Package de.lmu.ifi.dbs.elki.algorithm.clustering.gdbscan

Generalized DBSCAN Generalized DBSCAN is an abstraction of the original DBSCAN idea, that allows the use of arbitrary "neighborhood" and "core point" predicates.

See: Description

Interface Summary

Interface	Description
CorePredicate<T>	Predicate for GeneralizedDBSCAN to evaluate whether a point is a core point or not.
CorePredicate.Instance<T>	Instance for a particular data set.
NeighborPredicate<T>	Get the neighbors of an object Note the Factory/Instance split of this interface.
NeighborPredicate.Instance<T>	Instance for a particular data set.

Class Summary

Class	Description
AbstractRangeQueryNeighborPredicate<O,M,N>	Abstract local model neighborhood predicate.
AbstractRangeQueryNeighborPredicate.Instance<N,M>	Instance for a particular data set.
AbstractRangeQueryNeighborPredicate.Parameterizer<O>	Parameterization class
COPACNeighborPredicate<V extends NumberVector>	COPAC neighborhood predicate.
COPACNeighborPredicate.COPACModel	Model used by COPAC for core point property.
COPACNeighborPredicate.Instance	Instance for a particular data set.
COPACNeighborPredicate.Parameterizer<V extends NumberVector>	Parameterization class.
EpsilonNeighborPredicate<O>	The default DBSCAN and OPTICS neighbor predicate, using an epsilon-neighborhood.
EpsilonNeighborPredicate.Instance	Instance for a particular data set.
EpsilonNeighborPredicate.Parameterizer<O>	Parameterization class
ERiCNeighborPredicate<V extends NumberVector>	ERiC neighborhood predicate.
ERiCNeighborPredicate.Parameterizer<V extends NumberVector>	Parameterization class.
FourCCorePredicate	The 4C core point predicate.
FourCCorePredicate.Instance	Instance for a particular data set.
FourCCorePredicate.Parameterizer	Parameterization class
FourCNeighborPredicate<V extends NumberVector>	4C identifies local subgroups of data objects sharing a uniform correlation.
FourCNeighborPredicate.Instance	Instance for a particular data set.
FourCNeighborPredicate.Parameterizer<O extends NumberVector>	Parameterization class.
GeneralizedDBSCAN	Generalized DBSCAN, density-based clustering with noise.
GeneralizedDBSCAN.Instance<T>	Instance for a particular data set.
GeneralizedDBSCAN.Parameterizer	Parameterization class
LSDBC<O extends NumberVector>	Locally Scaled Density Based Clustering.
LSDBC.Parameterizer<O extends NumberVector>	Parameterization class
MinPtsCorePredicate	The DBSCAN default core point predicate -- having at least MinPtsCorePredicate.minpts neighbors.
MinPtsCorePredicate.Instance	Instance for a particular data set.
MinPtsCorePredicate.Parameterizer	Parameterization class
PreDeConCorePredicate	The PreDeCon core point predicate -- having at least minpts. neighbors, and a maximum preference dimensionality of lambda.
PreDeConCorePredicate.Instance	Instance for a particular data set.
PreDeConCorePredicate.Parameterizer	Parameterization class
PreDeConNeighborPredicate<V extends NumberVector>	Neighborhood predicate used by PreDeCon.
PreDeConNeighborPredicate.Instance	Instance for a particular data set.
PreDeConNeighborPredicate.Parameterizer<V extends NumberVector>	Parameterization class.
PreDeConNeighborPredicate.PreDeConModel	Model used by PreDeCon for core point property.
SimilarityNeighborPredicate<O>	The DBSCAN neighbor predicate for a SimilarityFunction, using all neighbors with a minimum similarity.
SimilarityNeighborPredicate.Instance	Instance for a particular data set.
SimilarityNeighborPredicate.Parameterizer<O>	Parameterization class

Package de.lmu.ifi.dbs.elki.algorithm.clustering.gdbscan Description

Generalized DBSCAN

Generalized DBSCAN is an abstraction of the original DBSCAN idea, that allows the use of arbitrary "neighborhood" and "core point" predicates.

For each object, the neighborhood as defined by the "neighborhood" predicate is retrieved - in original DBSCAN, this is the objects within an epsilon sphere around the query object. Then the core point predicate is evaluated to decide if the object is considered dense. If so, a cluster is started (or extended) to include the neighbors as well.

Reference:

Jörg Sander, Martin Ester, Hans-Peter Kriegel, Xiaowei Xu
Density-Based Clustering in Spatial Databases: The Algorithm GDBSCAN and Its Applications
Data Mining and Knowledge Discovery, 1998.