News for hackers. Stuff that matters.
Re:Recommending items to more than a billion people by sjtrny on Wednesday June 03, 2015 @03:21AM (#9651062)
The technical name for this is "collaborative filtering". I think they are basing their work on this paper

- http://www.jmlr.org/papers/volume10/takacs09a/takacs09a.pdf

EDIT: Actually looks like Eq (15) from

- http://public.research.att.com/~volinsky/netflix/BellKorICDM...

Anyway there are lots of papers around on the topic.

Re:Recommending items to more than a billion people by Rifu on Wednesday June 03, 2015 @05:19AM (#9651412)
They do mention that in the 2nd paragraph and repeatedly throughout the article as CF. I found the article to be a nice primer on the topic as it listed common approaches to the problem.
Re:Recommending items to more than a billion people by thebigjc on Wednesday June 03, 2015 @10:24AM (#9652763)
Lots of research here, but the most interesting part is their distributed graph approach. Scaling up CF is a challenging problem, and the current open source implementations have problems at really large scales.

I hope they open source their Giraph implementation, or at least part of it.