摘要
本文介绍了语料库研究的一些特点 ,并以交互式多策略的思想为背景 ,对 IHSMTS系统 CBMT翻译引擎中模式库的设计与实现问题进行了探讨 ,提出了面向对象的分类模式库的设计思想 ,并对模式的表示、模式库的组织进行了阐述 ,方便了模式库检索 ,添加等操作的实现。同时介绍了近似模式匹配算法 ,从句法功能相似的角度抽取出所比较事例的功能词和句法特征 ,作为检索模式库和相似度计算的依据。
This paper first introduces some aspects of the corpus-based research area and then describes the design and implementation of the Knowledge Base of the Corpus-Based Machine Translation (CBMT) Engine of Interactive Hybrid Strategies Machine Translation System (IHSMTS) as well as the idea of Object-Oriented Classified Pattern Base for easy retrieval and appending of patterns.An algorithm of similar pattern matching is proposed,which is used to retrieve the best matching case and measure their similaity according to the obtained function words and syntax features of compared cases.Finally,it presents the process of extracting knowledge from the pattern base and refining it.
出处
《微型电脑应用》
2002年第3期5-9,共5页
Microcomputer Applications
基金
国家自然科学基金资助
