Frequent pattern mining is one of the most important data mining tasks. Traditional frequent pattern mining algorithmsare executed in a "batch" mode, that is,all the data are mined in one time, so they cannotmeet the needs of the ever-growing bigdata mining. MapReduce is a popular parallel computing modeland has been widely used in the field of parallel data mining. In this paper, we migrate the traditional frequent pattern incremental mining algorithm CanTree to the MapReduce computing model,achieving a parallel frequent pattern incremental miningalgorithm. The experimental results show that the proposed algorithm achievesbetterload balancing and improvesthe execution efficiency significantly.

 

data mining, frequent pattern mining, incremental mining, MapReduce, Hadoop, PFP