clusterfly: Creates a convenient data structure for dealing with a...
In ggobi/clusterfly: Explore clustering interactively using R and GGobi

Description Usage Arguments Details See Also Examples

Once you have created a clusterfly object, you can add clusterings to it with cfly_cluster, and visualise then in GGobi with cfly_show and cfly_animate. Static graphics are also available: cfly_pcp will produce a parallel coordinates plot, cfly_dist will show the distribution of each variable in each cluster, and cfly_fluct compares two clusterings with a fluctuation diagram.

1	clusterfly(df, extra = NULL, rescale = TRUE)

`df`	data frame to be clustered
`extra`	extra variables to be included in output, but not clustered
`rescale`	rescale, if true each variable will be scaled to have mean 0 and variance 1.

If you want to standardise the cluster labelling to one group, look at clarify and cfly_clarify

vignette("introduction")

olives <- read.csv(ggobi_find_file("data","olive.csv"))
ol <- clusterfly(olives[, -(1:3)], olives[, 2:3])
ol <- cfly_cluster(ol, kmeans, 4, name="k4-1")
ol <- cfly_cluster(ol, kmeans, 4, name="k4-2")
ol <- cfly_cluster(ol, kmeans, 4, name="k4-3")

if (interactive()) {
ggobi(ol)
cfly_show(ol, "k4-1")
cfly_animate(ol, max = 5)
close(ol)
}