Trust Networks for Recommender Systems

Recommendation Systems are gaining tremendous prominence in the digital society, and are fast becoming the bastions of electronic commerce. “Will I like this book?”, “Is this a movie I can see with my kids?’, ‘Which hotel will suit me the best?”: we increasingly rely on the social aspects of the world wide web (WWW) to help us navigate through such questions in our everyday life. We are quick to judge, and even quicker to just imitate our friends and do what they did. The magic potion that casts a spell on us to imitate and even at times make irrational decisions is trust. Trust enhanced recommender systems are designed to help us to form an opinion on matters that are not entirely known to us, or evennot known at all.The social web allows people to express their views to other users of the system. We call the resulting network a social network. There are social networks in which the users can explicitly express their opinion as trust and distrust statements. We refer to these kinds of social networks as trust networks.The particular focus of this book, infusion of the theory of how online trust networks can be modeled and the utility of these models to enhance the quality of recommendations generated in the online recommendation systems arena is not only groundbreaking and innovative; it is likely to be the central pivot for the next generation of research in social network analysis. Think of any system where humans need subjective judgments from their peers and seers. As you start to read the book, it will be quickly evident that issues explored in this book are the backbone of any such system. Some of these broad issues are: who you know, who you don’t know, who you trust, why you trust them, how does this trust translate, aggregate, and propagate; and how to ef?ciently and correctly identify key trust ?gures to enhance the quality of generated recommendations.