| 中文天然说话处置多以古代汉语为研究对象，但古汉语作为中国传统文明的主要载体，其主动句法和语义剖析研究对于中国传统文明的西传具有主要意义。对于古汉语的主动句法和语义剖析研究也会增进古代汉语主动剖析的成长和提高。古代汉语的主动句法和语义剖析获得了很年夜提高，但以后通用语法和句法剖析器在应用于古汉语句法剖析和语义表现时存在必定缺乏。本文以数据库语义学为实际框架，以其主要构成部门一—联合语法为技巧支撑，基于时光线性准绳和描写性准绳，以左传及其英译本为语料来剖析古汉语的根本句法和语义关系。本文的研究任务重要分为以下几个方面：第一，依据数据库语义学句法剖析的详细请求，以左传及其英译本为语料起源，树立一部双语辞书。词条的存储情势为“命题粒”，“命题粒”是一种非递归特点构造，是一个属性值对的聚集。词的语法信息和语义信息分离作为其特点构造中响应属性的值停止具体标注，知足基于数据库语义学和左联合语法句法语义剖析须要的同时，法语论文范文，从词的条理上下降由一词多义和词类活用惹起的歧义产生率。第二，从天然说话处置的角度动身，对左传的语法特色、句式构造从新停止梳理，并联合基于左联合语法的主动句法剖析的须要，总结归结根本句律例则，从名词、动词、描述词三年夜词类的根本用法动身，研究和剖析了主谓构造、结合构造、动宾构造等根本构造，和宾语前置、语义主动、情势主动和成份省略等根本构造的变体情势。第三，在实词处置方面，提出有前提移植法。在虚词接收实词的进程中，有前提保存实词的焦点属性值和／或语义属性值。这和数据库语义学本来的处置办法分歧，可以或许防止说话生成和机械翻译进程中能够涌现的年夜量回溯。第四，在句法剖析进程中透过表层构造发掘说话内容，剖析深条理的语义关系和语用内在，经由过程规矩操作的办法为词的语义属性添加新值，法语毕业论文，以表示其施事、受事、历事、修辞、主动等语义脚色和语用功效。我们等待基于改良的数据库语义学的主动句法语义剖析办法可以或许在往后应用于其他年夜范围语料的研究和剖析上，好比和左传生成年月分歧且具有分歧语法特点的古汉语文本。别的，基于本文研究基本上的说话生成和机械翻译也是我们后续研究的偏向之一。 Abstract: Chinese natural language disposal multi based on ancient Chinese as the research object, but ancient Chinese as the main carrier of Chinese traditional culture, the active syntactic and semantic analysis and study about Chinese traditional civilization of the West pass has important significance. About the ancient Chinese active syntactic and semantic analysis will promote the analysis of ancient Chinese initiative to grow and improve. Ancient Chinese active syntactic and semantic analysis to obtain the very big improvement, but after universal grammar and parser in ancient Chinese syntactic analysis and semantic expression, there must be a lack in using. In this paper, the Database Semantics for practical framework, to the main part of a united grammar for technical support, based on linear time yardstick and description of principle, in "Zuo Zhuan" and its English version as corpus to analyze the ancient Chinese fundamental syntactic and semantic relations. The research of the following aspects: first, based on Database Semantics parsing detailed request, in "Zuo Zhuan" and its English version is derived from the corpus, establish a bilingual dictionary. The proposition in the grain storage situation of entry 1, "proposition particle" is a non recursive structure characteristics, is an attribute value of the aggregation. Grammatical and semantic information of separation as the structural features of the response properties of the values of specific annotation, contented need based on Database Semantics and United left grammar syntactic and semantic analysis, from the words of structured decreased by polysemy and word usage cause ambiguity rate. Second, set out from the angle of natural language disposal, on the grammar characteristic of the "Zuo Zhuan", sentence structure from the new carding, and joint for the needs left union syntax active syntactic analysis, summarizing the fundamental sentence statutes, from nouns, verbs, and describe the start with the basic usage of three of the eve of the word parts of speech, research and analysis of the subject predicate structure, combined with structural and dynamic object construction and basic structure, object and active front, semantic, situation active and component ellipsis and basic structure of variant forms. Third, in the substantive disposition, put forward the premise transplantation method. In the process of receiving the notional words, a notional value of focal attribute preserving and / or semantic attribute value. This database semantics and the original disposal method to differ, can perhaps prevent speaking in the process of generation and machine translation to the emergence of a large number of backtracking. Fourth, in the parsing process through the surface of the structure to explore speech content, analysis of the deep-seated semantic relations and pragmatic inner, via a process is operating rules for semantic properties of words to add new value, expressed in the agent, patient, calendar, rhetoric, the initiative for semantic feet color and pragmatic functions. We wait for based on improved Database Semantics active syntactic and semantic analysis method can perhaps in the later used in other large-scale corpus of research and analysis, like and Zuozhuan generation different eras and has different grammatical features of ancient Chinese text. One of the other, this paper studies the bias basically talk generation and machine translation is based on our follow-up study. 目录: |
