基于MapReduce的Apriori前后项约束关联规则改进算法  

Improved Algorithm for Association Rules of Apriori Before and After Items Based on MapReduce

作　　者：王伟 储泽楠[1,2,3] 韩毅[1,4] 吴朝霞[1,2] 焦清局[5] WANG Wei;CHU Zenan;HAN Yi;WU Zhaoxia;JIAO Qingju(School of Computer Science and Information Engineering,Anyang Institute of Technology,Anyang 455000,China;Anyang Information System Application Engineering Technology Research Center,Anyang Institute of Technology,Anyang 455000,China;Henan Province High Precision Spindle Engineering Laboratory,Anyang Institute of Technology,Anyang 455000,China;National CNC System Engineering Technology Research Center,Huazhong University of Science and Technology,Wuhan 430000,China;Ministry of Education Oracle Information Processing Key Laboratory,Anyang Normal University,Anyang 455000,China)

机构地区：[1]安阳工学院计算机科学与信息工程学院,河南安阳455000 [2]安阳工学院安阳市信息系统应用工程技术研究中心,河南安阳455000 [3]安阳工学院河南省高精密主轴工程实验室,河南安阳455000 [4]华中科技大学国家数控系统工程技术研究中心,湖北武汉430000 [5]安阳师范学院教育部甲骨文信息处理重点实验室,河南安阳455000

出　　处：《信阳师范学院学报（自然科学版）》2020年第3期448-453,共6页Journal of Xinyang Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(61806007);河南省科技计划项目(182102210197);河南省高等学校重点科研项目(2020ZDJH002)。

摘　　要：针对经典的Apriori算法依赖内存,只适用于小规模数据集,在面对海量数据集时显得无能为力以及该算法没有考虑用户的需求情况等问题,提出了基于MapReduce的Apriori前后项约束关联规则改进算法.该方法首先对经典Apriori算法挖掘过程进行了改进,加入了用户的前后项约束规则,使得在挖掘过程中剪枝的程度更大并且获取到更加精准的规则.然后利用云计算的MapReduce编程技术,对改进的Apriori算法的各个步骤并行化.实验结果表明,改进的算法在处理不同的数据集时有一定的优势,然后经过MapReduce模型并行化后,提高了对海量数据的处理能力和效率,并且具有良好的扩展性.Aiming at the memory dependence of the classic Apriori algorithm,it is only suitable for small-scale datasets,it seems to be powerless in the face of massive datasets,and the algorithm does not consider the user’s needs.The improved algorithm of Apriori pre-term constraint association rules based on MapReduce is proposed.Firstly,the method of the classic Apriori algorithm mining process is improved,and the user’s pre-and post-item constraint rules are added,which makes the pruning degree more in the mining process and obtains more precise rules.Then,using the MapReduce programming technology of cloud computing,the steps of the improved Apriori algorithm are parallelized.The experimental results show that the improved algorithm has certain advantages in dealing with different data sets.After parallelization by MapReduce model,it improves the processing ability and efficiency of massive data and has good scalability.

关 键 词：关联规则 APRIORI算法 项约束 MAPREDUCE 并行算法 HADOOP 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]