On Jan 1, 2010, at 10:45 AM, Bogdan Vatkov wrote: > I do not know how this updatable clustering works (using previous results as > centroids for new clusterings), is there an example I could see in action? > Additionally I would like to see an example of how could one combine Canopy > and k-means, I just saw this described in theory somehow but could not find > an example of it.
You should just be able to run Canopy and then use the output of that as the value for --clusters on the KMeansDriver.
