随着Twitter、新浪微博等社会网络的蓬勃发展,社会化短文本逐渐进入研究者的视野,在该方向上的研究成果正在逐年扩大.在分析社会化短文本摘要的意义之后,综述了社会化短文本的研究现状.社会化短文本摘要方面的研究成果主要有三类,面向web文档的自动摘要、面向事件的自动摘要、面向话题(或查询)的自动摘要.其中,面向话题(或查询)的社会化短文本自动摘要是对社会网络中海量短文本进行高效访问的重要手段.文中重点分析四种面向话题(或查询)的社会化短文本自动摘要的最新研究成果,包括:基于词频的方法、基于图的方法、基于合并的方法及基于优化的方法.最后指出社会化短文本自动摘要所面临的挑战和研究方向.
With the flourish of various kinds of social networks such as Twitter and Sina microblog, the socialized-short-text is gradu- ally entering into the study of the researchers. We precede the related works on socialized-short-text with the analysis of the signifi- cance of socialized-short-text summarization. We categorize the related works on socialized-short-text summarization into three groups, Web document oriented summarization, event oriented summarization, and topic/query oriented summarization. Topic/query oriented summarization is an effective approach to access information from vast amounts of text in social network. Hence, we put more attention on topic/query oriented socialized-short-text summarization and discuss four strategies proposed in the past two years, including "term frequency based strategy", "graph based strategy", "merging based strategy" and "optimizing based strategy". At the end of this paper, we propose some challenges and potential research directions.