MicroBlog recommendation based on user interaction

Microblog is a broadcast social network platform which uses "following" mechanism to share brief real-time information. Due to its simpleness and convenience, more and more people use microblog to communicate with each other. However, with the increasing number of users, there generates lots of data, which brings a challenge for users to obtain information they are interested in. Moreover, the reverse chronological order property and the homepage page limit also increase the difficulty of information retrieval. Therefore, the information publishers should recommend users based on the users\´ personal preference. In order to match user demands, we present a method using "user influence" to solve the recommendation problem. In this paper, we propose an algorithm, an extension of PageRank, which considers that the "user influence" is related to the link structure between user pages and the user interaction in mircoblog. We conduct an experiment using a public Sina search dataset and Kendall\´s r correlation analysis, and conclude that our algorithm performs reasonably.