摘要
本文在简要介绍篇章的向量空间模型表示的基础上,讨论了基于段间相似度和关系图进行篇章分析的方法,包括:结构分析,主题分析和聚类,浏览与跳段阅读.最后讨论所存在的主要问题及进一步改进的意见.
The vector space model (VSM) of text in arbitrary subject area is outlined, and text relation maps among paragraphs are obtained by similarity analysis. Based on the relation maps, several approaches for text analysis are described, which include structure analysis, theme analysis and chustering, theme browsing and text traversal. At last, some of the main problems are discussed and further improvements are suggested.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1997年第2期112-117,共6页
Pattern Recognition and Artificial Intelligence
基金
国家高技术863-306主题
国家自然科学基金
关键词
向量空间模型
信息检索
篇章理解
自然语言处理
Vector Space Model, Information Retrieval, Text Understanding, Natural Language Processing