语句级汉字输入技术
[作者]王晓龙; 王幼龙;
[摘要]本文讨论了包括声音输入、键盘输入、文字识别等各种形式的汉字输入技术的研究和发展，阐述了按照字、词、语句作为汉字输入技术发展阶段的思想，提出了适用于上述各种形式的类码语句岐义处理问题，该问题可描述为有向图求最短路径的问题。本文讨论了采用语法──语义分析和统计模型的最少元素概率推理方法和控制策略，在知识库完备或不完备的情况下均可进行正常的推理，并给出基于当时情况下的最佳结果。本文还简要介绍了几个应用事例。
[Abstract]This paper discusses the research and development of Chinese input technology in keyboard, speech and character recognition fields, which can be dividing as three stages:character-input, word-input and sentence-input. In the concept of a directed graph, a method is presented to solve the main problem of the third generation Chinese input system, i. e. the problem resulted by different meaning of the sentence inputted by keyboard or speech system or recognition system. Various experiments show that higher fl...
[关键字]人机接口; 自然语言理解; 中文信息处理; 语句输入;
|
汉字输入法类的设计与实现
[作者]赵雷; 吕强; 杨季文; 朱巧明;
[摘要]本文指出了汉字输入法实现技术不足，提出了一种全新的汉字输入法与汉字平台脱离的松散连接的结构体系，给出了用面向对象技术来实现汉字输入法类的方法，描述了汉字输入法类的结构，简要说们它的用法，最后给出了一个用汉字输入法类实现汉字输入法对象的实例。
[Abstract]his paper points out the shortcomings of the implementation of the Chinese Characetr input method, and puts forword a loosely coupled structure that makes the Chinese Character input method seperate from the Chinese Platform. It presents a method to realize a Chinese Character Input Method CLASS using Oriental Object techenology. In this article, we introduce the structure of the CLASS, and describe its usage briefly. An example of a Chinese Character Input Method OBJECT is also shown at the end of the pape...
[关键字]汉字输入法; 类;
|
数词的语义结构及通用翻译算法
[作者]郭宏蕾; 姚天顺;
[摘要]本文给出了数词的语义结构描述及知识表示，提出高层位数词、号位数词的概念，较好地刻化了各语种读数的语言规律。在此基础上，提出一种通用数词翻译算法，使数词翻译在实现上具有独立于语种的特点。
[Abstract]Abstract The paper presents the semantic structure and knowledge representation of the numeral. The concepts of superior position numeral and subordinate position numeral are given, which can exactly describe the linguistic principles of the numeration in the different native languages. Based upon the Principles, the authors put forward a general algorithm of the numeral translation. The algorithm is independent of each language in the architecture and flexible to be realized.
[关键字]位数词; 系数词; 词的语义结构;
|
汉字图象的小波分析
[作者]梁玉尧; 李多; 马争鸣;
[摘要]目前，我们进行汉字模式识别所使用的汉字是以图象的方式输入计算机。要将其转换为讨算机所能识别的字符，其关键在于输入汉字图象特征的提取。作为图象的汉字有其自身的特点，它是由较简单的笔划所组成的，每种笔划又有其较固定的方向性（即空间分步的固定性），所有的汉字都是由几种简单的笔划所组成的。而小波变换为我们提供了一个十分有效的分析图象信息的多分辨率方法，它可以将原始图象分解为模糊子图象和水平方向、垂直方向、斜方向上的子图象。因此，小波变换为我们分析汉字图象信息提供了一个十分有效的手段。本文正是基于汉字的上述特点，并利用小波对空间频率的多分辨率分析方法，对汉字图象处理而得到汉字图象特征的。
[Abstract]Abstract In tradition, Chinese characters recognition is a difficult problem in the computer characters recognition because of its complex structure. But Chinese characters have their own speciality, it can divide into five Basic strokes, each of them can project into horizontal, vertical and catercornered-direction. This is an important speciality in synactic pattern for Chinese characters recognition. This paper has proposed a method to obtain the Chinese characters speciality by using wavelet decompositi...
[关键字]小波变换; 快速小波变换; 汉字图象特征;
|
基于假设检验的手写印刷体汉字识别方法
[作者]黄铁军; 胡家忠;
[摘要]为将统计决策方法和句法方法有机结合起来，本文提出了以部件为基元的基于假设检验的手写印刷体汉字识别方法：由统计方法得到候补字集，利用部件特征的先验知识抽取待识字可能包含的部件并对假设进行验证，从而不断缩小候补字集，并逐步完善汉字的结构描述。初步实验表明其分类效果明显。
[Abstract]o combine statistical decision method and syntactic structural method, this paper advanced an idea about a leading syntactic structural method based on hypothesis testing that regards elements as basis units. This method refers to the classification result of decision method as the candidate set, then tries to take out possible elements by their structural features and tests the hypothesis. The process will reduce the candidate set step by step and give more and more comprehensive description of the charact...
