Click or drag to resize

AverageLinkageTInstance Class

Implements the unweighted pair-group average method or UPGMA, i.e., returns the mean distance between the elements in each cluster.
Inheritance Hierarchy
SystemObject
  Aglomera.LinkageAverageLinkageTInstance

Namespace:  Aglomera.Linkage
Assembly:  Aglomera (in Aglomera.dll) Version: 1.1.0
Syntax
C#
Copy
public class AverageLinkage<TInstance> : ILinkageCriterion<TInstance>
where TInstance : Object, IComparable<TInstance>
Request Example View Source

Type Parameters

TInstance
The type of instance considered.

The AverageLinkageTInstance type exposes the following members.

Constructors
  NameDescription
Public methodAverageLinkageTInstance
Creates a new AverageLinkageTInstance with given dissimilarity metric.
Top
Properties
  NameDescription
Public propertyDissimilarityMetric
Gets the metric used to measure the dissimilarity / distance between cluster elements according to this linkage criterion.
Top
Methods
  NameDescription
Public methodCalculate
Calculates the distance / dissimilarity between the two given clusters according to this linkage criterion.
Public methodEquals (Inherited from Object.)
Protected methodFinalize (Inherited from Object.)
Public methodGetHashCode (Inherited from Object.)
Public methodGetType (Inherited from Object.)
Protected methodMemberwiseClone (Inherited from Object.)
Public methodToString (Inherited from Object.)
Top
Remarks
Average linkage tries to strike a balance between SingleLinkageTInstance and CompleteLinkageTInstance. It uses average pairwise dissimilarity, so clusters tend to be relatively compact and relatively far apart. However, it is not clear what properties the resulting clusters have when we cut an average linkage tree at given distance. Single and complete linkage trees each had simple interpretations [1]. References: [1] - http://www.stat.cmu.edu/~ryantibs/datamining/lectures/05-clus2-marked.pdf.
See Also