基于语法纠错算法的自动机器在线翻译平台构建研究  

Research on the construction of automatic machine online translation platform based on syntax error correction algorithm

作　　者：王锐[1] WANG Rui(Shaanxi Energy Institute,Xi’an 710000,China)

机构地区：[1]陕西能源职业技术学院,西安710000

出　　处：《自动化与仪器仪表》2024年第9期280-284,共5页Automation & Instrumentation

基　　金：省级课题《高职院校大学英语课程开展“课程思政”的教学路径研究》(2023SZX396)。

摘　　要：英语作为全球交流的主要语言,其语法正确性尤为重要。鉴于许多非英语母语者在书写中常犯语法错误,此次研究致力于开发高效的自动机器在线翻译平台,以改善英语作文的语法纠错效率。采用基于Transformer框架和BERT预训练技术的系统,结合双编码器和混合注意力机制,以提升语法纠错精确性。结果显示,研究算法在CoNLL测试集上的精确率提高了9.8%,召回率提高了4.8%,F0.5值提高6.09%。在CLEC语料库的英语作文批改测试中,平均精确率达83.19%,召回率为71.40%,F1值为77.02%。可以看出,研究开发的自动机器在线翻译系统在语法纠错方面显示了显著的优势。该研究为非英语母语者提供了一个高效的语法纠错工具,有助于提升英语书写能力,促进全球交流和理解。As the main language of global communication,the grammatical correctness of English is particularly important.Many non-native English speakers often make grammatical mistakes in their writing.The aim of this research is to develop an efficient automated online translation platform to improve the efficiency of grammar correction in English compositions.A system based on the Transformer framework and BERT pre-training technology combines dual encoders and hybrid attention mechanisms to improve grammar error correction accuracy.The results show that the accuracy of the proposed algorithm on the CoNLL test set is increased by 9.8%,the recall rate is increased by 4.8%,and the F0.5 value is increased by 6.09%.In the English composition correction test of CLEC corpus,the average accuracy rate is 83.19%,the recall rate is 71.40%and the F1 value is 77.02%.It can be seen that the automatic machine online translation system has shown remarkable advantages in grammar correction.The study provides an effective grammar correction tool for non-native English speakers,which can help improve written English and promote global communication and understanding.

关 键 词：语法纠错算法 自动机器 在线翻译 BERT 英语 

分 类 号：TP131[自动化与计算机技术—控制理论与控制工程]