MMFI: An Effective Algorithm for Mining Maximal Frequent Itemsets

Existing algorithms for mining maximal frequent itemsets have to do superset checking, and some of them using FP-tree have to construct conditional frequent pattern trees recursively. We present a novel algorithm for mining maximal frequent itemsets from a transactional database. In the algorithm, the FP-Tree data structure is used and adapted, and a new strategy called ldquoNBNrdquo (Node By Node) is used for traversing the adapted FP-Tree. Neither superset checking nor constructing conditional frequent pattern trees is needed in the algorithm. We analyze the performance of the algorithm and compare our method with existing algorithms. Our technique works better for mining maximal frequent itemsets. It is also proved by experimental comparison that our algorithm is more fast and efficient.