Data is increasing rapidly day by day along with the transactional database. Dividing this data and storing it in a distributed manner is an effective way for storage and retrieval. Mining such distributed data with minimum dependence between sub-problems is a crucial task. Finding frequent itemsets and corresponding association rules is a big challenge while considering the aggregation in a distributed environment. To overcome these challenges, we propose a distributed frequent itemset generation and association rule mining algorithm using MapReduce programming model. The proposed scheme generates frequent itemset and mine association rules using a synthesized distributed technique. The rules are mined in a distributed manner, and then weights are assigned to subsets of data and association rules. A proper mixture of association rules that are generated in distributed manner is done using a weighted approach. This paper presents a novel MapReduce-based synthesis approach, which can work well over a distributed storage of large amount of data.

数据与事务数据库一起迅速增长。分割这些数据并以分布式方式存储它是一种有效的存储和检索方式。挖掘具有最小子问题依赖性的此类分布式数据是一项至关重要的任务。在考虑分布式环境中的聚合时，查找频繁的项目集和相应的关联规则是一个很大的挑战。为了克服这些挑战，我们提出了一种使用MapReduce编程模型的分布式频繁项集生成和关联规则挖掘算法。提出的方案使用综合的分布式技术生成频繁的项目集和矿山关联规则。以分布式方式挖掘规则，然后将权重分配给数据和关联规则的子集。使用加权方法可以适当地混合以分布式方式生成的关联规则。本文提出了一种新颖的基于MapReduce的综合方法，该方法可以在大量数据的分布式存储上很好地工作。