Multi-objective optimisation based fuzzy association rule mining method

Fuzzy association rule mining (FARM) is a mainstream method to discover hidden patterns and association rules in quantitative data. It is essential to improve performance metrics, including quantity performance (e.g., the number of rules, the number of frequent itemsets) and quality performance (e.g., fuzzy support and confidence). The current approaches inadequately support optimisation of both quantity and quality performance. We propose a multi-objective optimisation algorithm for FARM (MOOFARM), where quantity and quality performance metrics are improved and validated simultaneously. The experimental evaluation conducted on a real dataset showcases the outstanding performance of MOOFARM against state-of-the-art works. In particular, at minimum support = 0.1, minimum confidence = 0.7, our MOOFARM increases the quantity performance up to 11 times. The proposed method improves the quality performance up to 71.05%.