SIGIR09: The Wisdom of the Few: A Collaborative Filtering Approach Based on Expert Opinions from the Web
Xavier Amatriain of Telefonica research presented work on collaborative filtering. Usually you do collaborative filtering by finding the other users “similar” to your subject and combining their recommendations. This paper argued/demonstrated that sometimes you are better off figuring out who the experts art and only paying attention to their opinions. You might just create non-personalized recommendations from them, or you might personalize by finding the best _experts_ to recommend for a user. The experimented by exploring movie recommendation using the Netflix challenge mass ratings versus using the (expert) critics’ recommendations on Rotten Tomatoes. They found expert recommendations often worked better.
I asked about some past work on e.g. semisupervised learning suggests various approaches to combining small amounts of high-quality data (experts) with large amounts of messier data (mass user ratings). It suggests, for example, some sort of weighted combination of expert and mass user opinion. They know this could help a lot, but don’t have a general approach to separating the export from everyone else in a large mass of recommendations (Rotten Tomatoes did it for them).