汉译英机器翻译错误类型统计分析——以外宣文本汉译英为例

Statistical analysis on types of C-E machine translation errors:A case study on C-E translation of publicity text

通过平行测评国内常用的机器翻译平台,研究现有机器翻译平台在汉译英转换上的不足之处。通过搜狗、有道、谷歌和百度四个机器翻译平台对选取的案例文本进行汉译英转换,并通过和人工译文的比较,统计并分析国内机器翻译平台在汉译英转换中易出现的典型机器翻译问题。经对China Daily网站上一则演讲稿汉译英案例文本机器翻译译文和人工译文的平行比较,发现所测试四个机器翻译平台在汉译英转换上,存在词和句层面翻译错误、漏译、大小写格式错误等共性化的机器翻译问题。这些翻译问题有待机器翻译平台的完善和译者的译后编辑来处理。

The study aims to analyze deficiencies of existing machine translation platforms in Chinese-English translation by means of parallel evaluation of common machine translation platforms in China.Four common machine translation platforms namely Sogou,Youdao,Google and Baidu were used to carry out Chinese-English translation of the selected case.Through comparing C-E machine translation with the human translation,the study counts and analyzes typical problems that easily occur in the C-E machine translation by machine translation platforms.Based on the parallel comparison between the C-E machine translation and human translation of a speech draft case on the website of China Daily,it is found that the tested four machine translation systems have common limitations in C-E machine translation including syntactic translation errors,omissions and format errors,which need to be further solved by improvement of machine translation systems and post-translation editing.