From Collaborative Filtering to Implicit Culture: a general agent-based framework
Enrico Blanzieri, Paolo Giorgini · 2004
Collaborative Filtering bases its effectiveness as a recommender system on ratings about a set of items provided by a set of users. In our perspective, an agent behaves as a member of a group would do (the agent implicitly belongs to the same "culture" of the group) without extra-effort or direct interaction. In this paper, we introduce the concept of Implicit Culture and propose a general architecture for Systems for Implicit Culture Support. We show how Collaborative Filtering can be considered as an instance of our architecture, and finally, we consider the related work. 1. INTRODUCTION Given the problem of information overload, the building of recommender systems is a mayor issue. Collaborative Filtering (see [4] for a recent reference) demonstrated to be an effective approach from an applicative point of view. However, the ideas underlying Collaborative Filtering have a greater scope than filtering itself. In this paper we capture those ideas in the notion of Implicit Culture an...