对话系统数据预处理模块的设计与实现

发布时间：2018-01-27 09:25

本文关键词： 微服务架构 SpringCloud 知识图谱　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网技术的发展,人机交互的方式也发生了很大的改变。对话系统是人机交互技术发展到现在的产物,它允许用户用自然语言和计算机应用程序交互。近些年来,知识图谱的兴起也给对话系统带来了新的思考和新的想法。本人在实习公司参与的项目就是对话系统,该对话系统通过应用接口的方式为其他系统提供服务。本人负责的是数据预处理模块的工作,包括数据采集、数据存储、知识图谱的构建。本人具体的工作内容如下:(1)前期进行大量的调研和分析,了解行业背景和工作流程,分析现有系统的不足和问题,查阅文献资料,针对性解决问题,提出解决方案。(2)独立完成实现了数据采集的全部功能,参与了知识图谱构建的相关模块。(3)基于SpringCloud的微服务架构实现了数据采集,将数据采集流程中的各个部分变成了独立的服务。(4)在数据采集当中研究了基于XML(Extensible Markup Language)的页面可视化的配置,分析总结了不同数据之间的相互转换的规则,实现了抽取规则的可视化配置。(5)研究了知识图谱构建的相关文献,和同事总结了构建的一些步骤,通过自然语言处理的技术进行处理,最终构建成知识图谱,这些步骤还处在实践优化阶段。(6)参与并对相关模块进行了测试。本人参与的工作完成了现阶段的目标,还在进行着持续性的优化和开发。数据采集系统运行良好,知识图谱正在分领域的构建中。
[Abstract]:With the development of Internet technology, the way of human-computer interaction has changed greatly. Dialogue system is the product of human-computer interaction technology. It allows users to interact with computer applications in natural languages. The rise of knowledge map also brings new thinking and new ideas to the dialogue system. The project I participated in in the internship company is the dialogue system. The dialogue system provides services to other systems through the application interface. I am responsible for the work of the data preprocessing module, including data acquisition, data storage. The construction of knowledge map. My specific work content is as follows: 1) the previous stage of a large number of research and analysis, understanding of the industry background and work process, analysis of the shortcomings and problems of the existing system, consult the literature. Aiming at solving the problem, putting forward the solution. 2) accomplishing all the functions of data collection independently. Participate in the building of knowledge map related module. 3) based on the SpringCloud micro-service architecture to achieve data acquisition. Turn each part of the data acquisition process into a separate service. 4) in the data collection, we study the data acquisition based on XML(Extensible Markup language. The configuration of page visualization for. This paper analyzes and summarizes the rules of conversion between different data, realizes the visual configuration of extraction rules. (5) researches the related literature of knowledge map construction, and summarizes some steps of construction. Through the natural language processing technology to process, the final construction of knowledge map. These steps are also in the practical optimization stage. 6) participate and test the relevant modules. I participated in the work completed the goal of the present stage. The data acquisition system is running well and the knowledge map is being constructed in different fields.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP274;TP311.52

【参考文献】