The Benefits of Using Prefix Tree Data Structure in Multi-Level Frequent Pattern Mining

Finding frequent itemsets is one of the most investigated fields of data mining. In this paper, the horizon of frequent pattern mining is expanded by extending single-level algorithms for mining multi-level frequent patterns. There are presented two algorithms that extract multi-level frequent patterns from databases using two efficient data structures: FP-tree and AFOP-tree, to represent the conditional databases. A comparison study is made between using these data structures and algorithms and Apriori algorithm to reflect their benefits. The compared algorithms are presented together with some experimental data that leads to the final conclusions.