【作者】 徐芳；

【导师】 赵恒永；

【作者基本信息】 北京化工大学 ， 计算机应用技术， 2008， 硕士

【摘要】 搜索引擎技术以一定的策略在互联网中搜集、发现信息,对信息进行理解、提取、组织和处理,并为用户提供检索服务。目前的搜索引擎大多采用关键字匹配的方式,只要发现含有这个关键字,就将该文档或网页作为查询结果返回给用户。由于参与匹配的是字符的外形,而不是它们所表达的概念,因而经常出现检索不全,答非所问的情况。于是,需要采取一定的策略提高搜索引擎的知识处理能力和理解能力,这已经成为搜索技术未来发展的趋势。概念检索就是其中一种实现方式。概念检索是把信息检索从当前基于关键词检索的层面提高到基于知识(概念)检索的层面,从词所表达的内在涵义的层面上来认识和处理用户的检索请求。本文研究了实现概念检索的关键技术一知识库技术,并研究知识库的建立、表示和利用这三个方面的问题,选取了本体技术作为解决问题的方法。本文将本体技术、Jena推理工具和Lucene全文搜索技术相结合,并对化工专业词汇特点进行深入研究,确定词汇的层次关系结构,定义其中的类和类的属性,以及类与类之间的关系,采用本体开发工具Protégé对化工专业词汇的语义信息进行表示,使用OWL作为本体的描述语言,使用Jena工具针对所建立的本体自定义推理规则,对本体进行解析和推理查询,利用Lucene作为搜索引擎内核进行具体的索引检索,从而基本实现了概念检索主要的两个功能:同义检索和相关扩展检索。本文利用实验室搜索引擎系统Spider模块,从化工专业网站上抓取网页,并利用网页转换的文本文件进行实验。通过实验数据证实,基于概念的搜索引擎提高了查全率,使搜索引擎智能化。更多还原

【Abstract】 Search engine technology to a certain strategy collects and finds information in Internet. It understands, extracts, and processes the information, and provides retrieval services for the users. Most of the current search engines use keyword matching. If is found to contain the keywords, the file or the website as a literature search result back to the user. Because matching the shape character, and not by their expression of the concept, often get incomplete and fault retrieval results.So, it needs to adopt some strategies to improve the search engine knowledge processing capabilities and ability to understand. It has become the future development goals and trends of the search technology. Concept-based Retrieval is a technology available. Concept Retrieval is based on keywords from the current level to a level based on knowledge. From the words expressed by the concept understanding and dealing with users search request.This paper studies the key technology of realization the concept retrieval -Knowledge Base technology, and the establishment, use, and representation of Knowledge Base. Choose ontology technology as a way to solve problems. This paper will put ontology technology, Jena reasoning and Lucene full-text search tools together. This paper did depth studies to Chemical professional vocabulary to determine the relationship between the level of vocabulary structure, which defines the category and class attributes, as well as category and the relationship between the categories. This paper created ontology in Chemical field by using Protege tools and chose OWL as ontology language. Jena tools for use to establish rules of reasoning to analysis and reasoning ontology for enquiries. Specific Index Search is used as a search engine core Lucene. Basically realizes the concept retrieval of the two main functions: synonym searching and retrieval related to expansion.This paper uses the text file for experiments, which is converted by the website snatched from the chemical professional by the Spider module of search engine system in our laboratory. By the experimental data, it can be seen that concept-based search engine has high recall rate and make intelligent search engine.更多还原