online	If centroids should be updated during iterations.
[in]	maxit	Maximal number of iterations per replicate.

Definition at line 90 of file kmeans.cpp.

Member Function Documentation

calculate()

bool KMeans::calculate	(	const Eigen::MatrixXd &	X,
		qint32	kClusters,
		Eigen::VectorXi &	idx,
		Eigen::MatrixXd &	C,
		Eigen::VectorXd &	sumD,
		Eigen::MatrixXd &	D )

Clusters input data X.

Parameters

[in]	X	Input data (rows = points; cols = p dimensional space).
[in]	kClusters	Number of k clusters.
[out]	idx	The cluster indices to which cluster the input points belong to.
[out]	C	Cluster centroids k x p.
[out]	sumD	Summation of the distances to the centroid within one cluster.
[out]	D	Cluster distances to the centroid.

Returns

true if clustering succeeded, false otherwise.

Definition at line 140 of file kmeans.cpp.

---

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

• src/libraries/math/kmeans.h
• src/libraries/math/kmeans.cpp