Package elki.clustering.kmeans.initialization

Class RandomlyChosen<O>

  • Type Parameters:
    O - Vector type
    All Implemented Interfaces:
    KMeansInitialization, KMedoidsInitialization<O>
    @Reference(authors="D. J. McRae",title="MIKCA: A FORTRAN IV Iterative K-Means Cluster Analysis Program",booktitle="Behavioral Science 16(4)",bibkey="journals/misc/McRae71") @Reference(authors="E. W. Forgy",title="Cluster analysis of multivariate data: efficiency versus interpretability of classifications",booktitle="Biometrics 21(3)",bibkey="journals/biometrics/Forgy65") @Reference(authors="M. R. Anderberg",title="Nonhierarchical Clustering Methods",booktitle="Cluster Analysis for Applications",bibkey="books/academic/Anderberg73/Ch7")
public class RandomlyChosen<O>
extends AbstractKMeansInitialization
implements KMedoidsInitialization<O>
    Initialize K-means by randomly choosing k existing elements as initial cluster centers.

    Reference:

    D. J. McRae
    MIKCA: A FORTRAN IV Iterative K-Means Cluster Analysis Program
    Behavioral Science 16(4)

    E. W. Forgy
    Cluster analysis of multivariate data: efficiency versus interpretability of classifications
    Abstract published in Biometrics 21(3)

    M. R. Anderberg
    Hierarchical Clustering Methods
    Cluster Analysis for Applications

    This initialization is often attributed to Forgy (but this is also debated), but we were unable to verify neither McRae (not available online?) nor Forgy so far (apparently, only an abstract is available in print, so we mostly can rely on indirect references, such as Anderberg).

    Since:
    0.5.0
    Author:
    Erich Schubert

    • Nested Class Summary

      Nested Classes 
      Modifier and Type Class Description
      static class  RandomlyChosen.Par<V>
      Parameterization class.

    • Constructor Summary

      Constructors 
      Constructor Description
      RandomlyChosen​(elki.utilities.random.RandomFactory rnd)
      Constructor.

    • Method Summary

      All Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      double[][] chooseInitialMeans​(elki.database.relation.Relation<? extends elki.data.NumberVector> relation, int k, elki.distance.NumberVectorDistance<?> distance)
      Choose initial means
      elki.database.ids.DBIDs chooseInitialMedoids​(int k, elki.database.ids.DBIDs ids, elki.database.query.distance.DistanceQuery<? super O> distance)
      Choose initial means

      • Methods inherited from class java.lang.Object

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

    • Constructor Detail

      • RandomlyChosen

        public RandomlyChosen​(elki.utilities.random.RandomFactory rnd)
        Constructor.
        Parameters:
        rnd - Random generator.

    • Method Detail

      • chooseInitialMeans

        public double[][] chooseInitialMeans​(elki.database.relation.Relation<? extends elki.data.NumberVector> relation,
                                     int k,
                                     elki.distance.NumberVectorDistance<?> distance)
        Description copied from interface: KMeansInitialization
        Choose initial means
        Specified by:
        chooseInitialMeans in interface KMeansInitialization
        Parameters:
        relation - Relation
        k - Parameter k
        distance - Distance function
        Returns:
        List of chosen means for k-means

      • chooseInitialMedoids

        public elki.database.ids.DBIDs chooseInitialMedoids​(int k,
                                                    elki.database.ids.DBIDs ids,
                                                    elki.database.query.distance.DistanceQuery<? super O> distance)
        Description copied from interface: KMedoidsInitialization
        Choose initial means
        Specified by:
        chooseInitialMedoids in interface KMedoidsInitialization<O>
        Parameters:
        k - Parameter k
        ids - Candidate IDs.
        distance - Distance function
        Returns:
        List of chosen means for k-means