Symmetric Collaborative Filtering Using the Noisy Sensor Model

Abstract

Collaborative filtering is the process of making recommendations regarding the potential preference of a user, for example shopping on the Internet, based on the preference ratings of the user and a number of other users for various items. This paper considers collaborative filtering based on explicit multivalued ratings. To evaluate the algorithms, we consider only pure collaborative filtering, using ratings exclusively, and no other information about the people or items. Our approach is to predict a user's preferences regarding a particular item by using other people who rated that item and other items rated by the user as noisy sensors. The noisy sensor model uses Bayes' theorem to compute the probability distribution for the user's rating of a new item. We give two variant models: in one, we learn a classical normal linear regression model of how users rate items; in another, we assume different users rate items the same, but the accuracy of the sensors needs to be learned. We compare these variant models with state-of-the-art techniques and show how they are significantly better, whether a user has rated only two items or many. We report empirical results using the EachMovie database of movie ratings. We also show that by considering items similarity along with the users similarity, the accuracy of the prediction increases.