知识发现中可继承性问题的研究

On the Inheritability Problem in Knowledge Discovery in Database

摘要: 提出知识发现中的可继承性问题，通过对知识发现过程和挖掘算法形式化描述和分析，抽象出各个阶段的形式联系及其约束条件，在此基础上提出初等知识的概念．在引入初等知识后，对传统的挖掘算法、增量式挖掘算法、可继承性挖掘算法进行形式化描述和比较，得出如下结论：可继承性挖掘算法能够有效的提高数据集变化、参数变化情况下的数据挖掘效率．

Abstract: The inheritability problem in the KDD(knowledge discovery in database)process is presented.The KDD process,the DM(data mining)algorithms,and the relation between them are formally described and deeply analyzed.The basic concept named PK(primary knowledge)is put forward.Then the traditional DM algorithm,incremental DM algorithm and the inheritable DM algorithm are formally described and compared.Finally,it is concluded that the inheritable DM algorithm can improve the efficiency of data mining with the variable data set and parameters.