smile.clustering.linkage

Class Linkage

java.lang.Object

smile.clustering.linkage.Linkage

Direct Known Subclasses:

CompleteLinkage, SingleLinkage, UPGMALinkage, UPGMCLinkage, WardLinkage, WPGMALinkage, WPGMCLinkage

public abstract class Linkage
extends java.lang.Object

A measure of dissimilarity between clusters (i.e. sets of observations).

References

1. Anil K. Jain, Richard C. Dubes. Algorithms for clustering data. 1988.

See Also:

HierarchicalClustering

Constructor Summary

Constructors

Constructor and Description
Linkage(double[][] proximity) Initialize the linkage with the lower triangular proximity matrix.
Linkage(int size, float[] proximity) Initialize the linkage with the lower triangular proximity matrix.

Method Summary

Modifier and Type	Method and Description
float	d(int i, int j) Returns the distance/dissimilarity between two clusters/objects, which are indexed by integers.
abstract void	merge(int i, int j) Merge two clusters into one and update the proximity matrix.
static float[]	proximity(double[][] data) Calculate the proximity matrix (linearized in column major) with Euclidean distance.
static <T> float[]	proximity(T[] data, smile.math.distance.Distance<T> distance) Calculate the proximity matrix (linearized in column major).
int	size() Returns the proximity matrix size.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Linkage

public Linkage(double[][] proximity)

Initialize the linkage with the lower triangular proximity matrix.

Linkage

public Linkage(int size,
               float[] proximity)

Initialize the linkage with the lower triangular proximity matrix.

Parameters:

size - the data size.

proximity - column-wise linearized proximity matrix that stores only the lower half. The length of proximity should be size * (size+1) / 2. To save space, Linkage will use this argument directly without copy. The elements may be modified.

Method Detail

size

public int size()

Returns the proximity matrix size.

d

public float d(int i,
               int j)

Returns the distance/dissimilarity between two clusters/objects, which are indexed by integers.

merge

public abstract void merge(int i,
                           int j)

Merge two clusters into one and update the proximity matrix.

Parameters:

i - cluster id.

j - cluster id.

proximity

public static float[] proximity(double[][] data)

Calculate the proximity matrix (linearized in column major) with Euclidean distance.

proximity

public static <T> float[] proximity(T[] data,
                                    smile.math.distance.Distance<T> distance)

Calculate the proximity matrix (linearized in column major).