Discovery of Frequent Itemsets: Frequent Item Tree-Based Approach

Abstract

Mining frequent patterns in large transactional databases is a highly researched area in the field of data mining. Existing frequent pattern discovering algorithms suffer from many problems regarding the high memory dependency when mining large amount of data, computational and I/O cost. Additionally, the recursive mining process to mine these structures is also too voracious in memory resources. In this paper, we describe a more efficient algorithm for mining complete frequent itemsets from transactional databases. The suggested algorithm is partially based on FP-tree hypothesis and extracts the frequent itemsets directly from the tree. Its memory requirement, which is independent from the number of processed transactions, is another benefit of the new method. We present performance comparisons for our algorithm against the Apriori algorithm and FP-growth.