基于旋进原则的领域驱动数据挖掘方法研究

Research on Domain-Driven Data Mining Based on SPIRPRO Principles

领域驱动数据挖掘（Domain-Driven Data Mining,DDDM）是数据挖掘中的新方法,目的是挖掘用户感兴趣、可行动的知识,与传统的数据挖掘过程CRISP-DM相比,DDDM是基于约束的、人机结合、往复循环、不断逼近目标、深层次的知识发现过程。本文在剖析DDDM挖掘过程难度自增殖的复杂性特点的基础上,提出基于旋进原则的系统方法进行挖掘,提出从领域知识、数据和技术三个方面进行旋进挖掘,以使得挖掘出来的知识更满足用户在现实世界活动中对知识的需求。最后,文中结合名老中医学术思想挖掘进行了实证研究,开发了基于语义的Apriori算法,并用抽象语义、分类语义和组合语义结构化表示领域知识,挖掘结果表明基于旋进原则的DDDM方法是可行和有优势的。

Due to the disadvantage of traditional data mining method,a new method for mining interesting and actionable knowledge,called Domain-Driven Data Mining（DDDM）,was introduced.After analyzing the feature of self—increasing difficulty system of data mining process,we proposed a system method for Domain-Driven Data Mining based on SPIRPRO Principles,and analyzed from the following three aspects：domain knowledge,data and technique with the method of SPIRPRO Principles in order to meet user mining requirement.At last,a case related to Traditional Chinese Medicine（TCM） mining was introduced,a new priori-based on semantic was developed and proposed abstract semantic,classified semantic and compound semantic to formed domain knowledge.Result has shown that Domain-Driven Data Mining based on SPIRPRO principle is feasible and advantageous.