一种改进的分布式关联规则挖掘算法  

An Algorithm for Distributed Association Rules Mining

作　　者：曹文梁[1] 

机构地区：[1]东莞职业技术学院计算机工程系,广东东莞523808

出　　处：《东莞理工学院学报》2014年第3期35-40,共6页Journal of Dongguan University of Technology

摘　　要：数据挖掘本质上是一种新的商业信息处理技术,通过对数据进行统计、分析、综合和推理,发现数据间的关联性、未来趋势以及一般性的概括知识,用以指导高级商务活动。由于需要,对数据间的关联性的数据挖掘算法模型已成为数据库及相关领域的一个研究热点,给出了一种基于分布式数据库的挖掘模型及其相应的一种有效的挖掘算法,其由若干个站点集合而成,各个站点拥有各自的数据库、中央处理机、客户端,以及各自的局部数据库管理系统,依靠通讯网络连接。采用购物篮分析式关联规则,将各个数据库文件的数据合成,从而得到挖掘结果,对挖掘的方法又进一步挖掘,即将不满足条件的规则重新传送到各分布式站点上进行更加精确的挖掘处理,从而避免了频繁的网络通讯。该算法在减轻网络频繁的通讯负担,体现并行计算以及异构数据挖掘方面具有独特优点。Data mining is a new business information processing technology .Through microcosmic , medium even macro-scopic statistic, analysis, synthesizing and inference , it can find the co-relation between data, trends and generality of knowledge in order to guide senior business activities .Data mining algorithm model has been a hot topic in the database and relational field .This paper introduces a mining model based on distributed database and presents an efficient mining algorithm .It includes a few stations , and each station has individual database , CPU, client-side and the manage system of local database , connected by communication internet.Through shopping basket analysis association rule and integrating every database file , it gets a mining result, and then makes a further mining upon the mining method , transports the rules which are not fit with the requirements back to each distributed station to make a more accurate mining process , thus avoiding the frequent internet communication .This algorithm can reduce fre-quent communication burden , owning an distinguishing virtue in parallel arithmetic computing and asynchronous operation＆amp;hetero -geneous mining .

关 键 词：分布式 关联规则 挖掘模型 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]