Fast Distributed Mining Algorithm of Maximum Frequent Itemsets Based on Cloud Computing

The paper proposed a fast distributed mining algorithm of maximum frequent itemsets based on cloud computing, namely, FDMMFI algorithm. FDMMFI algorithm made nodes compute local maximum frequent itemsets by cloud computing, then the center node exchanged data with other nodes and combined, finally, global maximum frequent itemsets were gained by cloud computing. Theoretical analysis and experimental results suggest that under the same minimum support threshold, communication traffic and runtime of FDMMFI decreases while comparing with CD and FDM. The less the minimum support threshold, the better the three performance parameters of FDMMFI. FDMMFI algorithm is fast and effective.