Re: [R] Empty clusters in k-means - possible solution

[Luca Nanetti]

Dear Simon,

I respectfully disagree with you, when you say that it is NOT a bug. Empty
clusters can only happen if one provides a set of centers; if one enters
something like k=4, it cannot happen, so it must be a bug. Which one is
your case?
>From the help:

"Except for the Lloyd-Forgy method, k clusters will always be returned if a
number is specified. If an initial matrix of centres is supplied, it is
possible that no point will be closest to one or more centres, which is
currently an error for the Hartigan-Wong method"

This having been said, I suspect that there is a bug in the current
implementation of kmeans. I get the "empty cluster: try ..." answer even
when I enter a number of clusters, but only with R 2.15. I never had it
with R 2.14. Didn't try with R 3, yet.
I posted this problem on the list twice, at no avail (unless I forgot to
find the answer in the mailing list, of course).

Kind regards,
Luca

> Hello all,

> k-means algorithms can at times fail because one of the cluster become
> emmpty. In this case, the kmeans R function returns:
>"empty cluster: try a better set of initial centers"

> This has been discussed several times on several R-lists, and is NOT a
> bug, but can be annoying when using k-means in complex simulation where
> this error brings everything to a stop. One can use try() or tryCatch()
> to avoid this, but this is just a programming trick.

--
______________

Luca Nanetti, MSc, MRI
University Medical Center Groningen
Neuroimaging Center Groningen
Groningen, The Netherlands
Tel: +31 50 363 4733

[[alternative HTML version deleted]]

______________________________________________
R-h...@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

[Ranjan Maitra]

There is a documented bug in R 2.15.2 and R 2.15.3. The developers
decided not to fix this with an update, but it has been fixed with R
3.0.0 (perhaps also with R 3.0.1, which I have not tried myself). So,
upgrading R should take care of this problem.

HTH,
Ranjan

--
Important Notice: This mailbox is ignored: e-mails are set to be
deleted on receipt. For those needing to send personal or professional
e-mail, please use appropriate addresses.

____________________________________________________________
FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!