Research Of Uyghur N-gram Model And Smoothing Algorithm

In the technologically advanced 21st century, language modeling and machine translation, based on statistics, have been widely used. The English and Chinese translation software has already been studied. Translation software in the Latin languages, especially google's statistical language model and translation software, is mature and has been tremendously developed. In China, minority languages and Chinese translation work rather late. This paper studies mainly Uygur language model and achieve the Uygur language model by using the smoothing algorithm.This paper describes and compares five popular kind of international language models and smoothing algorithms in term of Uygur language characteristics, and then achieve the Uygur language model by using statistics-based N-gram language model and the five smoothing algorithm to. In the implementation process, the popular algorithm in the current research, such as Kneser-Ney and modKneser-Ney algorithm, will be described in details. The research of Uygur and Chinese sentence-alignment in the area of government report and references should be done before the model is applied, which is the basement of future Uygur and Chinese translation.The test of the system shows that the system has achieved good experimental results in the areas of government reports and documents. Thus, it is of a new help to the Uygur and Chinese language translation.