A mining algorithm for distributed global maximal frequent itemsets based on Sorted SCan-Tree

摘要

A new algorithm, named SCan-MAX for mining distributed maximal frequent itemsets from databases was proposed, the SCan-MAX used Sorted SCan-tree to store all the information of the transactions from the databases. SCan-MAX firstly scanned the local database and then gets the global 1-frequent itemsets, and then it created the SCan-tree on each node and used orderly sequence to store frequent itemsets. All the local SCan-tree was send to the master node to create the global SCan-tree. Then, the root's sub-tree was send to the slave node to mine local maximal frequent itemsets and send the information of itemsets to the master node to mine the global maximal frequent itemsets. Experimental results show that the algorithm makes full use of the characteristics of SCan-Tree, significantly reduces the amount of communication required between nodes, thus enhanced the efficiency of the algorithm. Theoretical analysis and experimental results show that the algorithm has a good time and space efficiency.