基于特征提取的电气专业英汉机器自动转换翻译系统设计  

Design of Automatic Conversion Translation System Based on Features Extraction

作　　者：王蕊[1] WANG Rui(Xi’an Si Yuan University,Xi’an 710038)

机构地区：[1]西安思源学院,西安710038

出　　处：《自动化与仪器仪表》2023年第2期237-241,共5页Automation & Instrumentation

基　　金：西安思源学院校级重点课程建设项目《英语听力》;西安思源学院横向课题合作项目《陕西开瑞建设工程项目管理有限公司高层领导队伍跨文化交际与商务英语会话能力提升的策略研究》(20211011)。

摘　　要：在电子专业词汇特征偏少的条件下,现有英汉自动翻译系统精准翻译输出的成功率明显降低。主要在于对电气专业语境特征的语义识别计算精度上过于偏低,导致整体翻译输出结果精准度偏低,影响系统翻译成功率。根据问题产生原因,采用特征提取算法,对硬件架构及其功能进行重构设计,使其能够适应电气专业翻译计算;在硬件架构设计基础上,软件采用逐层引入特征提取算法的设计思路,由翻译样本语义特征的初始模型、翻译语义特征提取、翻译特征相似度判定输出三部分构成完整的软件算法层,以此实现提升翻译准确度的效果。通过实验数据表明,采用特征提取算法后,设计系统整体翻译准确度相较对比系统有明显提升,同时具有提升效果稳定性好,系统收敛性好等特点,综合性能满足有效性设计要求,能够有效解决当前英汉自动翻译系统电气专业翻译精准度低、成功率低的问题。Under the condition of less vocabulary characteristics of electronic majors, the success rate of accurate translation output of the existing English-Chinese automatic translation system is significantly reduced.It is mainly that the calculation accuracy of electrical professional context characteristics is too low, which leads to the low accuracy of the overall translation output results and affects the translation success rate of the system.According to the causes of the problem, the feature extraction algorithm is reconstructed and designed to adapt to the electrical translation and functions. On the hardware architecture design, the initial model of translation sample semantic features, translation semantic feature extraction and translation feature similarity, so as to improve the translation accuracy.Through the experimental data show that the feature extraction algorithm, design system overall translation accuracy compared with contrast system has significantly improved, at the same time has good stability, good system convergence, comprehensive performance meet the requirements of effectiveness design, can effectively solve the current English-Chinese automatic translation system electrical professional translation low accuracy, low success rate.

关 键 词：特征提取 电气专业 英汉机器 自动转换翻译系统 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]