在基于知识的系统中,如何建立知识库是建立知识系统的核心问题。本文讨论了深层知识系统以及目前专家系统开发过程中的若干问题。提出了为克服这些问题的两种方法——概念匹配和经验学习,并描述了一种适于概念匹配和经验学习的知识存储结构。 In the knowledge-based system, how to build a knowledge base is the core issue of knowledge system. This article discusses several issues in deep knowledge systems and the current development of expert systems. Two methods for overcoming these problems, concept matching and experience learning, are proposed, and a knowledge storage structure suitable for concept matching and experiential learning is described.