Research Of Bilingual Sentence Alignment Served The Chinese-Uyghur Machine Translation System

Since the 80th of the last century, the fast development of computer science and the quick availability and application of computer technology in the research filed of linguistics make the corpus linguistics become the main stream of the research, in which parallel corpus and comparative corpus have became a main focus in the research of corpus linguistics.The research work of bilingual parallel corpus mainly includes construction ,alignment and tagging and so on, in which alignment of different level is the important core task, alignment is not only the essential prerequisite for further using parallel corpus to acquire some linguistic knowledge ,but also the important preliminary procedure of machine translation using of bilingual knowledge.The thesis begins with the problem of alignment, introduces some basic methods of alignment at present, then specifically analyzes the problem of Chinese -Uyghur sentence alignment.on this basis, bringing up the idea that using the method based on the anchor to accomplish Chinese-Uyghur bilingual sentence alignment ,analyzing the result of statistic experiment deriving form Chinese-Uyghur corpus which makes some improvements on weight coefficient of this method that will be more suitable for Chinese -Uyghur bilingual alignment technology.At last, the thesis introduces the method of anchor sentence pairs automatic proofreading which improves the correct rate of alignment.