摘要
人名是文本中表达关键信息的部分,机器翻译系统在人名翻译上的表现对翻译质量具有重要影响。人名翻译作为命名实体翻译的重要部分,具有重要的研究意义。设计与实现基于Django的汉语-维吾尔语人名自动翻译网络服务接口,供机器翻译、关键词翻译等系统使用。采用TensorFlow框架搭建翻译模型并实现Django Web服务接口。实验结果显示,该翻译模型的准确率达到93.7%,服务接口翻译速度达到每秒139.72个字符,具有良好的并发性。
Human name is the key part of the text to express the key information.The performance of machine translation system in human name translation has an important influence on translation quality.As an important part of named entity translation,name translation has important research significance.In this paper,we design and implement a Django-based Chinese-Uighur personal name automatic translation network service interface for machine translation,keyword translation and other systems.we use tensorflow to build a translation model and implement Django web services.The experimental results show that the accuracy of the translation model is 93.7%,and the translation speed of the service interface is 139.72 characters per second,which has good concurrency.
作者
西热艾力·海热拉
艾山·吾买尔
王路路
Xireaili Hairela;Aishan Wumaier;WANG Lu-lu(College of Information Science and Engineering,Xinjiang University,Urumqi 830046;Xinjiang Laboratory of Multi-Language Information Technology,Xinjiang University,Urumqi 830046)
出处
《现代计算机》
2020年第18期19-22,32,共5页
Modern Computer
基金
国家自然科学基金项目(No.61662077,61262060)
新疆多语种信息技术实验室开放课题(No.2016D03023)。