借助汉语以义索词的英语用法检索

Locating idiomatic English expressions in parallel corporabased on expanded Chinese queries of Chinese seed words

有了一定的表达意图却找不到合适的英语表述方式,是中国学生经常面临的问题。借助句对齐的英汉平行语料库,能够从汉语出发检索出与之对应的英语实例。然而当前基于语言形式的语料库检索,其检索结果必须与用户输入的特定检索项匹配。为尽可能全面准确地获取地道的英语表述,本文借鉴信息检索的查询扩展策略,利用同义词词林等语义资源,介绍一种基于汉语同义表述的意义检索,在检索过程同时利用浅层语法分析,使检索精度得以保证。

Finding appropriate English expressions is a stumbling block for Chinese learners of English It is possible to retrieve authentic English examples by searching for their equivalent Chinese terms in a sentence-aligned English-Chinese parallel corpus However,as the nature of current corpus retrieval strategy is based on exact matching between query terms and results returned,most English sentences,which can meet a searcher’s information needs but differ from the query terms in form,will not be included in the search results We propose a novel approach to deal with this problem To obtain more desired English expressions,inspired by the query expansion strategy commonly used in the information retrieval community,a thesaurus and an English-Chinese equivalent word list are employed to expand initial search terms,which brings about an enhancement on recall rate Meanwhile,the corpus,preprocessed by a shallow parser,enables grammatical relations to be integrated into corpus queries,also resulting in additional improvement in precision.