Ranking and Combining Social Network Data for Web Personalization

Various Web-based social network data reflect user interests from multiple perspectives in a distributed environment. They need to be integrated for better user modelling and personalized services. We argue that in different scenarios, different social networks play different roles and their degrees of importance are not equivalent. Hence, ranking strategies among different social network data sources are needed. In addition, combining different social network data can produce interesting subsets of these data with different levels of importance. In this paper, we propose social network data ranking and composition strategies, we validate the proposed methods by collaboration network data (Semantic Web Dog Food) and micro-blogging data (from Twitter), then we use the ranked and composed results for developing a Web-based personalized academic visit recommendation system to show their potential effectiveness.