IPTV recommender system using social network for multiple users

Abstract

With the advent of Internet Protocol television (IPTV), which combines broadcasting and telecommunications, TV is changing from the traditional broadcast-style into an interactive, two-way service that offers a vast range of content. As the volume of TV content increases, viewers face the difficulty of finding the right programming, and spend more time and effort to find content they like. This has accelerated the growth of recommendation systems that suggest content items the user is likely to find interesting. The present paper proposes a recommendation system that exploits user groups in online social networks to recommend IPTV content items to viewers. The proposed recommendation system manages a group of viewers with similar tastes rather than dealing with viewers individually. It automatically recognizes IPTV viewers within a specific range and categorizes them into groups that are, in turn, used to form a social network. The proposed system provides tailored recommendations for a group of viewers by analyzing the group's pattern of IPTV content consumption. The experimental results show that the proposed recommendation algorithm yields higher performance than the collaborative filtering approach. © 2013 International Information Institute.