【作者】 黄艳；

【导师】 潘云鹤； 耿卫东；

【作者基本信息】 浙江大学 ， 计算机应用， 2002， 硕士

【摘要】 随着电子技术和计算机技术的迅猛发展，集成电路和传感器件等数字化产品的选用成为一个关系到系统设计成败的重要问题。但是，这类产品种类多、参数多、变化快，导致传统查询模式往往无法满足各类查询用户的查询需求。为此，我们研制开发了面向数字化产品的智能柔性查询接口以支持自然语言查询方式。 本文的第一章首先介绍了信息化发展的现状，其中着重讨论了集成电路和传感器的发展现状，并由此引出本文的研究目的。 本文的第二章分析了与模糊自然语言处理相关的两种技术的发展现状：自然语言分析技术和模糊语言处理技术。 基于面向数字化产品的自然语言查询语句的冗余词汇的无效性和语义制约的歧义性的双重特点，本文的第三章提出了简化模糊语义文法的基本思想，它将模糊集合理论应用于上下文无关文法，并利用语义信息计算文法规则的隶属程度，从而确定句子的语法结构关系。 本文的第四章详细介绍了“去模糊”过程，即在模糊查询表述和精确查询条件之间架起一座桥梁，将模糊语素转化为等价的精确条件，从而构造可被数据库搜索引擎理解并执行的精确SQL查询语句。在这一过程中引入了自适应模糊处理技术以动态定制限定各类模糊词汇的隶属函数。 为了验证以上技术的实际可行性，本文第五章实现了一个面向数字化产品的自然语言查询接口。最后，文章的第六章对对当前研究工作进行总结，指出了现存的不足，并对系统应用和未来的工作提出了展望。更多还原

【Abstract】 With the rapid development of electric and computer techniques, the choice of digital products such as integrated circuits and sensors becomes a key problem in system designs. However, due to the various categories and parameters and the rapid changes of digital products, traditional query modes can no longer satisfy the query requirements of different kinds of end users. Therefore, we developed an intelligent and flexible digital product-oriented query interface which allows end users to employ natural language sentences to express queries. In chapter one, we reviewed the development of information industry, especially focusing on the industry of integrated circuit and sensors, based on which, the object of our current research is introduced. Chapter two gives a survey of the development of natural language parsing techniques and fuzzy language processing techniques, two key techniques that serve to process fuzzy natural language query sentences. In chapter three, the basic idea behind the proposed technique of Reduced Fuzzy Semantic Grammar (RFSG) is presented. RFSG is established based on the dual characteristics of digital product-oriented natural language query sentences, that is, ineffectiveness of redundant words and ambiguity restricted by semantics. This granunar parses a sentence structure by applying fuzzy sets.theory to the context-free grammar and exploring semantics to calculate the membership degree of each grammar rules. Deffuzification process is described in chapter four in detail, in this process, a bridge is built between imprecise and precise to transform a fuzzy term to its equivalent precise forms so that an executable SQL statement can be generated. Here, adaptive fuzzy process technique is explored to dynamically define the various membership functions that restrict corresponding fuzzy terms. To validate the feasibility of the proposed techniques, in chapter five, a digital product-oriented natural language query interface is implemented. Finally, chapter six summarizes the main contributes of our work and those work under improvement.更多还原