摘要
学术英文写作对于母语非英语的(English as a Second Language,ESL)学者而言是一项挑战.现有的ESL写作辅助系统均基于连续单词(n-gram)在语料库中出现的频率来构建,而忽略了语义信息对于ESL写作的作用.文中提出基于语义的写作辅助方法,支持按照词性、搭配类型和概念扩展等语义条件来检索短语和自动推荐搭配.此外,为了方便英语知识有限的ESL学者利用语义条件进行检索,文中还设计了易于理解的短语检索界面.该界面提供包含语义修饰符的查询表达式,并可根据用户输入的部分查询表达式在线提示和补全.最后,为验证语义方法的有效性,文中实现并部署了ESLWriter系统.用户实验表明,ESLWriter可以有效地推荐搭配并提供短语检索结果,其查询界面直观易用;ESL学者通过混合使用自动搭配推荐功能和短语检索功能,写作质量和写作信心得到有效提高.这些结果证明了语义信息在ESL写作辅助中的重要作用.
Academic writing in English is a challenge for scholars who use English as a Second Language(ESL).Existing English writing assistant systems are usually based on the frequency of consecutive words(n-gram)in the corpus,but ignore the role of semantic information in ESL writing.In this research,we propose a semantic approach to support ESL writing,which allows users to retrieve corpus phrases and automatically makes collocation suggestions using a number of semantic constraints,including part-of-speech,collocation type,and conceptual expansion.Furthermore,to address the limited linguistic knowledge of general ESL users,we design an easy-to-understand query interface for phrase retrieval,which supports query with semantic modifiers,and can automatically suggest query expressions according to users' input.To examine the validity of our semantic approach,we implement a real running system ESLWriter.User experiments show that ESLWriter is helpful in phrase retrieval and collocation suggestion,and the query interface for end users is intuitive and easy-to-use.Participants can significantly improve their writing quality and confidence by mixed use of the two functions.These results confirm the effectiveness of incorporating semantic information into ESL writing system.
出处
《计算机学报》
EI
CSCD
北大核心
2016年第4期822-834,共13页
Chinese Journal of Computers
关键词
ESL写作
语义
查询表达
搭配
语料库
社交网络
慕课
人机交互
ESL writing
semantics
query expression
collocation
corpus
social networks
massive open online course(MOOC)
human computer interaction