Inside UC Berkeley's High Tech Joke Recommender

alphadogg writes "Every day is something like April Fools' Day at the University of California, Berkeley joke recommendation site, dubbed Jester. Now on Version 4.0, the site tosses visitors a handful of jokes to rate on a scale of "less funny" to "more funny." It then recommends jokes based on the user's taste (or lack thereof), dynamically making recommendations based on the user's most recent ratings. Jester's more than a joke jukebox though. Underlying it is a Berkeley-patented "collaborative filtering algorithm" dubbed Eigentaste , now on Version 5.0. The more people who use the system and rate jokes, the more data Berkeley researchers have to advance their understanding of recommendation systems, like those used by Amazon.com and other Web sites."

Comments

Link to the actual thing...

[adpsimpson]

For those who want to actually see it, not a blog about it - Jester