Package smile.clustering

Class HierarchicalClustering

java.lang.Object
smile.clustering.HierarchicalClustering
All Implemented Interfaces:
Serializable
public class HierarchicalClustering extends Object implements Serializable
Agglomerative Hierarchical Clustering. Hierarchical agglomerative clustering seeks to build a hierarchy of clusters in a bottom up approach: each observation starts in its own cluster, and pairs of clusters are merged as one moves up the hierarchy. The results of hierarchical clustering are usually presented in a dendrogram.

In general, the merges are determined in a greedy manner. In order to decide which clusters should be combined, a measure of dissimilarity between sets of observations is required. In most methods of hierarchical clustering, this is achieved by use of an appropriate metric, and a linkage criteria which specifies the dissimilarity of sets as a function of the pairwise distances of observations in the sets.

Hierarchical clustering has the distinct advantage that any valid measure of distance can be used. In fact, the observations themselves are not required: all that is used is a matrix of distances.

References

  1. David Eppstein. Fast hierarchical clustering and other applications of dynamic closest pairs. SODA 1998.
See Also:

  • Constructor Summary

    Constructors
    Constructor
    Description
    HierarchicalClustering(int[][] tree, double[] height)
    Constructor.

  • Method Summary

    Modifier and Type
    Method
    Description
    static HierarchicalClustering
    fit(Linkage linkage)
    Fits the Agglomerative Hierarchical Clustering with given linkage method, which includes proximity matrix.
    double[]
    height()
    Returns a set of n-1 non-decreasing real values, which are the clustering height, i.e., the value of the criterion associated with the clustering method for the particular agglomeration.
    int[]
    partition(double h)
    Cuts a tree into several groups by specifying the cut height.
    int[]
    partition(int k)
    Cuts a tree into several groups by specifying the desired number.
    int[][]
    tree()
    Returns an n-1 by 2 matrix of which row i describes the merging of clusters at step i of the clustering.

    Methods inherited from class java.lang.Object

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

  • Constructor Details

    • HierarchicalClustering

      public HierarchicalClustering(int[][] tree, double[] height)
      Constructor.
      Parameters:
      tree - an n-1 by 2 matrix of which row i describes the merging of clusters at step i of the clustering.
      height - the clustering height.

  • Method Details

    • fit

      public static HierarchicalClustering fit(Linkage linkage)
      Fits the Agglomerative Hierarchical Clustering with given linkage method, which includes proximity matrix.
      Parameters:
      linkage - a linkage method to merge clusters. The linkage object includes the proximity matrix of data.
      Returns:
      the model.

    • tree

      public int[][] tree()
      Returns an n-1 by 2 matrix of which row i describes the merging of clusters at step i of the clustering. If an element j in the row is less than n, then observation j was merged at this stage. If j >= n then the merge was with the cluster formed at the (earlier) stage j-n of the algorithm.
      Returns:
      the merge tree.

    • height

      public double[] height()
      Returns a set of n-1 non-decreasing real values, which are the clustering height, i.e., the value of the criterion associated with the clustering method for the particular agglomeration.
      Returns:
      the tree node height.

    • partition

      public int[] partition(int k)
      Cuts a tree into several groups by specifying the desired number.
      Parameters:
      k - the number of clusters.
      Returns:
      the cluster label of each sample.

    • partition

      public int[] partition(double h)
      Cuts a tree into several groups by specifying the cut height.
      Parameters:
      h - the cut height.
      Returns:
      the cluster label of each sample.