Social streams recommendation in sina microblog with relation of user and interest

Social websites like Facebook and Twitter have been growing rapidly in terms of user and content. Users are faced with hundreds of social streams including those they are not interested in. Many studies have focused on recommending topics or discovering news and events to alleviate information overload. In this paper, we have developed a system to recommend Top-N social streams to individual user from his / her real-time social streams in Sina Microblog, the largest microblog system in China. We develop a novel model to define Tie-Strength between two users. We also introduce a new concept named Related-Interest which measures A´s interest in B more accurately. We have implemented a content-based recommender system with several algorithms based on these ideas mentioned above. These algorithms are tested by real world users we recruited from onsite application of Sina. Our algorithms outperform a baseline of chronological order and a related study on Twitter.