Topic-STG: extending the session-based temporal graph approach for personalized tweet recommendation

Jianjun Yu, Yi Shen, Zhenglu Yang

Abstract: Micro-blogging is experiencing fantastic success in the worldwide. However, during its rapid development, it has encountered the problem of information overload, which has troubled many users. In this paper, we mainly focus on the task of tweet recommendation to address this problem. We extend the session-based temporal graph (STG) approach as Topic-STG for tweet recommendation which comprehensively considers three types of features in Twitter: the textual information, the time factor, and the users' behavior. The experimental results conducted on a real dataset demonstrate the effectiveness of our approach.