org.carrot2.matrix.factorization.seeding
Class KMeansSeedingStrategy

java.lang.Object
  org.carrot2.matrix.factorization.seeding.KMeansSeedingStrategy

All Implemented Interfaces:

ISeedingStrategy

public class KMeansSeedingStrategy
extends Object
implements ISeedingStrategy

Matrix seeding based on the k-means algorithms.

Constructor Summary

KMeansSeedingStrategy()
Creates the KMeansSeedingStrategy.

KMeansSeedingStrategy(int maxIterations)
Creates the KMeansSeedingStrategy.

Method Summary

void	seed(org.apache.mahout.math.matrix.DoubleMatrix2D A, org.apache.mahout.math.matrix.DoubleMatrix2D U, org.apache.mahout.math.matrix.DoubleMatrix2D V) Initializes values of the provided U and V matrices.
String	toString()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

KMeansSeedingStrategy

public KMeansSeedingStrategy()

Creates the KMeansSeedingStrategy.

KMeansSeedingStrategy

public KMeansSeedingStrategy(int maxIterations)

Creates the KMeansSeedingStrategy.

Parameters:

maxIterations - maximum number of KMeans iterations.

Method Detail

seed

public void seed(org.apache.mahout.math.matrix.DoubleMatrix2D A,
                 org.apache.mahout.math.matrix.DoubleMatrix2D U,
                 org.apache.mahout.math.matrix.DoubleMatrix2D V)

Description copied from interface: ISeedingStrategy

Initializes values of the provided U and V matrices. The A matrix is the input matrix to be factorized.

Specified by:

seed in interface ISeedingStrategy

Parameters:

A - matrix to be factorized

U - left factorized matrix to be seeded

V - right factorized matrix to be seeded

toString

public String toString()

Overrides:

toString in class Object