面向综合集成研讨环境的主动信息获取方法  

Active information retrieval method for CWME

作　　者：倪娜[1] 刘凯[1] 李耀东[1] 

机构地区：[1]中国科学院自动化研究所复杂系统与智能科学重点实验室,北京100190

出　　处：《计算机应用研究》2010年第11期4058-4062,4066,共6页Application Research of Computers

基　　金：国家"973"重点基础研究发展计划资助项目(2007CB311007)

摘　　要：针对在综合集成研讨环境中,由于存在时间压力,传统的网络信息获取方法难以直接使用,提出了一种面向综合集成研讨环境的主动信息获取方法。该方法将领域词条与通用词条相结合,从发言文本流中实时提取话题,并在话题发生变化时自动生成检索词送入搜索引擎进行检索,再通过多个用户之间的协作推荐实现对重要检索结果的筛选。实验结果表明,这种方法可为综合集成研讨系统的用户提供及时、准确、上下文相关的信息服务。In CWME,the severe time pressure makes it impossible for participants to retrieve topic-related information timely and sufficiently from the Internet on traditional approaches.To meet this challenge,this paper proposed an active information retrieval method,which could extract topics online and judge change of topics during the discussion process.Once sensed a new topic,query terms generated from this new topic would be sent to a WWW search engine automatically.Identified the valuable search results from the WWW search engine based on collaborative filtering and then sent to the participants actively.Experimental results show that the proposed method is effective in CWME,and can provide timely,accurate and topic-related information for participants well.

关 键 词：综合集成研讨厅 领域词典 话题提取 协作过滤 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]