KMeans Class Reference

Implementation of the k-means algorithm. More...

#include <KMeans.hpp>

Collaboration diagram for KMeans:

Public Types
enum	InitializationMethod { RANDOM, SAMPLE }

Public Member Functions
	KMeans ()
MatrixD	predict (MatrixD data)
	Assigns elements of a data set to clusters. More...
void	fit (MatrixD data, unsigned int k, unsigned int iters=100, unsigned int inits=100, double distance=2, InitializationMethod initMethod=SAMPLE, bool verbose=false)
	Find the k centroids that best fit the data. More...
const MatrixD &	getY () const
const MatrixD &	getCentroids () const
unsigned int	getK () const
unsigned int	getTotalIterations () const
double	getDistance () const
double	getSse () const

Private Member Functions
double	SSE ()

Private Attributes
MatrixD	X
MatrixD	y
MatrixD	centroids
unsigned int	k
unsigned int	totalIterations
double	distance
double	sse
InitializationMethod	initMethod

Detailed Description

Implementation of the k-means algorithm.

Author

Douglas De Rizzo Meneghetti (dougl.nosp@m.asri.nosp@m.zzom@.nosp@m.gmai.nosp@m.l.com)

Date

2017-10-25 Implementaion of the k-means algorithm

Definition at line 17 of file KMeans.hpp.

Member Enumeration Documentation

InitializationMethod

enum KMeans::InitializationMethod

Enumerator
RANDOM
SAMPLE

Definition at line 19 of file KMeans.hpp.

Constructor & Destructor Documentation

KMeans()

KMeans::KMeans ( )

inline

Definition at line 34 of file KMeans.hpp.

Member Function Documentation

fit()

void KMeans::fit ( MatrixD  data, unsigned int  k, unsigned int  iters = 100, unsigned int  inits = 100, double  distance = 2, InitializationMethod  initMethod = SAMPLE, bool  verbose = false  )

inline

Find the k centroids that best fit the data.

Parameters

data	a Matrix containing the data
k	number of clusters to be generated
iters	number of maximum assignment/adjustment iterations
inits	number of algorithm reinitialization
distance	L norm of the distance measure to be used (1 for Manhattan, 2 for Euclidean etc.)
initMethod	centroid initialization method
verbose	whether to output progress or not

Definition at line 72 of file KMeans.hpp.

Here is the call graph for this function:

getCentroids()

const MatrixD& KMeans::getCentroids ( ) const

inline

Definition at line 129 of file KMeans.hpp.

getDistance()

double KMeans::getDistance ( ) const

inline

Definition at line 141 of file KMeans.hpp.

getK()

unsigned int KMeans::getK ( ) const

inline

Definition at line 133 of file KMeans.hpp.

getSse()

double KMeans::getSse ( ) const

inline

Definition at line 145 of file KMeans.hpp.

getTotalIterations()

unsigned int KMeans::getTotalIterations ( ) const

inline

Definition at line 137 of file KMeans.hpp.

getY()

const MatrixD& KMeans::getY ( ) const

inline

Definition at line 125 of file KMeans.hpp.

predict()

MatrixD KMeans::predict ( MatrixD  data )

inline

Assigns elements of a data set to clusters.

Parameters

data	a Matrix containing elements in rows and features in columns

Returns

column vector with the index of clusters each element is assigned to

Definition at line 41 of file KMeans.hpp.

Here is the call graph for this function:

Here is the caller graph for this function:

SSE()

double KMeans::SSE ( )

inlineprivate

Returns

Sum of squared errors between elements and their centroids

Definition at line 29 of file KMeans.hpp.

Here is the call graph for this function:

Here is the caller graph for this function:

Field Documentation

centroids

MatrixD KMeans::centroids

private

Definition at line 21 of file KMeans.hpp.

distance

double KMeans::distance

private

Definition at line 23 of file KMeans.hpp.

initMethod

InitializationMethod KMeans::initMethod

private

Definition at line 24 of file KMeans.hpp.

k

unsigned int KMeans::k

private

Definition at line 22 of file KMeans.hpp.

sse

double KMeans::sse

private

Definition at line 23 of file KMeans.hpp.

totalIterations

unsigned int KMeans::totalIterations

private

Definition at line 22 of file KMeans.hpp.

X

MatrixD KMeans::X

private

Definition at line 21 of file KMeans.hpp.

y

MatrixD KMeans::y

private

Definition at line 21 of file KMeans.hpp.

---

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

• include/KMeans.hpp