Optimization of fuzzy classification system by genetic strategies

A novel approach to construct fuzzy classification system based on fuzzy association rules is proposed in this paper. Competitive agglomeration algorithm is employed to partition quantitative attributes from each data record into several optimized fuzzy sets, resulting in an initial fuzzy classification system. A fuzzy classification system with high accuracy and interpretability can be further achieved by genetic strategies. Simulation applied to an existent diabetes dataset demonstrates the performance of the proposed approach is better than those of other popular classification methods.