面向语义的Web信息查询研究及应用-智能与分布计算实验室

面向语义的Web信息查询研究及应用

姓名	王渊
论文答辩日期	2005.05.11
论文提交日期	2005.05.14
论文级别	硕士
中文题名	面向语义的Web信息查询研究及应用
英文题名	Research on Semantic Oriented Web Information Query and its Application
导师1	卢正鼎
导师2
中文关键词	语义网;本体;Web信息查询
英文关键词	Semantic Web;Ontology;Web information query
中文文摘	Web作为全球性的网络信息系统，极大促进了Internet的可应用性和在其之上的信息传播。然而伴随着Web成功的同时，呈指数级增长的海量信息使来自各领域的用户对信息的查找、访问、表示以及维护变得更加困难。对于目前Web上的信息，计算机只能从格式上来处理和验证，并不能处理知识级别的问题。而不同团体对同一领域事物的认识和表示往往不同,使得来自不同数据源关于同一领域的数据描述存在着冲突。这种认识上的差异所产生的数据描述差异可被称作语义异构。为了最大程度地利用己有知识成果，从这些不同的数据源中有效获取信息，面向语义的Web信息查询研究就变得非常有意义。在对RDF/RDFS,DAML,OIL等Ontology语言相关规范进行细致研究之后,设计并实现了一个基于关系的搜索引擎OntoLook。该系统就是希望通过找出概念之间的关系，将关键字与关系进行组合查询，获得更加贴切用户需求的页面集。因为语义Web在基础架构一级提供了记录资源之间关系的能力，因此OntoLook通过在网页中加入符合RDF语法的标注信息并配之以相应的OWL文档作为解释来表达网页中所包含的机器可理解的语义信息。OntoLook提供了友好的界面以便于用户输入关键字并初略指定关键字的类别。由于用户提交给搜索引擎的关键字之间都是有关系的，因此OntoLook在处理用户提交的关键字组合时，首先通过本体找出用户指定概念的父类概念和子类概念，然后找出这些概念之间的所有关系，进而将关键字与各种关系进行组合后再提交给数据库。通过实验证明，该系统能够获得更加贴切用户需求的页面集。
英文文摘	World Wide Web,as a global web information system，accelerate the speed of information transmission with the development of www, when users want to find or interview the appropriate information, more and more information put users to inconvenience. The information on Web at present, the computer can only come up from the form and deal with and prove , can not deal with the problem of the rank of knowledge. After studying how to apply Ontology-based methods, which can solve problems caused by semantic diversity in the Web environment, we design and realize a Web intelligent search system based semantic web named OntoLook. It uses OWL as the Ontology language, and also gives powerful methods for ontology information repository storage. It solves problems caused by traditional key-word-based search engines, and it can provide intelligent search services. After attentive research on the specifications of RDF/RDFS, DAML, OIL, OWL, etc,we design and realize a Web intelligent search engine based on semantic web called OntoLook. Because there always has some relations between the key words user input, if we can find out the relations between these key words, combine key words with correspond relations and retrieve by them, the result web pages will be more appropriate than traditional key-word-based search engines. When processing the key words users input, OntoLook first finds out the father concept and subdivision concept of the concept users specified, then select the possible relations between the concepts by ontology. Last, send key word and various kinds of relations to the database after being made up. Through the experiment, this system can obtain the page collection which are more appropriate to user's needs .