新话题监测研究进展  

Review of Study on New Topic Detection

作　　者：饶洋辉[1] 叶良[2] 常红旭[2] 程洁[2] 

机构地区：[1]中国科学院文献情报中心,硕士研究生100190 [2]中国科学院计算机网络信息中心,100190

出　　处：《图书馆杂志》2009年第7期60-63,共4页Library Journal

基　　金：中国科学院规划与战略研究项目“21世纪科技发展前沿走势研究”,项目编号:KACX1-YW-0733

摘　　要：新话题监测是话题检测与跟踪领域的一项重要研究。对于事件层的新话题监测,主要阐述基于聚类的方法,包括文档的向量化表示以及文档聚类。对于句子层的新话题监测,重点介绍基于问答系统的方法,包括问题表述与预处理、相关句子检索以及新颖句子检测。并对新话题监测研究存在的问题及其发展趋势进行了探讨。New topic detection is an important research in the field of topic detection and tracking. For new topic detection at the event level, method based on clustering was discussed mainly, which included the representation of documents, and document clustering. For new topic detection at the sentence level, method based on question answering was presented mainly, which included question formulation and processing, relevant sentence retrieval, and new answer detection. Finally, problems and trends of new topic detection were discussed.

关 键 词：新话题监测 聚类 问答系统 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]