机器翻译评价系统研究与设计  

Research and Design of Machine Translation Evaluation System

作　　者：林颖 吾守尔·斯拉木[1] Lin Ying;Wu Shou Slam(Colleges of Information Science and Engineering,Xinjiang University,Urumqi 830046,Xinjiang;Colleges of Information Science and Engineering,Xinjiang University Urumqi 830046,Xinjiang)

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046

出　　处：《信息通信》2020年第5期26-28,共3页Information & Communications

基　　金：新疆维吾尔自治区重点实验室专项(2015KL013);国家自然科学基金项目(U1603262);国家973计划项目(2014CB340506)。

摘　　要：伴随着现代的计算机技术和研究的快速发展,机器翻译技术的研究已经在全球的研究内容范围内产生了新的热点。自动翻译评估技术不仅仅可以迅速并且能够准确的估计出来机器翻译双语语料内容的质量,人们还能够将机器翻译内容的质量评估的数据信息结果能够做为系统反馈的数据信息来调节其机器翻译系统的数据反馈的参数,从而使得机器翻译系统可以效果达到最优。因此,对于机器翻译内容评价的研究不仅具有日常生活应用的实际价值,而且在某种程度上是帮助促进了翻译理论研究的进一步发展。文章的主要工作包括以下部分:构建完成使用机器翻译得到的双语语料的语料库,并且训练词向量模型。利用系统进行文本内容相似度计算,并对此机器翻译评估技术进行实验,验证此系统有效性。With the rapid development of modern computer technology and research,the research of machine translation technology has generated new hotspots within the scope of global research.Automatic translation evaluation technology can not only quickly and accurately estimate the quality of machine translation bilingual corpus content,people can also use the data information results of machine translation content quality evaluation as system feedback data information to adjust its machine translation system The parameters of the data feedback,so that the machine translation system can achieve the best effect.Therefore,the research on machine translation content evaluation not only has practical value in daily life applications,but also helps to promote the further development of translation theory research to a certain extent.The main work of this paper includes the following parts:one is to summarize the existing methods of machine translation evaluation;the second is to build a corpus of bilingual corpora obtained using machine translation and train the word vector model.The third is to use the system to calculate the similarity of text content,and experiment with this machine translation evaluation technology to verify the effectiveness of this system.

关 键 词：机器翻译评价 文本相似度 软件设计 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]