An Automatic Tag Recommendation Algorithm for Micro-blogging Users

Online social networks such as Sina Weibo micro-blogging website allow user to annotate himself (or herself) using tags, which describe the characteristics of the user. Many researches have been done on photos, films, commodities but rare researches on users. In this paper, we try to recommend tags for users who can communicate directly with other users. Our method is based on interactive relations between users and is low cost. We present and evaluate tag recommendation method to support the user self-annotation task by recommending a set of tags for users. Experiment evaluations on real and large Sina Weibo dataset which contains more than 140 million users and distributed processing framework Hadoop shows that our method can effectively recommend relevant tags to users.