基于分治策略的关联规则增量式更新技术  

AN INCREMENTAL UPDATING ALGORITHM TO MINE ASSOCIATION RULES BASED ON DIVIDE AND CONQUER METHOD

作　　者：林振民[1] 叶仰明[1] 陈锦秀[1] 

机构地区：[1]厦门大学计算机系,厦门361005

出　　处：《模式识别与人工智能》2002年第1期103-107,共5页Pattern Recognition and Artificial Intelligence

摘　　要：关联规则的发现是数据挖掘的一个重要问题.当最小支持度和最小可信度变化时,如何利用原来的挖掘结果快速得到新的频繁项目集和关联规则在实践上具有重要的意义.本文提出一种基于分治策略的关联规则增量式更新技术,充分利用了原来获得的结果,同时在开采过程中使用对某些频繁项目集加标志tag的方法来提高开采效率.Mining association rules is an important issue in data mining. When the minimum support or the minimum confidence varies, it is of great use employing the once - gained association rules to mine new ones. In this paper we present an incremental updating algorithm to mine association rules based on divide and conquer method. In the algorithm, we add an tag to some frequent item sets to speed up the mining process. The experiments show that the presented algorithm is of great efficiency.

关 键 词：关联规则 频繁项目集 增量式更新 分治策略 数据挖掘 数据库 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]