Improving collaborative recommendation of coupons through digital TV by semantic inference of users' reputation

Recommender systems have proven to be an effective response to the information overload problem, by identifying items the users may be interested in. Trust and reputation are being increasingly incorporated in collaborative recommender systems in order to improve their accuracy and reliability, using network structures in which nodes represent users and edges represent trust statements. However, current approaches require the users to provide explicit data (about which other users they trust or not) to form such networks. In this paper, we apply a semantic approach to automatically build implicit trust networks and, thereby, improve the recommendation results transparently to the users. Even though our approach is not limited to any specific domain, we illustrate it within the recommendation of promotional coupons through Digital TV, which can be accessed from domestic and mobile consumer devices.