improvements made to the recommendation engine made it easier to identify people through supposedly anonymous information.I guess the modeling community gets a +1 for great improvements.
I was a big fan of the Netflix Prize even if it only brought marginal improvements to the actual recommendation system. The shared knowledge and collaborative spirit was impressive. It doesn't sound like Netflix is going to go out on a limb and suggest a new contest. So it looks like this might be the end of the contest. This is very unfortunate because I was hoping this would spark a lot of companies trying these types of contests.
This also brings up a good point about making data anonymous. There are lots of ways to get this done. Please share ways that you're modeling for business, academic, or clientele that required making data anonymous.
No comments:
Post a Comment