基于主题地图的异构知识集成*

大型企业和科研机构所处理的信息通常是分布式的，这些信息除了分散在不同的数据库中之外，还以类型各异的文档形式独立存在。为了抽取信息资源中隐含的知识，需要探索数据库及存档文件，把其中有用的知识单元一一挑选出来。为了从整体上对抽取的知识加以把握，还需要把这些知识片段有机地整合起来。本文提出基于主题地图的知识集成系统TMKIS，将信息资源的本体表示规范、存储方式、自动抽取方式、合法性验证以及浏览方式有机地结合起来，利用主题地图技术，处理异构的信息资源，实现异构知识集成的目标。

Heterogeneous Knowledge Integration Based on Topic Maps

Information handled by companies and institutions is usually distributed,which is stored in databases in different places or kept in archives of different types.In order to extract implicit knowledge from the information resource,databases and archives have to be explored.When the useful knowledge slice is found,the knowledge segment still needs to be integrated to obtain a general overview.A kind of Knowledge Integrated System based on Topic maps(TMKIS) is developed to integrate the specification of ontology,the way of storage,auto-extraction,validation and navigation for information resource.According to the technology of Topic maps,heterogeneous information resource is handled by TMKIS and the integration of heterogeneous knowledge is achieved.