Why do you want to use the EM `algorithm'?
It will be easier to directly maximize the log likelihood, and that is a
5-mins job in R.
Take a look at the examples in MASS4 ch 16 (and see the books in the FAQ
or the scripts in the MASS package).
On Sun, 16 Nov 2003, Saint-Jean Christophe wrote:
> I would like to fit a von mises mixture to my data using the EM
> Algorithm.
> Unfornately, existing packages (including circstats) doesn't
provide
> such method.
> So, does anybody have written such code ?
> Big thanks,
> Christophe, (R newbie)
--
Brian D. Ripley, ripley at stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595