Hi All,

I have map-reduce version of SOM, but it's not yet in the form of
K-means. Shall I go ahead and attach the patch, or should I work on
getting it in line with K-means first and then send the patch.

--Thanks and Regards
Vaijanath N. Rao 

-----Original Message-----
From: Ted Dunning [mailto:[email protected]] 
Sent: Wednesday, March 03, 2010 6:37 AM
To: [email protected]
Subject: Re: SOM implementation

No.

But we would be happy to help you write one!

In particular, it is likely that the k-means implementation that we
currently have would provide a very good basis for build a SOM
implementation.  There is also a large scale SVD implementation in
Mahout that would be useful for deriving initial conditions.

SOM should have similar properties as k-means in terms of scaling and
thus should be a good match to iterative map-reduce implementation.

Would you like to try your hand on the problem?

On Tue, Mar 2, 2010 at 4:54 PM, hifsa kazmi <[email protected]>
wrote:

> Does mahout provides an implementation for Self Organizing Maps??

Reply via email to