社交网络推荐系统Social Network-Based Recommender Systems

People increasingly use social networks to manage various aspects of their lives such as communication, collaboration, and information sharing. A user’s network of friends may offer a wide range of important benefits such as receiving online help and support and the ability to exploit professional opportunities. One of the most profound properties of social networks is their dynamic nature governed by people constantly joining and leaving the social networks. The circle of friends may frequently change when people establish friendship through social links or when their interest in a social relationship ends and the link is removed. This book introduces novel techniques and algorithms for social network-based recommender systems. Here, concepts such as link prediction using graph patterns, following recommendation based on user authority, strategic partner selection in collaborative systems, and network formation based on “social brokers” are presented. In this book, well-established graph models such as the notion of hubs and authorities provide the basis for authority-based recommendation and are systematically extended towards a unified Hyperlink Induced Topic Search (HITS) and personalized PageRank model. Detailed experiments using various real-world datasets and systematic evaluation of recommendation results proof the applicability of the presented concepts.