Netflix Progress Prize Awarded

Re:Netflix Progress Prize Awarded by FiReaNG3L on Wednesday December 10, 2008 @06:48PM (#393657)

As per the prize rule, they are under the obligation to describe and publish their mode (as they did last time). I really wish someone would release an open source recommender package, coded in C, with PHP bindings, based on the latest / greatest algorithms, or any algorithm (just plain SVD if anything).

Interesting links (from the winning team(s) websites about their latest research) : http://public.research.att.com/~volinsky/netflix/kdd08koren.... http://www.commendo.at/index.php?lang=0&_0=2&_1=3

[ Parent | on HN ]

Re:Netflix Progress Prize Awarded by icefox on Wednesday December 10, 2008 @07:09PM (#393680)

Well I am not going to do everything for you, but here is something to start with:

http://benjamin-meyer.blogspot.com/2006/10/netflix-prize-con...

It is a little framework I put together to play with the netflix prize framework. I implemented SVD as one of the examples.

On a side note I did think about the idea of putting together a small company that would sell recommendation systems to companies for their website. The value would not be in the algorithm, but in the ability to integrate with their software.

[ Parent | on HN ]

Re:Netflix Progress Prize Awarded by FiReaNG3L on Wednesday December 10, 2008 @07:14PM (#393689)

Very nice, I saw many Netflix Prize derived frameworks but I somehow missed this one.

Did you test RMSE of your SVD on Netflix Dataset?

[ Parent | on HN ]

Re:Netflix Progress Prize Awarded by icefox on Wednesday December 10, 2008 @09:00PM (#393808)

Nope, it is just a stock SVD, it wont win any awards, just get the job done and be an example. There are plenty of people who have tweaked SVD up and down so I didn't spent time doing it too.

[ Parent | on HN ]