A Graph-Based Algorithm for Mining Maximal Frequent Itemsets

Association rule mining is an important research branch of data mining, and computing frequent itemsets is the main problem. The paper is designed to find maximal frequent itemsets only. It presents an algorithm based on a frequent pattern graph, which can find maximal frequent itemsets quickly. A breadth-first-search and a depth-first-search techniques are used to produce all maximal frequent itemsets of a database. The paper also analyzes the complexity of the algorithm, and explains the computation procedure by examples. It has high time efficiency and less space complexity for computing maximal frequent itemsets.