Mining Frequent Itemsets Using a Pruned Concept Lattice

Mining frequent itemsets is a crucial step in association rule mining. However, most algorithms mining frequent itemsets scan databases many times, which decreases the efficiency. In this paper, the relationship between a concept lattice and frequent itemsets is discussed, and the model of pruned concept lattice (PCL) is introduced to represent frequent itemsets in a given database, and the scale of frequent itemsets is compressed effectively. An algorithm for mining frequent itemsets based on PCL is proposed, which prunes infrequent concepts timely and dynamically during the PCL´s construction according to the Apriori property. The efficiency of the algorithm is demonstrated with experiments.