中文情感语义资源管理平台的构建

发布时间：2018-04-23 23:30

本文选题：框架语义 + 语义词典　；参考：《山西大学》2016年硕士论文

【摘要】：随着Web2.0技术的发展成熟,人们的生活和工作都受到互联网应用的影响而发生着改变,Web2.0理念中将网站和浏览者之间的交流放在了很重要的位置,用户不仅是浏览网站信息的人,也是创建网站信息的主体。人们从被动接受信息逐渐转变为主动寻找和发布信息,这些信息成为宝贵的资源。网络上的信息根据其语义内容可以分为客观性信息和主观性信息。在Web2.0环境下,出现了大量的主观性信息,如用户在博客中表达的观点,在电子商务网站发布的商品评论信息等。对这类信息的处理对政府、相关机构、企业乃至电子商务运营商等都具有巨大的价值。而由于信息量的庞大,人工分析费时费力,我们迫切需要一些技术,能够有效地处理网络上的主观性信息。通过对文本的语义分析可以实现较高精度和深度的主观性内容挖掘,但是,中文信息处理缺少语义词典和语料库以及对语义资源的管理来对语义分析进行有效的支撑。论文以中文情感语义框架资源为基础,提供情感语义资源编辑系统和情感语义资源本体的方案,对语义资源进行有效的集合和管理。本文的具体研究内容如下:本文首先对国内外情感语义现状进行了简要介绍,并且阐述了现有的情感语义资源的不足,总结出一种新的中文情感语义表示方法,主要以框架语义理论为核心,介绍了以框架为基础的情感语义表达模型以及中文框架语义语料标注的方法。在充分分析的基础上,运用Java语言构建一个中文情感语义资源管理平台,它有两方面的功能,一是情感语义资源的编辑、修改和展示功能,二是用本体组织该资源。情感语义资源编辑系统是面向人的,既包括情感资源的构建者,也包括浏览者。在具体分析了系统需求的前提下,实现了用户管理、框架编辑、词元编辑、情感语料库标注和情感语义资源浏览的功能,方便用户对情感语义资源的了解和使用。中文情感框架本体的构建是面向机器的,用规范化的描述体系使其成为机器可读、可理解的情感语义资源,实现概念化的推理和智能化的情感挖掘应用,为语义的分析和观点挖掘提供了强大的支撑。
[Abstract]:With the development of Web2.0 technology, people's life and work are affected by the Internet application, so the concept of Web 2.0 has changed the communication between the web site and the viewer is very important, the user is not only the person who browses the website information. Also is the main body that creates the website information. People gradually change from passively receiving information to actively seeking and publishing information, which becomes a valuable resource. The information on the network can be divided into objective information and subjective information according to its semantic content. In the environment of Web2.0, a large number of subjective information appeared, such as the views expressed by users in blogs, commodity reviews published on e-commerce websites, and so on. The processing of this kind of information is of great value to government, relevant organizations, enterprises and even e-commerce operators. Due to the huge amount of information and the time-consuming and laborious manual analysis, we urgently need some techniques to deal with the subjective information on the network effectively. Through semantic analysis of text, we can realize subjective content mining with high accuracy and depth. However, Chinese information processing lacks semantic dictionary, corpus and management of semantic resources to effectively support semantic analysis. Based on the Chinese affective semantic framework resources, this paper provides a scheme of affective semantic resource editing system and affective semantic resource ontology, which can effectively collect and manage semantic resources. The main contents of this paper are as follows: firstly, the present situation of emotional semantics at home and abroad is briefly introduced, and the deficiency of existing affective semantic resources is expounded, and a new method of expressing emotional semantics in Chinese is summarized. Based on the frame semantic theory, this paper introduces the frame based emotional semantic expression model and the method of Chinese frame semantic corpus annotation. On the basis of full analysis, a Chinese affective semantic resource management platform is constructed by using Java language. It has two functions, one is editing, modifying and displaying emotional semantic resources, the other is organizing the resources with ontology. The affective semantic resource editing system is human oriented, including both the constructor and the viewer. Based on the detailed analysis of the system requirements, the functions of user management, frame editing, lexical element editing, emotional corpus annotation and emotional semantic resources browsing are realized, which is convenient for users to understand and use emotional semantic resources. The construction of Chinese affective frame ontology is machine-oriented. It is made into a machine-readable and comprehensible emotional semantic resource by using a standardized description system to realize conceptualized reasoning and intelligent affective mining applications. It provides a strong support for semantic analysis and viewpoint mining.
【学位授予单位】：山西大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.52;G250.74

【参考文献】