SINA Weibo, a micro-blogging service, provides users with an application to record their brief postings about their lives. They can tag themselves using free tags to show their personal characteristics, but 78.2% of all users do not tag themselves. In this paper, we try to mine user's personal interests based on the self-defined free tags. A directed weighted graph is constructed with the interactive relations between users. We suppose that if two users have interacted with each other, they may share latent common interests. So interests can be propagated from a user to its interacted friends. Experiments on three SINA Weibo datasets show that our method performs better than exiting methods in mining user's personal interests. Moreover, our method is more efficient than these methods since we do not use the content of user's tweets but the user self-defined free tags only.
展开▼
