kmeans.h File Reference

Declarations of kmeans clustering algorithm and associated functions. More...

#include <jwsc/config.h>
#include <stdlib.h>
#include <inttypes.h>
#include "jwsc/vector/vector.h"
#include "jwsc/matrix/matrix.h"

Include dependency graph for kmeans.h:

Go to the source code of this file.

Functions
void	kmeans_d (Matrix_d **means_out, Vector_i32 **members_out, const Matrix_d *features, uint32_t K, uint32_t N)
	K Means algorithm.
void	init_kmeans_randomly_d (Matrix_d **means_out, const Matrix_d *features, uint32_t K)
	Randomly chooses k feature vectors to initialize the means with.

---

Detailed Description

Declarations of kmeans clustering algorithm and associated functions.

Author:

Joseph Schlecht

License:

Creative Commons BY-NC-SA 3.0

Definition in file kmeans.h.

---

Function Documentation

void kmeans_d	(	Matrix_d **	means_out,
		Vector_i32 **	members_out,
		const Matrix_d *	features,
		uint32_t	K,
		uint32_t	N
	)

K Means algorithm.

Implements a fairly standard k-means algorithm.

Currently, the algorithm handles < N mean membership as follows: if a mean has n < N membership, N-n feature vectors are choosen at random to be members. Note that the random number generator is assumed to already be seeded (rand() is used).

Parameters:

means_out	Result parameter. If *means_out is NULL when the function is called, the means are initialized by calling init_means_randomly(). Otherwise, *means_out is assumed to contain initialized means. Contains the resulting means when the function returns.
members_out	Result parameter. When the function returns, (*members_out)->elts[i] contains the mean number (row index into *means_out) for the ith feature vector.
features	Feature vectors to assign membership to means.
K	Number of means.
N	Minimum number of records per mean.

Warning:

If *means_out is non-NULL, it is assumed to be a matrix of initialized means.

Definition at line 85 of file kmeans.c.

void init_kmeans_randomly_d	(	Matrix_d **	means_out,
		const Matrix_d *	features,
		uint32_t	K
	)

Randomly chooses k feature vectors to initialize the means with.

Parameters:

means_out	Result parameter. When the function returns, contains the initialized means.
features	Feature vectors to randomly choose from for mean initialization.
K	Number of means.

Note:

Assumes the random number generator has already been seeded (uses rand()).

Definition at line 237 of file kmeans.c.