Efficient Mining of Fuzzy Frequent Itemsets with Type-2 Membership Functions

摘要

In the past, the Apriori-based algorithm with fuzzy type-2 membership functions was designed for discovering fuzzy association rules, which is very time-consuming to generate-and-test candidates in a level-wise way. In this paper, we present a list-based fuzzy mining algorithm to mine the fuzzy frequent itemsets with fuzzy type-2 membership functions. A fuzzy-list structure and an efficient pruning strategy are respectively designed to speed up the mining process of fuzzy frequent itemsets. Several experiments are carried to verify the efficiency and effectiveness of the designed algorithm compared to the state-of-the-art Apriori-based algorithm in terms of runtime and number of traversal nodes (candidates).