Research for parallel apriori algorithm based on MPI

In order to improve the efficiency of Apriori mining algorithm for the Ultra-large-scale data sets, based on the partition for the candidate itemsets, this paper presents a parallel algorithm for mining association rules which directly using MPI for passing message base on the master-slave structural model. Simulation analysis showed that the mining time of the algorithm which proposed in this paper has a higher degree of shortening compared with the algorithm of Apriori. There have good parallelism and scalability especially for large-scale database mining.