Class PreDeCon

  • All Implemented Interfaces:
    Algorithm, ClusteringAlgorithm<Clustering<Model>>

    @Title("PreDeCon: Subspace Preference weighted Density Connected Clustering")
    @Description("PreDeCon computes clusters of subspace preference weighted connected points. The algorithm searches for local subgroups of a set of feature vectors having a low variance along one or more (but not all) attributes.")
    @Reference(authors="Christian B\u00f6hm, Karin Kailing, Hans-Peter Kriegel, Peer Kr\u00f6ger",
               title="Density Connected Clustering with Local Subspace Preferences",
               booktitle="Proc. 4th IEEE Int. Conf. on Data Mining (ICDM\'04)",
               url="https://doi.org/10.1109/ICDM.2004.10087",
               bibkey="DBLP:conf/icdm/BohmKKK04")
    public class PreDeCon
    extends GeneralizedDBSCAN
    PreDeCon computes clusters of subspace preference weighted connected points. The algorithm searches for local subgroups of a set of feature vectors having a low variance along one or more (but not all) attributes.

    Reference:

    Christian Böhm, Karin Kailing, Hans-Peter Kriegel, Peer Kröger
    Density Connected Clustering with Local Subspace Preferences.
    Proc. 4th IEEE Int. Conf. on Data Mining (ICDM'04)

    Since:
    0.1
    Author:
    Peer Kröger
    • Constructor Detail

      • PreDeCon

        public PreDeCon​(PreDeCon.Settings settings)
        Constructor.
        Parameters:
        settings - PreDeCon settings.