An Efficient Matrix Algorithm for Mining Frequent Itemsets

Abstract

Efficient algorithms for mining frequent itemsets are crucial for mining association rules as well as for many other data mining tasks. In this paper, we integrate the merits of the matrix algorithm and Index-BitTableFI algorithm, and design an efficient algorithm for mining the frequent itemsets. In the new algorithm, it may be generated directly some frequent itemsets which do not generate in the Index-BitTableFI. At the same time, we do not use recursive method which is time-consuming to compute the other frequent itemsets in Index-BitTableFI algorithm, and use breadth-first search strategy to generate all frequent itemsets. On the other hand, we use the method of the matrix algorithm to compute the supports of the frequent itemsets which do not generate with subsume index technology. Since there are many frequent itemsets which can be generated directly in the new algorithm, the efficiency of the new algorithm is improved. Then an example is used to illustrate the new algorithm. The results of the experiment show that the new algorithm in performance is more remarkable for mining the long and small supports frequent itemsets for sparse datasets and mining frequent itemsets in dense datasets.