基于句法和语义的英汉翻译记忆系统设计研究  

Research on E-C Translation Memory System Based on Syntax and Semantic

作　　者：董菊霞[1] DONG Ju-xia(Pingdingshan University,Pingdingshan 46700,Henan)

机构地区：[1]平顶山学院,河南平顶山467000

出　　处：《电脑与电信》2022年第8期84-89,共6页Computer & Telecommunication

摘　　要：为了提高英汉翻译系统译文精准性和作业效率,选取翻译记忆技术作为开发工具,提出一种基于句法和语义的英汉翻译记忆系统设计研究。该系统将句子拆分为多个短句,提取语义和句法,将其与记忆库中的翻译资源进行比较,计算相似度,判断译文句子是否达到直接使用标准。如果未能达到标准,采取人工修正的方式调整译文,并存储至记忆库中。系统测试结果显示,本系统译文的相似度在[0.95,1]之间句子更多。随着句子数量的增加,作业效率明显提升,且重复句子较多的情况下优势更为显著。In order to improve the translation accuracy and efficiency of E-C translation system,translation memory technology is chosen as a development tool,and a study on the design of E-C translation memory system based on syntax and semantic is proposed.The system divides sentences into short sentences,extracts semantic and syntax,compares them with translation resources in memory,calculates similarity,and judges whether the translated sentences meet the direct use standard.If the standard is not met,the translation is adjusted manually and is stored in memory.The results of the system test show that there are more sentences of which the similarities are between[0.95,1].With the increase of the number of sentences,the operation efficiency is obviously improved,and the advantage is more obvious when there are more repeated sentences.

关 键 词：相似度 记忆库 翻译记忆 语义 句法 

分 类 号：TP391.2[自动化与计算机技术—计算机应用技术]