kmeans_v1.m File Reference

k-means clustering algorithm for complex-valued data. More...

Go to the source code of this file.

Functions
function	kmeans_v1 (in initPosArray, in symbVector, in varargin)
	This function applies the k-means algorithm to find the centers of complex-valued data clusters. More...

Detailed Description

k-means clustering algorithm for complex-valued data.

This function is faster than matlab's built-in kmeans, but has worse performance in finding the cluster centroids.

Example

   centers = kmeans_v1(initialCenters, DataPoints);
   centers = kmeans_v1(initialCenters, DataPoints, 'iterations', 50);
   centers = kmeans_v1(initialCenters, DataPoints, 'iterations', 50, 'tol', 1e-5);
   centers = kmeans_v1(initialCenters, DataPoints, 'iterations', 50, 'tol', 1e-5, 'gamma', 0.1);
%> 

Author

Edson Porto da Silva

Version

1

Definition in file kmeans_v1.m.

Function Documentation

function kmeans_v1	(	in	initPosArray,
		in	symbVector,
		in	varargin
	)

This function applies the k-means algorithm to find the centers of complex-valued data clusters.

Parameters

initPosArray	Initial particles positions (array of complex numbers)
symbVector	Received constellation (array of symbols + noise, rotation, etc)
gamma	Adaptation step. [Default: 0.5]
iterations	Max number of allowed k-means interations. [Default: 10]
tol	Convergence tolerance, i.e., if the average delta in the position the centers is lower than tol, the algorithm stops. [Default: 2e-3]

Return values

centers

Clusters centroids.