jMEF.BregmanHierarchicalClustering Class Reference

List of all members.

Public Types
enum	LINKAGE_CRITERION { MINIMUM_DISTANCE, MAXIMUM_DISTANCE, AVERAGE_DISTANCE }
	Type of the likage criterion used. More...
Static Public Member Functions
static HierarchicalMixtureModel	build (MixtureModel f, Clustering.CLUSTERING_TYPE type, LINKAGE_CRITERION linkage)
	Builds a hierarchical mixture model (class HierarchicalMixtureModel) from an input mixture model and a clustering type.

---

Detailed Description

Author:

Vincent Garcia

Frank Nielsen

Version:

1.0

License

See file LICENSE.txt

Description

The Bregman hierarchical clustering is the generalization of the hierarchical clustering towards the exponential family. Given a set of weighted distributions (mixture model), the Bregman hierarchical clustering builds a hierarchical mixture model (class HierarchicalMixtureModel). It is then possible from this hierarchical mixture model to

• quickly deduce a simpler mixture model,
• automatically learn the "optimal" number of components in the mixture model.

---

Member Enumeration Documentation

enum jMEF::BregmanHierarchicalClustering::LINKAGE_CRITERION

Type of the likage criterion used.

Enumerator:

MINIMUM_DISTANCE
MAXIMUM_DISTANCE
AVERAGE_DISTANCE

---

Member Function Documentation

static HierarchicalMixtureModel jMEF.BregmanHierarchicalClustering.build	(	MixtureModel	f,
		Clustering.CLUSTERING_TYPE	type,
		LINKAGE_CRITERION	linkage
	)			[static]

Builds a hierarchical mixture model (class HierarchicalMixtureModel) from an input mixture model and a clustering type.

Parameters:

	f	input mixture model given in source parameters
	type	type of the Bregman divergence used: right-sided, left-sided, or symmetric
	linkage	linkage criterion used: minimum, maximum, or average distance

Returns:

a hierarchical mixture model

---

The documentation for this class was generated from the following file:

• BregmanHierarchicalClustering.java