Here’s an in-depth look at the type of technology that powers Amazon, Pandora, and MANY more services.
Two people who have been working on these systems for years, Joseph Konstan and John Riedl who were part of the GroupLens team at the University of Minnesota have written a great overview article about recommendation systems in the October issue of IEEE Spectrum.
From the Article:
Over the years, the developers of recommender systems have tried a variety of approaches to gather and parse all that data. These days, they’ve mostly settled on what is called the personalized collaborative recommender. That type of recommender is at the heart of Amazon, Netflix, Facebook’s friend suggestions, and Last.fm, a popular music website based in the United Kingdom. They’re “personalized” because they track each user’s behavior—pages viewed, purchases, and
ratings—to come up with recommendations; they aren’t bringing up canned sets of suggestions. And they’re “collaborative” because they treat two items as being related based on the fact that lots of other customers have purchased or stated a preference for those items, rather than by analyzing sets of product features or keywords.
Personalized collaborative recommenders, in some form or another, have been around since at least 1992. In addition to the GroupLens project, another early recommender was MIT’s Ringo, which took lists of albums from users and suggested other music they might like.
The full text article is available here.
A second IEEE spectrum article might also be of interest. “A Recommendation Engine for Politics”
offers a look at ElectNext that you can take look at online.
Direct to ElectNext