A Multi-Objective Nutcracker Optimization Algorithm Based on Cubic Chaotic Map for Numerical Association Rule Mining

Traditional numerical association rule mining optimization algorithms have limitations in handling discrete attributes, and they are susceptible to becoming trapped in local optima, uneven population distribution, and poor convergence. To address these challenges, we propose a multi-objective nutcracker optimization algorithm based on a cubic chaotic map (C-MONOA), specifically designed for mining association rules from mixed data (continuous and discrete). Unlike existing models, C-MONOA leverages a chaotic map for population initialization, alongside Michigan rule encoding, to dynamically optimize feature intervals during the optimization process. This algorithm integrates continuous and discrete data more effectively and efficiently. This article uses support, confidence, Kulc metric, and comprehensibility as evaluation indicators for multi-objective optimization. The experimental results show that C-MONOA performs well in rule scoring and can generate frequent, simple, and accurate rule sets. This study extends the association rule mining method for mixed data, demonstrating high performance and robustness and providing new technical tools for application fields such as market analysis and disease prediction.