mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Zhengguo 'Mike' SUN <zhengguo...@yahoo.com>
Subject Re: Fwd: A MapReduce Algorithm for Matrix Multiplication
Date Wed, 09 Dec 2009 03:36:59 GMT
Thanks for your pointers, Jake. 

I am still working on the implementation. Certainly I would like to contribute after I finish
my implementation.



________________________________
From: Jake Mannix <jake.mannix@gmail.com>
To: mahout-user@lucene.apache.org
Cc: j-norstad@northwestern.edu
Sent: Tue, December 8, 2009 10:16:31 PM
Subject: Re: Fwd: A MapReduce Algorithm for Matrix Multiplication

On Tue, Dec 8, 2009 at 6:49 PM, Zhengguo 'Mike' SUN
<zhengguosun@yahoo.com>wrote:

> I appreciate if you could provide some pointers.
>

Of course, we'd also appreciate it if you could contribute that code here to
Mahout, as well, if
possible.  It'd be great to have that here for everyone to use, as I'm sure
that having the choice of
doing NMF as well as SVD or LSI or pLSI would very helpful for the rest of
the community!

Patches welcome!  Even if your first impls don't do any of the tricks we're
describing, if you
contribute it, we can help you improve it.

  -jake


>
>
> ________________________________
> From: Jake Mannix <jake.mannix@gmail.com>
> To: mahout-user@lucene.apache.org
> Cc: j-norstad@northwestern.edu
> Sent: Tue, December 8, 2009 9:08:19 PM
> Subject: Re: Fwd: A MapReduce Algorithm for Matrix Multiplication
>
> On Tue, Dec 8, 2009 at 5:56 PM, Ted Dunning <ted.dunning@gmail.com> wrote:
>
> > NMF should be amenable to the stochastic decomposition trick.  That
> reduces
> > the problem to a much smaller factorization that you could plausibly do
> > using sequential techniques.  Jake Mannix is working on getting
> > factorizations going, but I don't know if he has gotten to NMF.
> >
>
> I'm not currently working on NMF, but the stochastic decomposition trick
> will be
> in there soon, which should allow all this pretty easily.
>
> Although... if you start with a positive matrix, you may want a specialized
> random
> projector which preserves positivity for this kind of thing.  But I'm not
> sure, I
> haven't looked too closely at what happens when you try to do this trick on
> NMF.
>
>  -jake
>
>
>
>
>



      
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message