@Reference(authors="G. N. Lance, W. T. Williams", title="A general theory of classificatory sorting strategies 1. Hierarchical systems", booktitle="The Computer Journal 9.4", url="https://doi.org/10.1093/comjnl/9.4.373", bibkey="doi:10.1093/comjnl/9.4.373") @Alias(value={"flex","beta","de.lmu.ifi.dbs.elki.algorithm.clustering.hierarchical.FlexibleBetaLinkageMethod"}) @Priority(value=-2) public class FlexibleBetaLinkage extends java.lang.Object implements Linkage
Beta values larger than 0 cause chaining, and are thus not recommended. Instead, choose a value between -1 and 0.
The general form of the recursive definition is: \[d_{\text{Flexible},\beta}(A\cup B, C) := \tfrac{1-\beta}{2} d(A,C) + \tfrac{1-\beta}{2} d(B,C) + \beta d(A,B) \]
Reference:
G. N. Lance, W. T. Williams
A general theory of classificatory sorting strategies
1. Hierarchical systems
The Computer Journal 9.4
Modifier and Type | Class and Description |
---|---|
static class |
FlexibleBetaLinkage.Parameterizer
Parameterization class.
|
Modifier and Type | Field and Description |
---|---|
(package private) double |
alpha
Alpha parameter, derived from beta.
|
(package private) double |
beta
Beta parameter
|
Constructor and Description |
---|
FlexibleBetaLinkage(double beta)
Constructor.
|
Modifier and Type | Method and Description |
---|---|
double |
combine(int sizex,
double dx,
int sizey,
double dy,
int sizej,
double dxy)
Compute combined linkage for two clusters.
|
public FlexibleBetaLinkage(double beta)
beta
- Beta parameterpublic double combine(int sizex, double dx, int sizey, double dy, int sizej, double dxy)
Linkage
combine
in interface Linkage
sizex
- Size of first cluster x before mergingdx
- Distance of cluster x to j before mergingsizey
- Size of second cluster y before mergingdy
- Distance of cluster y to j before mergingsizej
- Size of candidate cluster jdxy
- Distance between clusters x and y before mergingCopyright © 2019 ELKI Development Team. License information.