Seeders are responsible for initializing the starting clusters used by certain clusterers such as K Means, Mean Shift, and Gaussian Mixture. The choice of initializer can play an important role in determining the quality of the final solution derived by a learning algorithm.
To create k seeds from a dataset:
public function seed(Dataset $dataset, int $k) : array;
use Rubix\ML\Clusterers\Seeders\PlusPlus; $seeder = new PlusPlus(); $seeds = $seeder->seed($dataset, 10);