汉英机器翻译中从句复杂度的错误分析  

Error Analysis of Clause Complexity in Chinese-English Machine Translation

作　　者：成洁[1] CHENG Jie(Department of Basic Science Shanxi Institute of International Trade and Commerce, Xi’an 712046, China)

机构地区：[1]陕西国际商贸学院基础课部,陕西西安712046

出　　处：《微型电脑应用》2020年第9期51-53,69,共4页Microcomputer Applications

基　　金：2019陕西省教育厅专项科研计划项目(19JK0090);2019陕西省社会科学基金项目(2019M032)。

摘　　要：误差分析是衡量机器翻译系统性能的重要手段。在从句复合模型的分析框架下,对从句复合级错误进行了分析,并从汉英语法差异的角度对错误进行了解释。统计结果表明,复句级错误发生的概率相对较高,而且错误的发生在不同的体裁上是不同的,并且与复句的长度呈正相关。Error analysis is an important method for performance estimationof machine translation systems.This article analyzes clause-level compound errors under the analysis framework of clause-complex model,and explains the errors from the perspective of differences between Chinese and English grammars.Statistical results show that the probability of complex sentence-level errors is relatively high,and that errors occur in different genres,and are positively related to the length of complex sentences.

关 键 词：机器翻译 从句复杂度 从句-复杂度级别错误 汉译英 

分 类 号：H315.9[语言文字—英语] TP391.2[自动化与计算机技术—计算机应用技术]