Non-iteration Parallel Algorithm for Frequent Pattern Discovery

摘要

For the high time overhead problems of Apriori algorithm while solving for the long length frequent patterns, using the MapReduce distributed programming ideas, the paper breaks the original idea of Aproiri which discovers the frequent item sets through gradually increasing the element numbers in the frequent item sets. It proposes a new non-iteration parallel algorithm about frequent pattern discovery, which can get arbitrary length frequent pattern at random. The experimental results show that the proposed algorithm has better time performance than such parallel algorithms which are under the ideas of traditional Apriori algorithm.