Fast mining maximal frequent itemsets based on Sorted FP-tree

In order to improve the efficiency of mining maximal frequent itemsets, a new algorithm SFP-Miner based on Sorted FP-Tree(SFP-Tree) was proposed. The algorithm scanned the database twice and generated the SFP-Tree first, then mined maximal frequent itemsets by depth-first method and directly in SFP-Tree by merging sub-trees, finally obtained the maximal frequent itemsets. The experimental results indicate that SFP-Miner is an efficient algorithm. It has superior to the algorithm Mafia in performance.