Distributed Apriori on cloud plateform without exposing individual transaction data

Privacy preserving and security issue becomes more and more important in data mining problems, especially in cloud computing environment. On cloud computing platform, data always distributed over different servers, It is not secure for each server to share data with other servers. Association rule mining is a core problem in machine learning domain, the widely used Apriori algorithm as a useful tool for personalized recommendation must address the privacy security problem in cloud computing environment. This paper addresses association rule mining issue with distributed data. Each server node is with some part of each transaction. Servers should corporate to get global association rules. In this process, no individual information should be revealed. One algorithm for finding frequent itemsets without revealing privacy data is proposed in this paper. In addition, one association rule mining algorithm in Hadoop which preserves privacy information has been proposed. The experiment presents the effectiveness of our method.