Research On Rule-based Uyghur-Kazak Machine Translation

Since the invention of the computer, people have been very interested in machine translation. Machine translation is called the computer refers to the use of a natural language(source language) into another natural language(target language) process. Machine Translation includes not only the non-close(non-closed) Language: machine translation(eg English, French) between, but also close to the(closed) Language(Uighur, Kazakh) between the machine translation research. Close to the machine translation between languages than not close to machine translation between languages easier. Their grammar and vocabulary common reducing development close to the machine language translation system between the effort required. Translation system requires morphological analysis, a simple conversion rules and context-dependent bilingual dictionary. Semantic analysis is usually required. Since the Uighur and Kazakh languages are close, all belong to Altaic Turkic, they have many similarities in terms of the same or structure, configuration, syntax, etc., thus achieving Uyghur machine translation into Kazakh easier.This paper presents a rule-based Uighur, Kazakh machine translation methods. According to correspondence between the Uighur, Kazakh between the two, built Uighur, Kazakh stem bilingual libraries and Uighur, Kazakh suffix mapping library, translation rules. The source language(Uighur) morphological analysis, based on stem bilingual library suffix mapping libraries and translation rules to generate the target language(Kazakh), the initial realization of the rule-based Uighur- Kazakh machine translation system.In this paper, four experiments were first conducted experiments on the rules of the present paper-based test platform and Tilmach translation online translation platform based on statistical experimental results BLEU values of 0.4567 and 0.4826. The present paper first and then translated and then translated translation test platform on Tilmach online translation platform and the first translation and then translate the Tilmach online translation platform in this thesis translation test platform, the translation results BLEU values of 0.4954 and 0.5060. Experimental results show that the rule-based Uighur, Kazakh statistical machine translation based on the combination of methods, the larger the pool of bilingual stem, the more complete the translation rules, we will improve the quality of translation.