Since iTunes has so many users, many of them decide to share their listening habits with Apple through iTunes Genius. In order to get personalized, automatic Genius playlists created, you have to share your listening data with Apple. All of this data is aggregated on Apple's servers to decipher similar songs based on listening habits.

That's why you have to update Genius whenever you add something to your library because Apple can't possibly store all of their listening data on your computer -- there's millions of songs out there! Only listening data relevant to you is kept on your computer when you update iTunes Genius (or after it updates once per week).

This same technology is what's used to recommend friends to you on Facebook, songs on Last.fm, software on Wakoopa, and on other sites that provide recommendations. This type of technology is called social discovery. If you're interested on learning more of the fine details of recommendation technologies like this, be sure to look at the slides from the Recked event -- an event all about recommendation technologies. The slides from that event go over the mathematics behind calculating interests.
Source(s):
http://www.recked.org/


Permalink | Report