N-gram统计模型在机器翻译系统中的应用

Statistical N-gram Method Used in Machine Translation System

文章提出了N-gram模型在机器翻译系统中的几个应用。模型是在语料库的基础上统计连续几个词的出现概率,以此来筛选翻译过程中的侯选元素,并可以对译文的语序进行纠正。由于此种方法是建立在语料库的基础之上的,从而具有真实可靠和实时等特点。实验表明,这种方法具有良好的性能,且与被处理的语言无关。

This paper introduces the statistical N-gram method,and used it in the Chinese-English Machine Translation System.Based on the parallel corpora,it performs statistics about consequent words.According to the statistic results,it can eliminate the wrong translations and rearrange the word sequence of the translation results.Experiments show that this method has good performance.