Memory-Based Collaborative Filtering Works Best

This infovis article on Collaborative Filtering provides a good summary of various approaches to ACF and concludes that :

"Memory Based algorithms outperform the others in yield and quality of preference prediction. It turns out, on the other hand, that they are the most conceptually simple and the easiest to implement, too."

Effectively, if you have enough data about people's preferences, you can predict someone's preference for an object they haven't rated from how the object is rated by other people who have similar preference profiles.

People who like what you like (your "taste buddies" in Walter Logeman's words), also like this.