consider: Recommendations for Possible Merging of Clusters

In optpart: Optimal Partitioning of Similarity Relations

Description

Presents an ordered list of possible cluster combinations to consider for merging to simplify a classification.

Usage

consider(part)

Arguments

part	an object of class ‘partana’ from functions partana, optpart or bestopt

Details

Simply sorts the cluster-to-cluster mean similarity matrix of a ‘partana’ object into a list sorted by mean similarity.

Value

a data.frame with three elements:

row	the current cluster number
col	the cluster to which it is most similar
vals	the mean similarity of the two clusters

Note

The listed combinations are not suggested to be optimal by any specific criterion. In fact, if the ‘partana’ object was generated by optpart or bestopt it is known that the suggested combinations are sub-optimal. Nevertheless, sometimes it is desirable to simplify a classification for other reasons.

Author(s)

David W. Roberts [e-mail]

References

http://ecology.msu.montana.edu/labdsv/R

See Also

optpart, bestopt

Examples

data(shoshveg)                    # returns a vegetation data.frame
dis.bc <- dsvdis(shoshveg,'bray') # calculates a Bray/Curtis
                                  # dissimilarity matrix
opt.5 <- optpart(5,dis.bc)        # generates a 5 cluster partition
consider(opt.5)                   # recommends possible clusters to merge

optpart documentation built on March 26, 2020, 6:18 p.m.