基于统计的维文汉文人名音译研究  

Research of the Uyghur-Chinese Personal Name Transliteration Based on Statistics

作　　者：谭煜辉[1] 吐尔根·依布拉音[1] 艾山·吾买尔[1] 买合木提·买买提 

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046 [2]新疆新能信息通信有限责任公司,新疆乌鲁木齐830026

出　　处：《新疆大学学报（自然科学版）》2012年第1期108-111,共4页Journal of Xinjiang University(Natural Science Edition)

基　　金：工信部电子发展基金-维哈柯语言文字软件开发及产业化-维哈柯文辅助翻译软件;新疆维吾尔自治区自然科学基金(2011211B07);新疆多语种信息技术开放课题;青年教师科研培育基金(XJEDU2010S07);新疆大学博士启动基金

摘　　要：传统的维汉人名音译大多是基于规则的,不同于基于语音的音译,本文在直接正字匹配(DOM)的框架下,将统计的思想引入到维汉人名的音译中,采用信源信道模型,将人名看成是特殊的句子,实现了维汉人名的自动音译.实验结果表明,直接正字匹配减少了中间过程,从而提高了音译的准确率,而统计的机器翻译方法比较灵活,能利用外来的信息,更加适合进行维汉人名音译.The most traditional Uyghur-Chinese personal name transliteration bases on rules, this paper introduces the idea of statistics into the Uyghur-Chinese personal name transliteration. Different from phoneme-based transliteration, this paper achieves Uyghur-Chinese personal name automatic transliteration, in the framework of Direct Orthographical Mapping, adopting the Source Channel Model, treating person name as the specical sentence. The experimental results show that Direct Orthographical Mapping reduces intermediate process, thus improving the accuracy rate of transliteration. Moreover, the statistical machine translation method is more flexible, and can use external information, and it is more suitable for Uyghur- Chinese personal name transliteration.

关 键 词：自然语言处理 统计 音译 维汉人名对 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]