面向小词典的高效英汉双语语料对齐算法  

Efficient Small-dictionary-oriented Algorithm for Alignment of English-Chinese Parallel Corpora

作　　者：熊伟[1] 陈蓉[1] 刘佳[1] 徐淼[1] 于中华[1] 

机构地区：[1]四川大学计算机学院,成都610064

出　　处：《计算机工程》2007年第13期210-212,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60473071);高等学校博士学科点专项科研基金SRFDP资助项目(20020610007);四川大学计算机学院青年基金资助项目

摘　　要：双语语料自动对齐是自然语言处理的一个重要研究课题。该文针对基于词典译文的英汉句子对齐算法存在的缺点,提出了面向小词典的高效英汉句子对齐算法,该算法在小词典的情况下仍具有较高的准确率,效率比传统算法提高近一倍。通过理论分析、对比实验可知,该算法是有效的。Automatic alignment of parallel corpora is an important research subject in natural language processing area. An efficient small-dictionary-oriented algorithm for alignment of English-Chinese parallel corpora is proposed in this paper in order to improve the efficiency of translation-based approach to parallel corpora alignment, and enhance its precision when the dictionary used is small. The algorithm has competitive precision under the condition of small dictionary, and doubles the efficiency of the traditional approach. The validity of the algorithm is argued through theoretical analysis and experiments.

关 键 词：机器翻译 局部对齐 补偿 双语语料 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]