A two-phase fuzzy mining approach

Abstract

In this paper, we propose a two-phase fuzzy mining approach based on a tree structure to discover fuzzy frequent itemsets from a quantitative database. A simple tree structure called the upper-bound fuzzy frequent-pattern tree (abbreviated as UBFFP tree) is designed to help achieve the purpose. The two-phase fuzzy mining approach can easily derive the upper-bound fuzzy supports of itemsets through the tree and prune unpromising itemsets in the first phase, and then finds the actual frequent fuzzy itemsets in the second phase. Experimental results also show the good performance of the proposed approach.