摘要
建立了一种表达用户信息需求的界模型,通过计算信息需求的上边界、下边界,确定信息需求的需求域.该模型从语言学的角度出发,通过挖掘语言的内涵和外延来表达语言的含义.用户的查询请求有其内涵和外延,信息需求域正是通过表达查询请求的内涵和外延来反映用户的真实信息需求.在信息检索时,计算文档与信息需求域的相似度,并根据相似度对文档进行排序.在实际检索时,采用用户真实相关反馈机制,该相关反馈机制得到的是真相关文档,因而可以更好地表达用户的信息需求.使用Lemur工具进行的对比分析实验表明,界模型下的真相关反馈信息检索具有较理想的效果.
An expression of user information need called bound model is introduced.The model determines the information need domain through calculating the upper bound and the lower bound of information need.From the linguistic point of view,the model expresses the connotation and extension of the meaning of the language.User＇s query request has its connotation and extension.The information domain uses connotation and extension to reflect the user＇s actual information needs.In information retrieval（IR）,similarity between document and the information need domain is calculated,and documents are sorted according to the similarity.The model uses user＇s relevance feedback techniques which can obtain real relevance documents,so it can better express user＇s information need.Experiments done using Lemur tool show that the bound model IR with real feedback relevance techniques has good retrieval results.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2010年第A02期301-306,共6页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金资助项目(60865003)
关键词
信息需求域
下边界
上边界
界模型
相关反馈
information need domain
lower bound
upper bound
bound model
relevance feedback
