摘要
本文回顾了自动文摘技术的发展历史,介绍了三种主要的文摘方法,即基于文本物理信息分析的方法、基于自然语言理解的方法以及基于文本结构分析的方法。同时指出了自动文摘技术目前存在的问题,包括摘要内容冗余的问题、摘要的语言缺乏连贯性的问题等,并探讨了文摘技术未来的发展方向。
This paper introduces the history of automatic abstracting, and three kinds of abstracting methods. The first method is based on text physical information analysis, the second method is based on natural language understanding and the third method is based on text structure analysis. At the same time, the paper points out the problems in automatic abstracting, including the coutent redundancy, the language coherence, and some others. In the end, the future of the abstracting technique is discussed.
出处
《情报科学》
CSSCI
1998年第1期63-69,共7页
Information Science
关键词
情报科学
自然语言处理
自动文摘
文摘
计算机
Information science Natural language processing Automatic abstracting