Implicit Feedback through User-system Interactions for Defining User Models in Personalized Search

Personalized search is aimed at tailoring the search outcome to the user context. A user profile represents the user interests and preferences that can be captured either explicitly or implicitly. Since a long time it has been advocated that Information Retrieval is an interactive activity, and various types of user-system interactions are more and more considered and used to improve search. In particular implicit feedback techniques are applied to the purpose of collecting user interests and preferences via user-system interactions. In this paper a short synthesis of the implicit techniques used to define user models is presented.