中文摘要：

报道关系识别是话题识别与跟踪TDT（Topic Detection and Tracking）研究内容中的基本任务之一,根据新闻话题的几大要素：时间、地点、人物、内容等,提出了一种基于话题要素的话题报道表示模型,并给出了基于话题要素相似度计算的报道关系识别方法。实验证明这种方法特别适用于同主题下不同话题的报道关系识别。

英文摘要：

Story link detection（SLD） is one of the five basic tasks in topic detection and tracking（TDT）.According to several major elements of news topic：when,where,who and what,this paper presents a topical story representation model based on topic elements,and then proposes an approach for story link detection based on topic elements similarity computation.This method is proved by the experiments reported in the paper to be particularly effective when dealing with SLD with different topics in same subject.