Abstract
Recommender systems help people in retrieving information that match their preferences by recommending products or services from a large number of candidates, and support people in making decisions in various contexts: what items to buy, which movie to watch or even who they can invite to their social network. They are especially useful in environments characterized by a vast amount of information, since they can e.ectively select a small subset of items that appear to fit the user's needs. We present the main points related to recommender systems using multimedia data, especially for social networks applications. We also describe, as an example, a framework developed at the University of Naples Federico II. It provides customized recommendations by originally combining intrinsic features of multimedia objects (low-level and semantic similarity), past behavior of individual users and overall behavior of the entire community of users, and eventually considering users' preferences and social interests.