X

Netflix's casting call for better recommendations

CBSNews
2 min read

Algorithms aren't exactly the stuff of Hollywood legend, but the right one could point you to the movies you really want to watch, and away from the clunkers.

Netflix contest

That's the premise of the contest that DVD rental star Netflix unveiled Monday. CEO Reed Hastings calls the company's current ratings system for movies "the Model T version of what is possible"--some bloggers said much the same, albeit less graciously--and he's issuing a call for someone to help him do better. Whoever figures out how to improve the Web-based software by 10 percent will win $1 million. (Lesser rewards will be given annually in the meantime.)

Netflix subscribers rate the films they watch, and the company's recommendation engine then tries to steer them to other movies particularly suited to them. The service currently comes to within one star (out of five), the company says, of accurately predicting what a consumer would really feel about, say, "X-Men: The Last Stand."

No, you don't have to figure it all out on your own. To get you started toward building that better algorithm, Netflix is ready to share with you a data set in the form of 100 million of its customers' movie ratings, duly anonymized.

Blog community response:

"Netflix currently lets users view the movies selected by friends, but as a poster child for moving services to the web I'd love to see them take up the most current social technologies. Tagging allows for exploration by theme in a much more flexible way than formal site created categories. It would be good if instead of asking for improvements to its black box recommendation engine Netflix began by making movie discovery more transparently social."
--TechCrunch

"I'm surprised that the ratings database is a download rather than an API. I'd bet that if they made it an API, the $1 million prize wouldn't be necessary at all."
--O'Reilly Radar

"If you are thinking of entering the contest, you might be interested to know that much of the Internet Movie Database (IMDb) database is available for download. Another good source for movie content is Amazon Web Services."
--Geeking with Greg