| 基于机器翻译的专利术语翻译获取方法研究 |
| |
| 引用本文： | 何彦青,刘建辉,屈鹏,李颖,徐红姣.基于机器翻译的专利术语翻译获取方法研究[J].图书情报工作,2014,58(19):25-30. |
| |
| 作者姓名： | 何彦青 刘建辉 屈鹏 李颖 徐红姣 |
| |
| 作者单位： | 1. 中国科学技术信息研究所;
2. 石家庄经济学院 |
| |
| 基金项目： | 本文系国家自然科学基金“面向专利文献的统计机器翻译语境分析”（项目编号：61303152）和中日国际合作项目“面向科技文献的日汉双向实用型机器翻译合作研究”（项目编号：2014DFA11350）研究成果之一。 |
| |
| 摘 要： | 鉴于专利术语的翻译要求高度的准确性和专业性，而专利术语的自动获取翻译对于机器翻译、词典自动编纂、跨语言信息检索等自然语言处理具有重要的实用价值，从双语的专利摘要中分别抽取术语，之后融合多术语识别方法，采用规则翻译和统计机器翻译来动态地辅助词汇化方法进行术语对齐，以期尽可能多地在双语的专利文献中获取准确的专利术语翻译对。在专利文摘中进行实验验证的结果是：专利术语翻译对的准确率达到80%。
|
| 关 键 词： | 机器翻译 术语识别 术语对齐 |
| 收稿时间： | 2014-07-24 |
Study on Acquisition Method of Patent Term Translation Based on Machine Translation |
| |
| He Yanqing,Liu Jianhui,Qu Peng,Li Ying,Xu Hongjiao.Study on Acquisition Method of Patent Term Translation Based on Machine Translation[J].Library and Information Service,2014,58(19):25-30. |
| |
| Authors: | He Yanqing Liu Jianhui Qu Peng Li Ying Xu Hongjiao |
| |
| Institution: | 1. Institute of Scientific and Technical Information of China, Beijing 100038;
2. Shijiazhuang University of Economics, Shijiazhuang 050031 |
| |
| Abstract: | Patent term translation requires high degree of accuracy and professionalism. Research on automatically obtaining patent term translation has important practical value on natural language processing tasks such as machine translation, compilation of bilingual dictionaries, cross-language information retrieval. This article extracts terms respectively from the bilingual patent abstracts, then combines multiple methods of terms recognition, finally uses rule-based machine translation and statistical machine translation to dynamically help lexical method realize term alignment, so as to obtain more accurate term translation as much as possible. The experiment on patent abstract shows that the precision of term translation pairs reaches 80%. |
| |
| Keywords: | machine translation term recognition term alignment |
|
| 点击此处可从《图书情报工作》浏览原始摘要信息 |
| 点击此处可从《图书情报工作》下载免费的PDF全文 |
