摘要
随着社会对英语翻译机器人同声传译功能要求的增加,许多翻译机器人的同声传译功能亟需提升。针对这一问题,研究利用基于密度的聚类算法对隐马尔科夫模型进行改进,并在改进算法的基础上构建新型的翻译机器人同声传译模型。对改进后的隐马尔科夫模型算法进行对比实验发现,改进算法在数据集中的最高准确率和召回率分别为0.913和0.912,显著优于对比算法。之后对同声传译模型的性能进行实证分析,结果显示,提出的同声传译模型的翻译耗时和翻译稳定性分别为78.3 s和0.79,显著优于传统模型。上述结果说明,研究提出的改进隐马尔科夫模型算法和基于该算法的同声传译模型均具有较优的性能,故将其应用于英语翻译机器人同声传译中可促进同声传译领域的发展。
With the increasing demand from society for the simultaneous interpretation function of English translation robots,the simultaneous interpretation function of many translation robots urgently needs to be improved.To address this issue,a density based clustering algorithm is studied to improve the hidden Markov model,and a new simultaneous interpretation model for translation robots is constructed based on the improved algorithm.Comparative experiments on the improved hidden Markov model algorithm showed that the highest accuracy and recall rates of the improved algorithm in the dataset were 0.913 and 0.912,respectively,which were significantly better than the comparison algorithm.Subsequently,an empirical analysis was conducted on the performance of the simultaneous interpretation model,and the results showed that the proposed simultaneous interpretation model had a translation time and stability of 78.3 seconds and 0.79 seconds,respectively,significantly better than traditional models.The above results indicate that both the improved hidden Markov model algorithm proposed in the study and the simultaneous interpretation model based on this algorithm have excellent performance.Therefore,its application in English translation robot simultaneous interpretation can promote the development of the field of simultaneous interpretation.
作者
陈琳
CHEN Lin(Xianyang Normal University,Xianyang Shaanxi 712000,China)
出处
《自动化与仪器仪表》
2024年第5期145-149,共5页
Automation & Instrumentation
基金
咸阳师范学院校级陕西教育学会教改项目《基于“1-2-1-1”教学理念的<英语公众演讲>课程思政之路径与实践》(2021Y032)。
