Summarizing timelines based on content and social network

With the rapid growth of social network such as Twitter and SinaWeibo, more and more applications are designed for users to manage and serve for their social network platform. In this paper, we focus on the problem of summarization of timelines, which is useful for filtering out replicated posts and organizing posts in a more structured way. The content of short text is combined with social network for clustering posts. A corpus of Sina Weibo is annotated. Intensive experiments are conducted based on the corpus. We show that our method may achieve high precision and recall. The corpus is also shared for research community for further research.