[关键字]手写印刷体汉字识别; 部件; 假设检验;
|
计量语言学统计分析软件系统
[作者]郑玉玲; 沈米遐; 徐昂;
[摘要]本文介绍近期完成的国家自然科学基金项目＜藏缅语语料库及比较研究的计量描写＞的软件系统。该系统建立了我国境内藏缅语族五大语支８２个语言点１６万词条的开放性词汇语音数据库。研制了语言特征统计，语言比较研究软件。设计了应用于多种语言谱系分类比较研究的语音对应关系“全方位交叉”算法。对藏语方言的音节、音位、声母、韵母、声词、词素、构词能力和语音结构等１０余项特征做了分布和对比统计。对藏语１５个方言点做了语音对应关系和音系对比关系的量化描述，并在此基础上做出具有历时与共时比较研究意义的Ｒ相关和Φ相关分析，得出了语言分类的相关矩阵和聚类分析图表。
[Abstract]This paper makes an introduction to the statistical software of the project" the Tibeto-Burman Corpus and the Quantitative Description of the Comparative Studies", Sponsored by the National Natural Sciences Foundation of China. This project establis hes an overt lexical and Phonetic corpus with 160, 000 vocabulary entries of 82 different languages or dialects bclonging to the five branches of the Tibeto--Burman Group wi thin China, develops a package for the statistics of linguistic features and comparativ...
[关键字]藏缅语; 语料库; 计量分析;
|
建立现代汉语依存关系的层次体系
[作者]刘伟权; 王明会; 钟义信;
[摘要]依存语法［１］通过分析语言单位内成分之间的依存关系揭示其句法结构。本文针对从短话到句群的各级单位内部的各种依存关系展开讨论。提出依存关系普遍存在于各级单位之中，初步建立了汉语依存关系的层次体系。这一体系覆盖了大部分常见的语言现象，经检验可应用于句法分析过程中，作为表示句子结构的一种手段。为了增进体系的完备性和正确性，显然还有许多工作要做。
[Abstract]Dependency Grammar provides a method to analyse the syntax structures of every usage of language, through phrases to sentences. To parse a structure, Dependency Relation Trees (DRTs) can be employed to depict the dependency relations between different components. In this paper, it is discovered that dependency relations are prevalent in grammar units of every level. A hierarchy structure of dependency relations in Chinese is pro posed and discussed in detail. To improve the completeness and correctness o...
[关键字]句法分析; 依据语法; 依存体系; 汉语;
|
高速中文PostScript系统研究
[作者]廖恒; 吴昭; 李三立;
[摘要]ＰｏｓｔＳｃｒｉｐｔ为桌面设计过程中，排版印刷的后端处理的开放标准。高速中文ＰｏｓｔＳｃｒｉｐｔ系统的实现涉及到多方面的技术。我们一方面研究了中文ＰｏｓｔＳｃｒｉｐｔ系统的软件构成和算法；另一方面，设计并实现了基于Ｉｎｔｅｌ８０９６０ＣＡ超标量亿次处理器的高速中文ＰｏｓｔＳｃｒｉｐｔ处理硬件系统，进一步将其扩充为分布式并行处理系统，并在通用ＰＣ平台上实现单机和并行中文ＰｏｓｔＳｃｒｉｐｔ系统；此外，我们还提出了存储压缩等相关技术。本文探讨了上述系统的构成，并对系统进行了测试和性能评价。测试表明，我们通过采用这些高速的硬件系统及与其相适应的软件算法，基本解决了中文ＰｏｓｔＳｃｒｉｐｔ处理速度慢和成本高的问题。
[Abstract]PostScript is a standard of page description in Desk Top Publishing area. Building high speed Chinese PostScript System involves technology of many aspects. We study the software structure and algorithm of Chinese PostScript system, as well as the hardware system for PostScript processing. We designed a hardware platform based on Intel 80960CA superscalar processor, then we extend it to a network parallel processing system. Single processor and multiprocessor system based on PC platform is also studied; We...
[关键字]中文PostScript; Desktop publishing; Parallel Processing; cluster computing supercomputing memory enhancement;
|
一种基于语言理解的输入方法──智能拼音输入方法
[作者]吴军; 王作英; 郭进; 王政贤;
[摘要]Abstract In this paper, smart spelling inputting method, a high-speed Chinese Character imputting method based upon language understanding is proposed. Upon this approach, the translation from Chinese spellings of a sentence to Chinese characters is done automatically by studying the relations among the contexts. When using this method, the users do not need to select Chinese characters after input spellings as when using other spelling inputting method, because the characters of a sentence are give out co...
[关键字]智能拼音输入方法; 拼音－汉字自动转换; 台尔可夫过程;
|
现代汉语语法信息词典规格说明书
[作者]俞士汶; 朱学锋; 王惠; 张芸芸;
[摘要]《现代汉语语法信息词典》是为计算机实现汉语分析和汉语生成而研制的一部电子词典。这部电子词典可以在语言信息处理的广泛领域中得到应用。本词典的详细规格说明书（ｓｐｅｃｉｆｉｃａｔｉｏｎ）的初稿制订于１９９０年。在八五攻关期间（１９９１年至１９９５年），一方面严格按照规格说明书进行词典内容的开发，一方面在开发过程中又对规格说明书进行了局部的调整与修订，于１９９５年１１月形成了现在的版本、这份规格说明书也是汉语信息处理研究的一项重要成果。现在发表的这份规格说明书共分以下五章：第一章介绍词典的设计目标与结构。第二章介绍总库的属性字段。第三章介绍各类词库的共同属性字段、第四章介绍各类词库专有的属性字段。第五章介绍《现代汉语语法信息词典》于１９９５年１１月通过专家鉴定时所达到的规模。附录给出了面向信息处理的现代汉语词语分类体系的词类代码表。
[Abstract]
[关键字]现代汉语; 语法信息; 词类; 电子词典;
|
共95页 当前第62页 