A logical framework for fuzzy collaborative filtering

Systems which predict the items a user would like, on the basis of the preferences given by other users, are attracting growing attention as automated recommender services on the Internet. Collaborative filtering techniques are widely used to implement such recommender systems. We show how fuzzy sets and many-valued logics can be fruitfully applied in the description and design of collaborative filtering methods.