cluster.cpp File Reference

#include "geoslib_old_f.h"
#include "Basic/Law.hpp"
#include <limits.h>

Macros
#define	DBL_MAXIMUM   1.e30
#define	DATA(iech, ivar)   (data[(iech) * nvar + (ivar)])
#define	DATA1(iech, ivar)   (data1[(iech) * nvar + (ivar)])
#define	DATA2(iech, ivar)   (data2[(iech) * nvar + (ivar)])
#define	CDATA(icl, ivar)   (cdata[(icl) * nvar + (ivar)])
#define	DISTMATRIX(i, j)   (distmatrix[(i) * nech + (j)])

Functions
static double	st_distance (int nvar, const double *data1, const double *data2, int index1, int index2)
static void	st_randomassign (int nclusters, int nech, int *clusterid)
static void	st_printclusterlist (int nclusters, int nech, int *clusterid)
static void	st_printclustercount (int nclusters, int nech, const int *clusterid)
static void	st_getclustermeans (const double *data, int nvar, int nech, int nclusters, const int *clusterid, double *cdata, int *cmask)
static void	st_getclustermedian (const double *data, int nvar, int nech, int nclusters, const int *clusterid, double *cdata, double *cache)
static double *	st_get_distmatrix (double *data, int nvar, int nech)
static void	st_getclustermedoids (int nech, int nclusters, const double *distmatrix, const int *clusterid, int *centroids, double *errors)
double *	kclusters (double *data, int nvar, int nech, int nclusters, int npass, int mode, int verbose)
int *	kmedoids (double *data, int nvar, int nech, int nclusters, int npass, int verbose)

Macro Definition Documentation

CDATA

#define CDATA	(		icl,
			ivar
	)		(cdata[(icl) * nvar + (ivar)])

DATA

#define DATA	(		iech,
			ivar
	)		(data[(iech) * nvar + (ivar)])

DATA1

#define DATA1	(		iech,
			ivar
	)		(data1[(iech) * nvar + (ivar)])

DATA2

#define DATA2	(		iech,
			ivar
	)		(data2[(iech) * nvar + (ivar)])

DBL_MAXIMUM

#define DBL_MAXIMUM   1.e30

DISTMATRIX

#define DISTMATRIX	(		i,
			j
	)		(distmatrix[(i) * nech + (j)])

Function Documentation

kclusters()

double* kclusters	(	double *	data,
		int	nvar,
		int	nech,
		int	nclusters,
		int	npass,
		int	mode,
		int	verbose
	)

Perform k-means clustering on a given set of variables. The number of clusters is given in input

Returns

Array of cluster coordinates

Parameters

[in]	data	Array of values
[in]	nvar	Number of samples
[in]	nech	Number of variables
[in]	nclusters	Number if clusters
[in]	npass	Number of times clustering is performed
[in]	mode	0 for k-means and 1 for k-medians
[in]	verbose	Verbose option

kmedoids()

int* kmedoids	(	double *	data,
		int	nvar,
		int	nech,
		int	nclusters,
		int	npass,
		int	verbose
	)

Perform k-medoids clustering on a given set of variables. The number of clusters is given in input

Returns

Array of cluster indices

Parameters

[in]	data	Array of values
[in]	nvar	Number of samples
[in]	nech	Number of variables
[in]	nclusters	Number if clusters
[in]	npass	Number of times clustering is performed
[in]	verbose	Verbose option

st_distance()

static double st_distance ( int  nvar, const double *  data1, const double *  data2, int  index1, int  index2  )

static

Calculate the distance between two samples

Parameters

[in]	nvar	Number of variables
[in]	data1	Array of values for first part
[in]	data2	Array of values for second part
[in]	index1	Rank of the first sample
[in]	index2	Rank of the second sample

st_get_distmatrix()

static double* st_get_distmatrix ( double *  data, int  nvar, int  nech  )

static

Evaluate the distance matrix

Parameters

[in]	data	Array of values
[in]	nvar	Number of samples
[in]	nech	Number of variables

st_getclustermeans()

static void st_getclustermeans ( const double *  data, int  nvar, int  nech, int  nclusters, const int *  clusterid, double *  cdata, int *  cmask  )

static

Find the center of a cluster

Parameters

[in]	nclusters	Number if clusters
[in]	nvar	Number of samples
[in]	nech	Number of variables
[in]	data	Array of values
[in]	clusterid	Array of cluster number for each sample
[out]	cdata	Array containing the centroids
[out]	cmask	Array containing the number of sample per cluster

st_getclustermedian()

static void st_getclustermedian ( const double *  data, int  nvar, int  nech, int  nclusters, const int *  clusterid, double *  cdata, double *  cache  )

static

Find the median of a cluster

Parameters

[in]	nclusters	Number if clusters
[in]	nvar	Number of samples
[in]	nech	Number of variables
[in]	data	Array of values
[in]	clusterid	Array of cluster number for each sample
[out]	cdata	Array containing the centroids
[out]	cache	Array for storing data of a cluster

st_getclustermedoids()

static void st_getclustermedoids ( int  nech, int  nclusters, const double *  distmatrix, const int *  clusterid, int *  centroids, double *  errors  )

static

Find the center of a cluster

Parameters

[in]	distmatrix	Matrix of distances
[in]	nech	Number of variables
[in]	nclusters	Number if clusters
[in]	clusterid	Array of cluster number for each sample
[out]	centroids	Array of centroids
[out]	errors	Array containing within-cluster distances

st_printclustercount()

static void st_printclustercount ( int  nclusters, int  nech, const int *  clusterid  )

static

Print the number of samples per cluster

Parameters

[in]	nclusters	Number if clusters
[in]	nech	Number of samples
[in]	clusterid	Array of cluster number for each sample

st_printclusterlist()

static void st_printclusterlist ( int  nclusters, int  nech, int *  clusterid  )

static

Print the list of cluster per sample

Parameters

[in]	nclusters	Number if clusters
[in]	nech	Number of samples
[in]	clusterid	Array of cluster number for each sample

st_randomassign()

static void st_randomassign ( int  nclusters, int  nech, int *  clusterid  )

static

Initial random assignment of samples to clusters

Parameters

[in]	nclusters	Number if clusters
[in]	nech	Number of samples
[out]	clusterid	Array of cluster number for each sample