On Mon, Mar 10, 2008 at 2:52 AM, Ted Dunning <[EMAIL PROTECTED]> wrote:

> Sorry.  The first year prize is gone (the $50,000 progress prize for the
> first yet) and the subsequent yearly prizes are going to be very hard to
> get
> (squeezing water from a stone).  The asymptotic progress curve tells the
> story.


+1

There is still room for improvement---the algorithms that were applied in
the first progress prize weren't particularly fancy.  But, good luck running
fancy algorithms on a data set of that size.  And, it appears that no single
algorithm will get you to the top---as with the progress prize winners,
you'll likely need to combine 100+ algorithm outputs.  I'm still anxious to
try MMMF on the data, but the Netflix data is 100x larger than the data sets
I've run MMMF on.  Matlab code I have just won't cut it...

Jason

-- 
Jason Rennie
Head of Machine Learning Technologies, StyleFeeder
http://www.stylefeeder.com/
Samantha's blog & pictures: http://samanthalyrarennie.blogspot.com/

Reply via email to