- All Implemented Interfaces:
- AgglomerationMethod
public final class CompleteLinkage
extends Object
implements AgglomerationMethod
The "complete", "maximum", "clique",
"furthest neighbor", or "furthest distance" method is a graph-based approach.
The distance between two clusters is calculated as the largest distance
between two objects in opposite clusters.
This method tends to produce well separated, small, compact spherical clusters.
The cluster space is dilated.
[The data analysis handbook. By Ildiko E. Frank, Roberto Todeschini]
This method tends to produce compact clusters. Outliers are given more weight with this method.
It is generally a good choice if the clusters are far apart in feature space, but not good if the data are noisy.
- See Also:
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]|
For the "single linkage" method:
ai = 0.5
aj = 0.5
b = 0
g = 0.5
Thus:
d[(i,j),k] = 0.5*d[i,k] + 0.5*d[j,k] + 0.5*|d[i,k]-d[j,k]|
= 0.5*d[i,k] + 0.5*d[j,k] + | 0.5*d[i,k] - 0.5*d[j,k] |
= d[i,j]