Package elki.clustering.em

Class KDTreeEM.Par

  • All Implemented Interfaces:
    Parameterizer
    Enclosing class:
    KDTreeEM
    public static class KDTreeEM.Par
extends java.lang.Object
implements Parameterizer
    Parameterization class.
    Author:
    Erich Schubert, Robert Gehde

    • Field Summary

      Fields 
      Modifier and Type Field Description
      protected double delta
      Stopping threshold
      static OptionID DELTA_ID
      Parameter to specify the termination criterion for maximization of E(M): E(M) - E(M') < em.delta, must be a double equal to or greater than 0.
      static OptionID EXACT_ASSIGN_ID
      Parameter to produce more precise final assignments
      (package private) boolean exactAssign
      Perform the slower exact assignment step.
      protected int k
      Number of clusters.
      static OptionID K_ID
      Parameter to specify the number of clusters to find.
      protected int maxiter
      Maximum number of iterations.
      static OptionID MAXITER_ID
      Parameter to specify a maximum number of iterations
      protected double mbw
      construction threshold
      static OptionID MBW_ID
      Parameter to specify the termination criterion for kd-tree construction.
      protected TextbookMultivariateGaussianModelFactory mfactory
      Cluster model factory.
      protected int miniter
      Minimum number of iterations.
      static OptionID MINITER_ID
      Parameter to specify a minimum number of iterations
      (package private) boolean soft
      Retain soft assignments?
      static OptionID SOFT_ID
      Parameter to specify the saving of soft assignments
      protected double tau
      cutoff threshold
      static OptionID TAU_CLASS_ID
      drop one class if the maximum weight of a class in the bounding box is lower than tauClass * wmin_max, where wmin_max is the maximum minimum weight of all classes
      static OptionID TAU_ID
      Parameter to specify the pruning criterion during the algorithm.
      protected double tauclass
      cutoff safety threshold

    • Constructor Summary

      Constructors 
      Constructor Description
      Par()  

    • Method Summary

      All Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      void configure​(Parameterization config)
      Configure the class.
      KDTreeEM make()
      Make an instance after successful configuration.

      • Methods inherited from class java.lang.Object

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

    • Field Detail

      • K_ID

        public static final OptionID K_ID
        Parameter to specify the number of clusters to find.

      • DELTA_ID

        public static final OptionID DELTA_ID
        Parameter to specify the termination criterion for maximization of E(M): E(M) - E(M') < em.delta, must be a double equal to or greater than 0.

      • MBW_ID

        public static final OptionID MBW_ID
        Parameter to specify the termination criterion for kd-tree construction. Stop splitting nodes when the width is smaller then mbw * dataset_width. Must be between 0 and 1.

      • TAU_ID

        public static final OptionID TAU_ID
        Parameter to specify the pruning criterion during the algorithm. Stop going down the kd-tree when possible weight error e < tau * totalweight. Must be between 0 and 1. Low for precise, high for fast results.

      • TAU_CLASS_ID

        public static final OptionID TAU_CLASS_ID
        drop one class if the maximum weight of a class in the bounding box is lower than tauClass * wmin_max, where wmin_max is the maximum minimum weight of all classes

      • MINITER_ID

        public static final OptionID MINITER_ID
        Parameter to specify a minimum number of iterations

      • MAXITER_ID

        public static final OptionID MAXITER_ID
        Parameter to specify a maximum number of iterations

      • SOFT_ID

        public static final OptionID SOFT_ID
        Parameter to specify the saving of soft assignments

      • EXACT_ASSIGN_ID

        public static final OptionID EXACT_ASSIGN_ID
        Parameter to produce more precise final assignments

      • k

        protected int k
        Number of clusters.

      • mbw

        protected double mbw
        construction threshold

      • tau

        protected double tau
        cutoff threshold

      • tauclass

        protected double tauclass
        cutoff safety threshold

      • delta

        protected double delta
        Stopping threshold

      • miniter

        protected int miniter
        Minimum number of iterations.

      • maxiter

        protected int maxiter
        Maximum number of iterations.

      • soft

        boolean soft
        Retain soft assignments?

      • exactAssign

        boolean exactAssign
        Perform the slower exact assignment step.

    • Constructor Detail

      • Par

        public Par()

    • Method Detail

      • configure

        public void configure​(Parameterization config)
        Description copied from interface: Parameterizer
        Configure the class.

        Note: the status is collected by the parameterization object, so that multiple errors may arise and be reported in one run.

        Specified by:
        configure in interface Parameterizer
        Parameters:
        config - Parameterization

      • make

        public KDTreeEM make()
        Description copied from interface: Parameterizer
        Make an instance after successful configuration.

        Note: your class should return the exact type, only this very broad interface should use Object as return type.

        Specified by:
        make in interface Parameterizer
        Returns:
        a new instance