基于统计的藏文音节字校对系统开发研究  

Development of proofreading Tibetan syllables word system based on statistics

作　　者：普布旦增[1] 关白[1] 

机构地区：[1]西藏大学藏文信息技术研究中心,西藏拉萨850000

出　　处：《西藏大学学报（社会科学版）》2015年第2期74-78,共5页Journal of Tibet University

基　　金：2012年度国家自然科学基金青年项目"现代藏文自动校对研究"(项目号:10YJCZH033);2012年度国家自然科学基金项目"藏语依存树库的构建"(项目号:61163043)阶段性成果

摘　　要：音节字是构成藏文文本的基本单位,也是整个现代藏文自动校对的基础。音节字校对的结果直接影响词校对的正确率,从而对词、格助词、句法和语义校对等产生重要影响。文章以语料库统计得到的藏文音节字字表、二元同现表和纠错建议表等作为语言知识库,以藏文音节字的错误类型为依据,设计了字表匹配和查找同现概率为侦错方法、用纠错建议表进行纠错的音节字校对系统。Syllables word is basic unit of constituting Tibetan text and is also the foundation of modern Tibetan automatic proofreading. Result of syllable word proofreading directly affect the accuracy of the word proofreading thus leading to bring an important impact on proofreading of word, preposition, syntactic and semantic. Tibetan syllable word table summed from corpus, binary co-occurrence tables and correction suggestion tables as the lan- guage knowledge baseline, according to error types of Tibetan syllable word, the Syllables word proofreading sys- tem of using error correction suggestions table for correcting was designed using word table matching and finding probability of co-occurrence as debugging method.

关 键 词：音节字 校对 侦错 纠错 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]