基于矩阵压缩的加权关联规则挖掘算法  

A Weighted Association Rules Mining Algorithm Based on Matrix Compression

作　　者：肖红光[1] 邓国群 谭雯 向德华 李宁 XIAO Hong-guang;DENG Guo-qun;TAN Wen;XIANG De-hua;LI Ning(School of Computer&Communication Engineering,Changsha University of Science&Technology,Changsha 410114,China;Hunan Institute of Metrology and Test,Changsha 410014,China)

机构地区：[1]长沙理工大学计算机与通信工程学院,湖南长沙410114 [2]湖南省计量检测研究院,湖南长沙410014

出　　处：《测控技术》2018年第3期10-13,共4页Measurement & Control Technology

基　　金：国家自然科学基金青年科学基金项目(41201468);国家公益性行业科研专项(201510003-5)

摘　　要：关联规则挖掘作为近年来的研究热点之一,其经典算法Apriori算法因需要多次扫描数据库且会产生大量候选项集,严重影响了关联规则的挖掘效率。在此基础上提出了一种基于矩阵压缩的加权关联规则挖掘算法,只需扫描一次数据库,并将其转换为0-1矩阵,根据相关性质对矩阵进行压缩,从而降低了算法执行过程中的计算量;同时,考虑到项目的重要性,采取加权的方法,用求概率的方式设置项目属性的权值,同Apriori算法相比,本算法在挖掘过程中能直接查找高阶频繁项集。实验结果表明,本算法能有效提高关联规则的挖掘效率。As one of the research hotspots in recent years,the association algorithm Apriori algorithm has a lot of candidate items because it needs to scan the database several times,which seriously affects the mining efficiency of association rules.On this basis,a weighted association rule mining algorithm based on matrix compression was proposed.The database only needs to be scanned once and converted into a 0-1 m atrix,and the matrix is compressed according to the related properties,which reduces the amount of calculation of the algorithm execution.At the same time,taking into account the importance of the project,a weighted approach is adopted with the probability of the way to set the weight of the project attributes.Compared with Apriori algorithm,this algorithm can find high order frequent itemsets directly in the mining process.Experimental results show that this algorithm can effectively improve the mining efficiency of association rules.

关 键 词：关联规则挖掘 APRIORI算法 矩阵压缩 加权 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]