Hadoop’s architecture is shown in figure 1. The master node includes JobTracker, NameNode,

Abstract. This paper implemented effective associate rule mining based on Hadoop parallel computing. First, the parallel FP-growth algorithm was run on Hadoop platform to find the frequent item sets of the transaction data. Second, the strong association rules was generated from the frequent item sets by a designed algorithm. Then, redundant rules were deleted according to filtering conditions to make model evaluation. After those steps, all the funny and non-redundancy strong association rules were mined out. In addition, this paper also analyzed the efficiency of the Hadoop parallel computing and explained the superiority of the Hadoop parallel computing when it handles big data. 1．Hadoop Platform Hadoop is an open source distributed computing platform[1] using a lot of inexpensive computers to build a Hadoop cluster. Through deploying applications to every computer of the Hadoop cluster to implement the parallel processing, speed up the processing speed and extend the storage space. Its superiority is to handle of large data in parallel. In addition, Hadoop has a lot of merits, such as high reliability, high scalability, high efficiency, and high fault tolerance and so on. Fig 1. Hadoop’s architectur