FST,词性标注,HMM,命名实体,TokenList,正则表达式,基于FST技术修正中文词性标注的探讨与实现
基于FST技术修正中文词性标注的探讨与实现
专业****团队→竭诚为你服务→论文请联系****→
  • 臂力论文网精英****团队,竭诚为你服务
当前位置:臂力论文网 >> 计算机论文 >> 人工智能 >> 浏览文章基于FST技术修正中文词性标注的探讨与实现

基于FST技术修正中文词性标注的探讨与实现

http://www.zidir.com/html/jsjlw/rgzn/498257.html


最新
  • 多数据源信息融合的电网故障诊
  • 孔子周游列国事迹及系年考
  • 网络动态文字与人类情感的智能

  • FST,词性标注,HMM,命名实体,TokenList,正则表达式,
    正文开始
    自然语言理解又被称为自然语言处理或计算语言学,它是人工智能领域中的前沿难题之一。自然语言的识别和处理是人工智能研究的最重要的课题之一,也是人工智能研究的关键。其中在自然语言处理中的汉语词性标注是中文信息处理技术中的一项基础性课题,一个确切精准的词性标注对自然语言的理解有着极其广泛的意 义,特别是在对输入文本进行句法浅析浅析、语义浅析浅析时,词性标注是一项必不可或缺的处理任务,因此,研究和实现汉语词性标注器具有重要的理论意 义和实用价值。词性标注的策略主要有基于规则策略和基于统计的策略两大类。一般的情况下,为了达到更好的词性标注结果,往往都是采取基于统计和基于规则相结合。在基于统计的策略中,主要是采取隐马尔科夫模型(HMM),而基于规则的策略中系统中主要是采取有限状态转换机(FST)的策略,目前在自然语言处理上的应用上,FST策略在理论上还比较欠缺。在本文中就如何把FST应用到自然语言处理的词性标注上做了详细研究,并最终给出了实现的结果。最近几年来,在国际新一代计算机激烈竞争的影响下,自然语言理解的研究在国内得到了越来越多的重视,研究单位在逐渐增加,研究队伍也在逐渐壮大。目前在国内的研究中比较有代表的研究成果主要有机器翻译、语料库的研究、篇章理解研究、受限汉语研究等。但是不管怎样,所有的研究的前端都必须有词性标注这一项。

    【Abstract】WwW.zIdiR.coM Natural Language Understanding is also known as natural language processing or computational linguistics, it is one of the forefronts of problems in the field of artificial intelligence. Natural language recognition and processing is one of the most important topics in artificial intelligence research and is also the key to artificial intelligence research. Chinese Part-of-Speech Tagging is a fundamental subject to Chinese information processing technology in natural language processing; a precise part of speech tagging has a very wide range of meaning for accurate understanding of natural language, POS tagging is an essential task especially in syntactic analysis, semantic analysis. Therefore, research and implementation of Chinese tagging device is of great importance both in theoretical and practical aspect.There are two kinds of method in speech tagging, one is based on rules and another is based on statistics. Generally, in order to achieve better results of speech tagging, we often combine these two methods in practical application. Based on statistical methods, Hidden Markov Model (HMM) is mainly taken by, and we take Finite-State-Transducer (FST) approach at rule-based approach. So far, the theory reservation in the application of Natural language processing is deficient. In this paper I have done an further study on how to apply FST to natural language processing of speech tagging and given the results achieved ultimately.In recent years, under the influence of a new generation of computers in the fierce international competition, the study of this field has caused more and more attention. Research units and research teams are also gradually expanding. Currently in China, machine translation, corpus research, understanding research chapter and restricted Chinese studies are on behalf of the main results of the research. However, all these researches must have the front-end research on Part-of-Speech tagging.

    【关键词】 FST; 词性标注; HMM; 命名实体; TokenList; 正则表达式;
    【Key words】 FST; Part-of-Speech Tagging; HMM; NE; TokenList; Regular Expression;
      http://www.zidir.com/html/jsjlw/rgzn/498257.html
      推荐
      关于人工智能的应用与发展 
      对人工智能的后现代主义反思 
      对人工智能及其发展的哲学思 
      略论人工智能技术在财务管理 
      J**A技术与人工智能在搜索引 
      本科高年级人工智能教学的几 
      从复杂性系统理论视角看人工 
      当代人工智能表征的分解方法 
      人工智能技术在计算机辅助教 
      电子商务与人工智能技术 
      语音识别中消除混响影响的方 
      基于Nios II的自动指纹识别 
      展望仿真技术在我国经济社会 
      人工智能在电气传动中运用的 
      走近人工智能1 
      植人式人工电动智能掌指关节 
      请您记住我们臂力论文网,竭诚为您服务 WWW.zIdir.com
      FST 词性标注 HMM 命名实体 TokenList 正则表达式
      基于FST技术修正中文词性标注的探讨与实现 》由臂力论文网 整理提供