Mining Maximal Frequent Itemsets with Frequent Pattern List

Mining frequent itemsets is a major aspect of association rule research. However, the mining of the complete of frequent itemsets will lead to a huge number of itemsets. Fortunately, this problem can be reduced to the mining of maximal frequent itemsets. In this paper, we adopt frequent pattern list (FPL) and bit string technique, propose a novel algorithm for mining maximal frequent itemsets based on frequent pattern list (FPLMFI-Mining). It conducts various operations on FPL according to the frequency of frequent items. Moreover, it utilizes bit string and-operation to test maximal frequent itemsets. This algorithm can be scaled up to very large databases by parallel projection and compress technique.