面向机械产品设计领域中自然语言理解的动词聚类

发布时间：2018-06-07 22:57

本文选题：自然语言理解 + 篇章分析　；参考：《西安电子科技大学》2011年硕士论文

【摘要】：本文对自然语言理解中篇章动词聚类模型进行了研究,并将其成功应用于机械产品设计过程中。利用整个自然语言理解系统,着重研究篇章分析中的动词聚类过程,最终将以自然语言表示的用户需求转化成概念设计要求,为后续设计提供支持。对自然语言理解中的动词进行了全面的分析,并根据动词的阶段性,对动词进行了二级分类。基于动词的分类,研究了事件的分类,以及动词和事件的关系。然后依据分类设计了动词聚类的相应处理方案,重点研究了过程聚类处理方案。在此处理方案的基础上,结合本系统的知识表示方法,建立动词聚类的处理模型。该模型以知识模板的形式来表示动词聚类过程中需要的静态知识和动态知识,从概念的内涵与外延出发,通过确定动词聚类的核心事件,依据领域过程树以及关联分析最终实现了基于篇章层面围绕核心事件的动词聚类。最后,结合自然语言理解的总体框架将此模型整合进去,并应用于机械产品的设计领域,并建立该领域的专业知识库,进行测试,取得了满意的效果。
[Abstract]:In this paper, the text verb clustering model in natural language understanding is studied, and it is successfully applied in the process of mechanical product design. With the help of the whole natural language understanding system, the verb clustering process in text analysis is studied. Finally, the user needs expressed in natural language are transformed into conceptual design requirements. In this paper, the verbs in natural language understanding are comprehensively analyzed, and the verbs are classified into two levels according to the stages of the verbs. Based on the classification of verbs, this paper studies the classification of events and the relationship between verbs and events. Then, according to the classification, the corresponding processing scheme of verb clustering is designed, and the process clustering scheme is studied emphatically. On the basis of this scheme and the knowledge representation method of this system, the processing model of verb clustering is established. In this model, the static knowledge and dynamic knowledge needed in the verb clustering process are expressed in the form of knowledge templates. From the connotation and extension of the concept, the core events of verb clustering are determined. According to domain process tree and relevance analysis, the verb clustering based on text level around core events is finally realized. Finally, the model is integrated with the general framework of natural language understanding and applied to the field of mechanical product design. The professional knowledge base in this field has been established and tested, and satisfactory results have been obtained.
【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】：2011
【分类号】：TH122

【参考文献】