A Novel Approach for Generating Personalized Mention List on Micro-Blogging System

Online social networks provide us a convenient way to access information, which in turn bring the information overload problem. Most of the previous works focused on analyzing user's retweet behavior on the micro-blogging system, and diverse recommendation algorithms were proposed to push personalized tweet list to users. In this paper, we aim to solve the overload problem in the mention list. We firstly explore the in-depth differences between mention and retweet behaviors, and find the users' various actions for a piece of mention. Then we propose a personalized ranking model with consideration on multi-dimensional relations among users and mention tweets to generate the personalized mention list. The experiment results on a micro-blogging system data set show that the proposed method performs better than benchmark methods.