Affiliations: [a] College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, China
| [b] Department of Computer Science, Electrical Engineering and Mathematical Sciences, Western Norway University of Applied Sciences, Bergen, Norway
| [c] Department of Computer Engineering, Sejong University, Seoul, Korea
| [d] Department of Computer Science and Information Engineering, Tamkang University, New Taipei City, Taiwan
| [e] Department of Information and Finance Management, National Taipei University of Technology, Taipei, Taiwan
| [f] Department of Mathematics & Computer Science, Brandon University, Brandon, Canada
| [g] School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen), Shenzhen, China
Correspondence:
[*]
Corresponding author. Jerry Chun-Wei Lin, Department of Computer Science, Electrical Engineering and Mathematical Sciences, Western Norway University of Applied Sciences, Bergen, Norway. E-mail: jerrylin@ieee.org.
Abstract: Association-rule mining (ARM) has concerned as an important and critical research issue in the field of data analytics and mining that aims at finding the correlations among the items in binary databases. However, the conventional algorithms considered the frequency of the item(set) in binary databases for ARM, which is not sufficient in real-life situations. Mining of useful information is not an easy task especially if the item(set) consists of the added values. Moreover, the discovered knowledge is not easy to understand if you are not the domain experts. For the past decades, several intelligent systems involved the fuzzy-set theory for many domains and applications due to it is interpretable for human reasoning. Before, the Apriori-based method for discovering fuzzy frequent itemsets (FFIs) based on the type-2 fuzzy-set theory was proposed, which requires the amount of computations with enormous candidates. In this study, we then first present a fast list-based multiple fuzzy frequent itemset mining (named as LFFT2)algorithm under type-2 fuzzy-set theory. It is developed by the type-2 membership functions to retrieve the multiple fuzzy frequent itemsets for presenting more useful and meaningful knowledge for making the efficient strategies or decisions. From the results shown in the experiments, it is clear to see that the developed LFFT2 outperforms the conventional Apriori-based approach regarding the execution time and the number of examined nodes in the search space.
