异质工程文档多维关联的语义检索方法  

Approach for multi-dimensional associated heterogeneous engineering document semantic retrieval

作　　者：姚原岗[1] 林兰芬[1] 董金祥[1] 

机构地区：[1]浙江大学计算机科学与技术学院,浙江杭州310027

出　　处：《浙江大学学报（工学版）》2011年第2期267-272,共6页Journal of Zhejiang University：Engineering Science

基　　金：国家"863"高技术研究发展计划资助项目(2009AA04Z151);浙江省自然科学基金资助项目(Z1090461)

摘　　要：针对现有工程信息检索中文档语义理解和关联不足的问题,根据工程环境下信息检索的特点,提出一个异质工程文档的检索方法,包括内容分析、语义建模、多维关联、语义推理与扩展以及查询处理等过程.采用基于本体的理解标注多源异质工程文档,获取工程语义信息,建立语义标注库,实现异质文档内容的统一表达.采用以文档为中心的多维关联机制,在文档内容语义理解的基础上,将基于本体的内在关联扩展为融合文档及内容的多重关联.结果表明,多维关联机制能够启发式地关联产品生命周期中的文档,增强查询的导航能力,为语义推理和检索扩展提供服务.Aimed at the problem of insufficiency of document semantic comprehension and association in current engineering information retrieval,according to the analysis of information retrieval characteristics in engineering environment,a heterogeneous engineering document information retrieval method was proposed,which included content analysis,semantic modeling,multi-dimensional association,semantic inference,expansion,and query process.Ontology based comprehension was utilized to capture engineering semantic and build semantic annotation base,which realized unified description of heterogeneous document content.Multi-dimensional association architecture was introduced focusing on documents.On the base of semantic comprehension of document content,ontology based internal association was expanded to multiple associations.The result shows that multi-dimensional association can associate documents in product life cycle heuristically,improve query navigation,and support semantic inference and retrieval expansion.

关 键 词：语义 异质文档 信息检索 多维关联 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]