Skip to content

[source]

Plus Plus#

This seeder attempts to maximize the chances of seeding distant clusters while still remaining random. It does so by sequentially selecting random samples weighted by their distance from the previous seed.

Parameters#

# Name Default Type Description
1 kernel Euclidean Distance The distance kernel used to compute the distance between samples.

Example#

use Rubix\ML\Clusterers\Seeders\PlusPlus;
use Rubix\ML\Kernels\Distance\Minkowski;

$seeder = new PlusPlus(new Minkowski(5.0));

References#

  1. D. Arthur et al. (2006). k-means++: The Advantages of Careful Seeding. ↩

  2. A. Stetco et al. (2015). Fuzzy C-means++: Fuzzy C-means with effective seeding initialization. ↩

Last update: 2021-03-03