No. And buckshot is probably not particularly scalable. Instead, we have an implementation of k-means++ that is nearly ready to release. That should provide comparable (or better) benefits compared to buckshot.
On Mon, Mar 1, 2010 at 3:18 AM, Arshad Khan <[email protected]> wrote: > I'm sorry if this questions has been asked before but does Mahout provides > implementation of Buckshot algorithm to improve KMeans performance? > -- Ted Dunning, CTO DeepDyve
