A New Parallel Algorithm for Mining Maximum Frequent Itemsets Based on Maximum Complete Subgraph
Zhixin Ma · Microelectronics & Computer · 2007
This paper presents a parallel algorithm for mining maximum frequent itemsets based on maximum complete subgraph. The algorithm discovers maximum complete subgraph by traversing the frequent 2-itemsets map presented by adjacent matrices, and divides itemsets with maximum complete subgraph vertex sequence. The innovation is dividing maximum frequent itemsets into LMFI, PMFI and SMFI. This method reduces the cost of merging maximum frequent itemsets. Meanwhile the algorithm has been implemented and optimized.