【作者】 杜昂；

【导师】 滕弘飞；

【作者基本信息】 大连理工大学 ， 机械设计及理论， 2009， 硕士

【摘要】 近年来,在产品设计领域,创新设计具有越来越重要的地位。产品创新设计中的一个关键问题是知识获取问题,即如何从海量信息中提取出符合设计需求的知识。例如在TRIZ创新概念设计的前期模糊前端阶段,原理解或领域解的求解阶段,都需要大量相关知识。当前知识获取的一个重要方式是从专利中获得,常用的方法是用关键词检索,但由于没有采用基于语义的检索方法,所以检索质量尚待进一步提高。本文重点研究基于语义的关键词检索方法,以提高检索的质量和效率,具有理论意义和应用价值。本课题首先研究专利的文本预处理方法,并构建了供语义关键词检索的专利数据库,然后研究基于语义关键词检索方法,并在此基础上给出了应用示例。主要研究工作如下:(1)给出了一种专利数据的文本预处理方法。本文对国家知识产权局网站下载的专利进行文本格式转换和存储,通过Jtidy方法实现HTML文本的转化,并通过商业软件(Adobe Acrobat Professional 8.0)实现PDF文本的转换,然后通过数据库建模将其存储到本课题组构建的专利数据库中,为用关键词或语义关键词从该专利数据库检索提供数据准备。(2)给出了一种基于语义的关键词检索方法。主要用于以下两方面:针对专利文本不带关键词的问题,用语义关键词检索方法对一个给定的专利文本,在语义理解的基础上自动提取专利文本关键词;针对用一般关键词检索范围有限的问题,利用语义关键词扩展相关专利检索的范围。最后通过编程实现了语义关键词的检索模块。(3)通过除冰雪车辆创新方案设计相关知识的专利检索实例,说明了本文给出的语义关键词检索方法应用。上述研究有助于面向创新设计的知识获取的专利检索研究,将上述成果作为模块纳入本课题组开发的“面向创新的专利知识检索系统”,用语义关键词从该专利数据库中检索所需的知识,为设计师进行创新设计的方案设计提供参考。更多还原

【Abstract】 In recent decades,innovational design is more and more important in the field of mechanical product design.One of the key problems of innovational design is how to get the needed information which meets the design requirements from massive information,that is, the problem of knowledge acquisition.For example,in the fuzzy front-end stage of the TRIZ innovative conceptual design,the obtainment of the principle solution and the domain solution requires a large number of relevant knowledge.At present,the knowledge is mainly acquired from the patents by keyword search.However,the field of search is limited due of the absence of semantic search.This paper focuses on the semantic-based keyword search in order to improve the quality and efficiency of retrieval,and has a certain theoretical significance and reference value of the application.This paper firstly researches the patent text information extraction method and builds a database,which is searched by semantic keywords,then researches the retrieval method based on semantic keywords,and gives the application examples on above-mentioned basis finally. There are several aspects of the work:(1) This paper proposes a patent(download from State Intellectual Property Office website) text preprocessing method,which performs the transformation of HTML documents by Jtidy method and transforms PDF documents by business software.Then,the data is stored in the database through the program,which can be searched by keywords or semantic keywords in the patent database and prepare for the further extension of the database using patents.(2) This paper presents a keyword retrieval method based on semantic,which can be used in two aspects as following:To deal with the problem that some patents have no keywords,the approach using semantic keywords extracts the keywords of the patent automatically according to the statistic on the basis of semantic understanding;To deal with the field limitation of retrieval on keywords,the semantic keywords is employed to extend the coverage of patent retrieval.(3) The patent search instance serving for innovative design of snow removal vehicles is used to illustrate the validation of the patent text information extraction method and semantic keyword retrieval method proposed in this paper.The above research benefits the patent retrieval of knowledge acquisition oriented to innovative design,which is used as a part of the "The patent retrieval system oriented to innovative design" developed by our research group.The approach employs semantic keywords to obtain the required knowledge by searching patent database,which can be used as the reference of scheme design in the innovative design for the designer.更多还原