Study of Algorithms for Mining Fuzzy Association Rules and Applications

Abstract

Knowledge discovery in the form of rules has become a useful and meaningful practice in Data Mining. In the last three decades, Association Rule Mining has been considered one of the primary research activities. Introducing fuzzy mathematical theories provides another dimension to make association rules more user-friendly and expressive. Membership degrees and Linguistic Terms play a significant role in the fuzzification process. The use of Fuzzy set concepts in rule generation, however, needs more attention from the researchers in different application domains. This paper is based on the study and findings of some important Fuzzy Association Rule Mining (FARM) algorithms and their applications. Beginning with the overview of Fuzzy Set theory and FARM algorithms, this study gives an analysis and account of the intricacies of algorithm development and applications-based activities in the last three decades. Out of many algorithms, some important FARM algorithms are undertaken for the study and explained with examples. The examples are prepared from both real and synthetic data. Based on the study, some key features of leading algorithms are observed and highlighted. Different research issues and challenges related to FARM are found out and discussed for further research.