基于科技文献资源的跨语言信息检索系统研究

Research on the Cross-language Information Retrieval System Based on Scientific and Technical Literature Resources

本文分析了跨语言信息检索的现状,介绍和描述跨语言信息检索的主要问题,并针对万方数据和NSTL的现有资源,构建出基于科技文献的跨语言信息检索原型系统,此原型系统通过字典查询得到检索词翻译,应用Google在线翻译进行标题、摘要翻译,并预留中国科学院自动化所提供的翻译引擎扩展。跨语言信息检索强调检索结果翻译的可读性,帮助用户理解基本大意,不刻意追求准确性,力求在应用过程中促进机器翻译技术的进步和发展,辅助科研。

The status quo of Cross-Language Information Retrieval （CLIR） is analyzed, and the main problems in CLIR is described. Based on the current resources of Wanfang Data and NSTL, this paper constructs a CLIR prototype system based on scientific and technical literature resources. The system gets the keyword translation by consulting dictionaries, and uses Google translator to translate titles and abstracts. It reserves the extensions for translation engines of the Institute of Automation of the Chinese Academy of Science. CLIR emphasizes the readability not the accuracy of the search results, helps users understand the main ideas, and strives to improve and develop machine translation technologies in application process.