摘要
隐喻是一种在语篇中普遍存在的语言现象.用计算机模拟隐喻这一带有创造性的人类语言,对隐喻计算化乃至自然语言处理的研究都具有十分重要的意义.本文介绍了一个汉语隐喻自动生成系统.在对大规模语料库进行统计研究的基础上,建立基于向量空间的模型,进行了词义相似度和词义相关度的计算,并构建了支持三种输入模式的隐喻句生成系统.实验结果表明,该系统初步实现了生成汉语隐喻句的目的,并具有一定的自学习功能.
Metaphor is commonly used in our daily speech and writing. The computer simulation of human metaphors – a typical phenomenon of human creativity – will not only contribute to metaphor computational research, but will also enrich the studies of natural language processing. This paper introduces a Chinese metaphor generation system. After statistical processing of a large-scale Chinese corpus, we propose a vector-space-based model, and combine results of lexical similarity computation and lexical relevancy computation. Then we built a Chinese metaphor generation system that supports three input patterns. As isshown by a certain number of tests, the system constructed on the basis of the computing model designed in this paper is basically capable of generating metaphorized Chinese sentences, and has a certain level of automatic learning function.
出处
《心智与计算》
2007年第1期133-141,共9页
Mind and Computation
基金
国家自然科学基金(60373080)资助
关键词
隐喻
自然语言生成
计算模型
系统实现
metaphor
natural language generation
computational model
system implementation
