Distributed fuzzy clustering based association rule mining: design, deployment and implementation

With the explosive growth of transactional big data in various scenarios such as manufacturing chain and supply chain etc., fuzzy association rule mining (FARM) algorithms as a practical method in data analysis and decision making are confronted with unprecedented challenges. Commonly used centralized FARM algorithms are no longer suitable for current numerical big data. In this paper, a distributed fuzzy clustering based association rule mining (DFARM) framework is proposed where outside-layer and inside-layer distribution are employed to realize the parallel operation of the whole FARM algorithm. In order to implement the proposed framework, we specifically design an implementation algorithm by the ‘MapReduce’ paradigm. Through the proposed algorithm, distributed association rule mining can be carried out for any form of data such as Boolean data or numerical data with various volumes. Furthermore, various experiments are conducted to validate that our algorithm outperforms the centralized one in terms of device utilization and time performance.

Recommended citation: Wu, J., Dai, L., Ma, Y., Zou, W., & Xia, Y. (2021, October). Distributed fuzzy clustering based association rule mining: Design, deployment and implementation. In 2021 China Automation Congress (CAC) (pp. 4366-4372). IEEE.
Download Paper