Online social networks can be used as networks of human sensors to detect important events. It is important to detect important events as early as possible. Microblogs provide a new communication and information sharing platform for people to report daily-life events, and express their views on various issues. Because of the quickness of microblogs, microblog data can be used to predict popular topics. In this paper, we propose a short-term prediction model of topic popularity. With data from Sina Weibo, the most popular microblog service in China, we test our algorithm and our data shows that the proposed model could give a short-term prediction on topic popularity.
展开▼
