ch.usi.inf.sape.hac.agglomeration

Interface AgglomerationMethod

All Known Implementing Classes:

AverageLinkage, CentroidLinkage, CompleteLinkage, MedianLinkage, SingleLinkage, WardLinkage, WeightedAverageLinkage

public interface AgglomerationMethod

An AgglomerationMethod represents the Lance-Williams dissimilarity update formula used for hierarchical agglomerative clustering. The general form of the Lance-Williams matrix-update formula: d[(i,j),k] = ai*d[i,k] + aj*d[j,k] + b*d[i,j] + g*|d[i,k]-d[j,k]| Parameters ai, aj, b, and g are defined differently for different methods: Method ai aj b g ------------- ------------------ ------------------ ------------------------ ----- Single 0.5 0.5 0 -0.5 Complete 0.5 0.5 0 0.5 Average ci/(ci+cj) cj/(ci+cj) 0 0 Centroid ci/(ci+cj) cj/(ci+cj) -ci*cj/((ci+cj)*(ci+cj)) 0 Median 0.5 0.5 -0.25 0 Ward (ci+ck)/(ci+cj+ck) (cj+ck)/(ci+cj+ck) -ck/(ci+cj+ck) 0 WeightedAverage 0.5 0.5 0 0 (ci, cj, ck are cluster cardinalities)

See Also:

http://www.mathworks.com/help/toolbox/stats/linkage.html, http://www.stanford.edu/~maureenh/quals/html/ml/node73.html, data analysis handbook. By Ildiko E. Frank, Roberto Todeschini. Pages 152-155]

Method Summary

Modifier and Type	Method and Description
double	computeDissimilarity(double dik, double djk, double dij, int ci, int cj, int ck) Compute the dissimilarity between the newly formed cluster (i,j) and the existing cluster k.

Method Detail

computeDissimilarity

double computeDissimilarity(double dik,
                            double djk,
                            double dij,
                            int ci,
                            int cj,
                            int ck)

Compute the dissimilarity between the newly formed cluster (i,j) and the existing cluster k.

Parameters:

dik - dissimilarity between clusters i and k

djk - dissimilarity between clusters j and k

dij - dissimilarity between clusters i and j

ci - cardinality of cluster i

cj - cardinality of cluster j

ck - cardinality of cluster k

Returns:

dissimilarity between cluster (i,j) and cluster k.