从知识库中发现新知识KDK（knowledge discovery in knowledgebase）是一个新课题，它的成功将直接作用于大型知识库的构建，并将为解决目前机器学习的瓶颈问题——知识获取起到重要影响。笔者的主要工作是：基于知识库中事实的KDK归纳结构、算法及其验证；基于知识库中规则的KDK归纳结构、算法及其验证。

Knowledge discovery in knowledge base (KDK) is a brand-new task. Its success will directly act on the construction of large knowledge base, and, at present, it is important to the solving of the bottleneck of machine study—discovering knowledge. The main work of this paper is： The inductive structure of KDK based on the facts in knowledge base, and its algorithm and experimental verification; The inductive structure algorithm of KDK for the rules in knowledge base and its experimental verification.