That should be quite possible.  K-means is an approximate implementation of
an EM algorithm.


On 3/7/08 10:51 AM, "Isabel Drost (JIRA)" <[EMAIL PROTECTED]> wrote:

> As far as I know EM from clustering tasks it should be possible to port the
> algorithm to a Hadoop setting in a similar way as the k-means implementation,
> right?

Reply via email to