CAJ | 학술논문

研究针对汉语（二语）教学会话体语料语义功能的检索需求，基于汉语国际教育动态语料库，探讨了一种语料库语言信息标注框架。围绕日常口语交际的主要目的以及教学范围，提出19类对答结构，刻画了引发语与应答语的基本形式。以此为基础，开展对答结构的自动识别算法研究，选取问候、感谢、祝贺、赞扬、介绍5个类别进行试验，实验测试在准确率和召回率上均取得较好的成绩。对答结构的标注框架对于会话体语料相关表达式的抽取具有较好的适应性，自动识别算法可用于语言信息的自动抽取以及语料库扩展应用软件研发等。
연구침대한어（이어）교학회화체어료어의공능적검색수구，기우한어국제교육동태어료고，탐토료일충어료고어언신식표주광가。위요일상구어교제적주요목적이급교학범위，제출19류대답결구，각화료인발어여응답어적기본형식。이차위기출，개전대답결구적자동식별산법연구，선취문후、감사、축하、찬양、개소5개유별진행시험，실험측시재준학솔화소회솔상균취득교호적성적。대답결구적표주광가대우회화체어료상관표체식적추취구유교호적괄응성，자동식별산법가용우어언신식적자동추취이급어료고확전응용연건연발등。
With a consideration of the retrieval requirement of the semantic functions for the conver- sation style in the dynamic corpus of the international Chinese education, this paper explores a tagging frame for the corpus-based language information. Focusing on the major goals in daily conversations and the relevant teaching domains, this frame brings up nineteen categories of question-answer struc- tures and describes the basic forms of questions and answers. With this, it discusses an automatic rec- ognition algorithm for the question-answer structures by selecting five categories such as greeting, thanking, congratulating, praising, and introduction as the research objects. The results show that it is fairly satisfactory in terms of accuracy and recall rate as well as applicable in the automatic retrieval of the language information and related studies.