A Personalized Microblog Search Model Considering User-Author Relationship

Although users can obtain real-time information from Microblog, it is more and more difficult to retrieve contents to satisfy their personalized needs because of the massive, irrelevant and repeated information. Different from webpage-oriented search, Microblog contents are much shorter and users' information selection behavior is influenced by comprehensive factors such as users' preference, Microblog contents, author characteristics, and the relationship between user and author. This paper proposes a personalized Microblog search model by integrating the comprehensive factors. The proposed model combines theme model and language model to analyze users' preference, and integrates the relationship between user and Microblog author to analyze the user-author dimension preference. The search feedback experiment shows that the proposed model characterized by user-author relationship can improve the effect of Microblog search.