A depth-first search algorithm of mining maximal frequent itemsets

Mining maximal frequent itemsets is a fundamental and important issue in many data mining application. A new depth-first search algorithm for mining maximal frequent itemsets called DFMFI (depth-first search for maximal frequent itemsets) is proposed, which can reduce the number of candidate itemsets and the cost of support counting. DFMFI projects the dataset information stored by the compressed FP-tree into the conditional matrix, and improves efficiency of support counting by using vector logic operation. Global 2-itemset pruning and local extension pruning used to prune the search space effectively. The experiments results verify the efficiency and advantage of this DFMFI.