An efficient algorithm for mining complete share-frequent itemsets using BitTable and heuristics

This paper proposes a new efficient algorithm for mining share-frequent itemsets from BitTable knowledge - extracted once from a transaction database. The knowledge contains sufficient information for such a mining task and provides efficient interactive access. The algorithm finds all share-frequent itemsets by level-wise generating complete promising candidates from a BitTable using heuristics and testing for desired solutions. Simulation results reveal that the proposed algorithm perform significantly better than ShFSM and DCG both runtime and a number of generated candidates.