限定领域的汉语-维吾尔语句子级对齐研究  

Limited the Field of Chinese-Uighur Statement Sub-level Alignment

作　　者：张亚军[1] 贺琛琛[2] 香丽芸[1] 

机构地区：[1]昌吉学院计算机工程系,新疆昌吉831100 [2]昌吉学院人事处,新疆昌吉831100

出　　处：《软件》2014年第3期62-64,共3页Software

基　　金：国家自然科学基金(项目编号:60963018);新疆维吾尔自治区高校科研计划青年教师科研基金项目(项目编号:XJEDU2011S42);昌吉学院项目(项目编号:2012YJQT03;12JYYB004)~~

摘　　要：针对政府文献的汉语维吾尔语语料库特点,充分利用汉语和维吾尔语的句子特性,提出一种汉维句子级别的对齐方法。该方法重点分析政府领域的汉语和维吾尔语的句型,分别对汉语和维吾尔语的语料进行边界识别,避免了复杂句型对汉语-维吾尔语句子对齐的影响,使得该方法取得句子对齐达到97%与99%之间的正确率。对齐的汉语-维吾尔语句子对可以充实语料库的规模,为汉语-维吾尔语短语对齐以及汉维机器翻译提供翻译语料。Chinese Uighur Corpus characteristics for government documents, and take full advantage of the features of Chinese and Uighur’s sentence, and put forward Chinese-Uighur sentence level alignment. The method analyzes government areas of Chinese and Uighur’s sentence boundary identification of the corpus of Chinese and Uighur, to avoid the impact of complex sentences in Chinese and Uygur sentence aligned, which makes the method to obtain the sentence alignment up to 97%and 99%correct rate. Alignment of the Chinese-Uighur statement that the child can enrich the corpus of the scale, Chinese-Uighur phrase alignment and machine translation in Chinese and Uygur translation corpus.

关 键 词：平行语料库 句子对齐 机器翻译 汉语-维吾尔语 句子边界识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]