集成概念空间与潜在语义索引的文本聚类检索研究  

作　　者：韩毅[1,2,3] 张克菊[1,2] 金碧辉[1] 

机构地区：[1]中国科学院国家科学图书馆,北京100190 [2]中国科学院研究生院,北京100190 [3]西南大学计算机与信息科学学院,重庆400715

出　　处：《情报理论与实践》2009年第6期102-105,共4页Information Studies:Theory & Application

摘　　要：信息环境的异构性、动态性与海量性使传统基于自然文本的信息检索方法与技术面临极大挑战,集成概念空间理论与潜在语义索引技术能为这种困境提供一些解决方案。在分析概念空间内涵与特征的基础上,利用潜在语义索引原理讨论了概念提取方法、同义词近义词处理方法及基准向量的生成方法,分析了网络条件下基于概念空间的文本分类、聚类检索基本机制,最后给出了完善概念空间的自学习机制。The traditional information retrieval method based on natural text has confronted with great challen- ges in the heterogeneous, dynamic and magnanimous information environment, and the integration of concept space and latent semantic index may be a solution to this predicament. Based on the analysis of the connotation and char- acteristics of concept space, this paper uses the principle of latent semantic index to discuss the methods of extrac- ting concept, dealing with synonym and near synonym, and creating normal vector. The basic mechanism of text classification and clustering retrieval based on concept space under the network conditions are analyzed. Finally, the self - learning mechanism of improving concept space are given.

关 键 词：概念空间 潜在语义索引 文本检索 聚类检索 

分 类 号：G354[文化科学—情报学]