Mining fuzzy frequent itemsets based on UBFFP trees

Abstract

In the past, most algorithms proposed for mining association rules handle items with binary values. Transactions with quantitative values are, however, commonly seen in real-world applications. Fuzzy mining algorithms for inducing rules from quantitative databases have thus been developed, most of which are based on the Apriori algorithm. Fuzzy mining is seldom based on frequent pattern (FP) trees because fuzzy-set processing from an FP tree is much more complicated than crisp-set processing. In this paper, a two-phase fuzzy mining approach based on a tree structure to obtain fuzzy frequent itemsets from a quantitative database is proposed. A simple tree structure called the upper-bound fuzzy frequent-pattern (UBFFP) tree is designed. The two- phase fuzzy mining approach can easily derive the upper-bound fuzzy counts of itemsets using the tree. It prunes unpromising itemsets in the first phase, and then finds the actual fuzzy frequent itemsets in the second phase. Experimental results show that the proposed approach has good performance.