Uyghur-Uzbek Language Machine Translation

Since the invention of the computer,people are increasingly interested in machine translation.Machine translation refers to the process by which a computer converts a natural language(source language)into another natural language(target language).Uyghur and Uzbek languages have many similarities in word structure,configuration,part of speech,syntax,and so on.Thus,it is easier to implement machine translation in Uyghur and Uzbek language.This paper proposed a method of Uygur and Uzbek machine translation based on the combination of dictionaries and rules.First of all,the Uygur and Uzbek language letters and their correspondence,syllable division rules and syllable relations,word structure and word class relations,sentence structure and compositional relationship are studied respectively.On this basis,two kinds of language matching dictionaries and letter corresponding rule base are constructed and the Uygur and Uzbek translation methods based on the dictionary and the Uygur and Uzbek translation methods based on the letter conversion rules are realized.Then,combined the two translation methods and implement the Uygur and Uzbek language translation systems based on the combination of dictionaries and rules.Uzbek language to Uygur language translation found that the the ambiguity of between the Uzbek and Uyghur alphabetes,such as one Uzbek alphabet corresponding to two or more Uyghur alphabet.For this ambiguity,the maximum entropy feature selection method is used to determine the Uygur alphabet corresponding to the ambiguous letter and generate the target language.The Uzbek language-Uyghur machine translation system based on the combination of dictionary and rule is optimized.The results of small-scale experiments show that the translation accuracy of the system is higher in Uygur and Uzbek language,and the translation quality of Uzbek to Uygur language is higher than that of Uygur language.In the future work,the system can also be through the bilingual stem library and the size of the library continues to expand,the continuous improvement of bilingual translation rules,and can use the existing dictionary and the combination of rules based on the analysis of part of speech changes and other methods effectively improve the quality of translation.