中文信息自动分类用知识库的设计与构建

Design ＆ construction of knowledge database for automatic classification in Chinese

在计算机智能处理技术远未成熟的情况下,基于概念语义网络的自动分类采用知识库技术,仍将是一种实用的选择。本项研究根据分类语言、主题语言、自然语言三者兼容互换的原理,以众多标引员的主题标引和分类标引的经验,即文献数据库实体中大量存在的文献分类号和主题词双重标引数据为基础,建立一个以《中图法》为基础的分类知识库——分类法与主题词表对照数据库。论文对构建分类知识库的思路、步骤及主要技术,包括关联度测度方案、标引词模式匹配、新词增补等进行了讨论。

This project of automatic indexing & classification in Chinese is based on knowledge database.According to the principles of compatibility among the Library Classi-fication,Thesaurus and Keywords,the knowledge database is constructed by the indexingdata of records of bibliographic database.So it is a concordance database between classnumber of Chinese Library Classification and descriptors of Chinese Thesaurus,The articlediscusses the technical lines,steps and methods of design and construction of knowledgedatabase,including provide the location weighting algorithm for concept mining and a newmethod to enhance the ability of recognition of the unregistered words.