AVATAR: an improved solution for personalized TV based on semantic interface

The search engines in Internet and the recommender systems in the digital TV domain, pursue to light the burden of users with access to massive amounts of information, by offering only data (and TV programs) of interest for them. In this paper, we emphasize the advantages of using the so-called semantic Web technologies in the development of an intelligent TV assistant, named AVATAR. Its main advantage is a great improvement with respect to previous TV recommenders, obtained by combining two personalization strategies with a novel common nexus related to semantic inference capabilities. By the inference, AVATAR discovers appealing and complex semantic associations between the user preferences and the finally recommended TV shows. It is worth noting that this inference process overcomes a drawback identified in the existing tools, which suggest programs too similar to those the user watched in the past. In this regard, our inference strategy provides the viewers with suggestions clearly enhanced, diversified and permanently updated to their personal preferences.