R Graphical Manual

Browse All

Last data update: 2014.03.03

R: Best Of Set Optimal Partitions From Random Starts

bestopt

R Documentation

Best Of Set Optimal Partitions From Random Starts

Description

Produces a specified number of optpart solutions from random starts, keeping the best result of the set

Usage

bestopt(dist,numclu,numrep,maxitr=100)

Arguments

`dist`	an object of class ‘dist’ from `dist`, `vegdist`, or `dsvdis`, or a symmetric dissimilarity matrix
`numclu`	the number of clusters desired
`numrep`	the number of random starts requested
`maxitr`	the maximum number of iterations per replicate

Details

calls function optpart with an random initial assignment of items to clusters ‘numitr’ times, keeping the best result (highest within/among ratio observed). See optpart for more details.

Value

an object of class partana, with components:

`ptc`	the mean similarity of each item to each cluster
`ctc`	the mean similarity of each cluster to other clusters
`musubx`	the membership of each item in each cluster
`clustering`	the best ‘crisp’ partition from musubx
`ratio`	the within-cluster/among-cluster similarity ratio achieved at each iteration of the selected result.

Note

This is a simple wrapper function to automate independent random starts of function optpart.

Author(s)

David W. Roberts droberts@montana.edu

Examples

data(shoshveg)
dis.bc <- dsvdis(shoshveg,'bray/curtis')
x <- bestopt(dis.bc,5,10)
summary(x)
## Not run: plot(x)