计算机应用与软件Issue(10)：21-25,112,6.DOI:10.3969/j.issn.1000-386x.2015.10.005
基于维基知识的微博事件追踪方法
MICROBLOGGING EVENT TRACKING METHOD BASED ON WIKIPEDIA KNOWLEDGE
摘要
Abstract
The length of microblogging text is short and the cyber-words emerge constantly.Traditional methods usually ignore the semantic information of the text while tracking microblogging events,which often results in unsatisfactory results.To solve this problem,we proposed a Wikipedia knowledge-based method for tracking microblogging event.First,we defined the strong correlation entities and alias entities of Wikipedia entries respectively,and represented the Wikipedia knowledge as quintuple forms.Then,we constructed a map from word space to Wikipedia entry space to indicate the initial event vector and subsequent microblogging text vector.Finally,we calculated the semantic simi-larities between microblogging text vectors and the initial event vector,and made judgement according to the pre-set threshold to complete event tracking.Experimental results show that the proposed method can make full use of semantic information of the Wikipedia knowledge and effectively improve the performance of microblogging events tracking compared with traditional methods.关键词
维基知识/微博/短文本/事件追踪/语义扩展Key words
Wikipedia knowledge/Microblogging/Short text/Event tracking/Semantic extension分类
信息技术与安全科学引用本文复制引用
唐浩浩,席耀一,周杰,郭志刚,陈刚..基于维基知识的微博事件追踪方法[J].计算机应用与软件,2015,(10):21-25,112,6.基金项目
国家高技术研究发展计划项目（2011A A7032030D）；全军军事研究生课题 （）
