情感隐喻语料库构建与应用  

Construction and application of affective metaphor corpus

作　　者：张冬瑜[1] 杨亮[1] 郑朴琪 徐博[1] 林鸿飞[1] 

机构地区：[1]大连理工大学信息检索研究室,大连116023

出　　处：《中国科学：信息科学》2015年第12期1574-1587,共14页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:61572102;61402075);辽宁省自然科学基金(批准号:201202031;2014020003)资助项目

摘　　要：隐喻普遍存在于人类语言中,是对复杂而抽象的情感进行概念化的重要手段,也是情感计算的重要组成部分.无论面向计算机科学还是语言学的应用,情感计算都需要大规模、高质量的语料库作为资源支撑,才能进行更加深入、广泛的研究.因此,本文构建一个大规模汉英双语情感隐喻语料库,用以满足情感计算对语言资源建设的迫切需求.首先归纳情感隐喻语料库构建的理论框架,然后从概念界定、设计原则、语料收集、语料标注、质量监控5个方面分别进行说明,最后论述了该语料库在机器翻译、外语教学及情感分析方面的应用.Metaphors are commonly used in human languages, and are an important way of conceptualizing complicated and abstract emotions for humans. Meanwhile, metaphors also play an important role in affective computing. Affective computing requires a large amount of high-quality corpus for application to both computer science and linguistics in order to perform additional extensive research. Based on this idea, we construct a large-scale Chinese-English bilingual emotional metaphor corpus to meet the urgent demand for affective computing. First, we introduce the theoretical framework for the construction of emotional metaphor corpora. Then,we introduce the framework from five perspectives： concept definitions, design principles, data collection, data annotations, and quality control. Finally, we analyze the application of the constructed corpus on the area of machine translation and English teaching.

关 键 词：情感计算 情感隐喻语料库 机器翻译 外语教学 情感分析 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]