1种基于密歇根编码的模糊分类系统设计方法

A Method of Designing Fuzzy Classification System Based on Michigan Encoding

兼顾模糊系统精确性和解释性,提出1种基于遗传算法的模糊分类系统设计方法.该算法在考虑模糊分类系统解释性的前提下,基于数据样本构建完整的规则集,并采用密歇根编码方式优化规则集和隶属函数参数,在保证系统解释性的同时提高了系统的精确性,仿真实验结果验证了该方法的有效性.

Considering the tradeoff between the accuracy and the interpretation of the fuzzy system, a design method of fuzzy classification system based on genetic algorithm is proposed. The new algorithm in consideration of fuzzy classification system interpretability, a complete set of rules is constructed based on the sample data, and the Michigan encoding of rule sets and membership function parameters were optimized in ensuring the interpretative system and improving the accuracy of the system. Simulation results verify the effectiveness of the proposed method.