基于规则和统计的名词句法分析研究及其在机械产品设计中的应用

发布时间：2018-03-01 10:27

本文关键词： 名词短语句法分析规则模型统计模型产品设计　出处：《西安电子科技大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着计算机科学的发展和对机械设计自动化的要求不断提高，智能化的系统将是未来的发展趋势。本文尝试利用自然语言理解技术实现机械产品的需求设计自动分析，通过对以自然语言形式表达的用户需求中的名词短语进行理解和分析，并将分析结果转化成概念设计要求，为后续设计提供支持。本文分别详细地介绍了基于规则和基于统计的句法分析方法，分析和比较了这两种句法分析方法的优缺点。根据机械产品概念设计领域自然语言的特点，确定了以产生式规则和框架表示法为主的知识表示方法。结合机械产品概念设计领域自然语言的特点，分析了名词性短语的构成方式和句法结构，将名词短语按词语的组合结构分为无组合、偏正结构和联合结构三类。基于上述分类，，设计了以规则方法为主、以统计方法为辅的名词句法分析模型，该模型的规则方法按名词短语的类型进行分类处理，建立了相关词和句子的模板，采用基于历史的统计模型计算名词短语的成词概率，在自然语言理解的后续处理中，根据概率大小对不同的句法结构进行差异化处理。在上述工作的基础上，实现了基于该模型的部分关键处理过程。最后，将规则和统计结合的名词句法分析模型应用到机械产品概念设计的自然语言理解系统中，辅助设计人员进行概念设计。
[Abstract]:With the development of computer science and the increasing requirement of mechanical design automation, intelligent system will be the development trend in the future. The noun phrases expressed in natural language are understood and analyzed, and the analysis results are translated into conceptual design requirements, which provide support for subsequent design. This paper introduces in detail the syntactic analysis methods based on rules and statistics, analyzes and compares the advantages and disadvantages of these two syntactic analysis methods, according to the characteristics of natural language in the field of conceptual design of mechanical products, The knowledge representation method, which is based on the production rule and frame representation, is determined. Combining with the characteristics of natural language in the field of conceptual design of mechanical products, this paper analyzes the structure and syntactic structure of nominal phrases. The noun phrases are divided into three categories according to the combination structure of the words: no combination, positive structure and joint structure. Based on the above classification, a noun syntactic analysis model is designed, which is mainly based on rule method and supplemented by statistical method. The rules of this model are classified according to the types of noun phrases, and the templates of related words and sentences are set up. The probability of noun phrases is calculated by using a statistical model based on history, which is used in the subsequent processing of natural language understanding. Different syntactic structures are differentiated according to the probability size. Based on the above work, some key processes based on the model are realized. Finally, The noun syntactic analysis model, which combines rule and statistics, is applied to the natural language understanding system of mechanical product conceptual design.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：H146;TP391.72

【参考文献】