基于规则的维吾尔人名汉文机器翻译算法研究

ON RULES-BASED CHINESE MACHINE TRANSLATION ALGORITHM FOR UYGHUR PERSONAL NAME

人名机器翻译传统的方法是依赖于预先建立的对照词库。这种方法的缺点是,因为不断出现新的人名而建立丰富、全面的对照词库是不可能的。研究了维吾尔人名汉文音译规则:读音规则,并在此基础上提出了一种新的基于规则的维吾尔人名汉文机器翻译方法。实验结果表明,该方法不需要建立两种语言的对照词库,完全不用考虑会出现的新的人名,算法实现简单,准确度高,完全克服了传统方法的缺点。

Traditional method of personal name machine translation relies on pre-established comparative dictionary. Owing to continuous e- mergence of new names, the disadvantage of this approach is the impossibility of establishing a rich and comprehensive dictionary. In this pa- per, we study the rule of Uyghur to Chinese personal names transliteration：the pronunciation rule, and propose a new Uighur persona/names to Chinese machine translation method based on it. Experimental results indicate that our method does not require to establishing a Uyghur- Chinese bilingual comparative dictionary. There is no need to consider that there will be new personal names, it is simple to implement and has high accuracy, thus completely overcomes the shortcoming of traditional method.