基于协同进化算法的高维模糊分类系统的设计

Design of High-dimensional Fuzzy Classification Systems Based on Cooperative Coevolutionary Algorithm

基于协同进化算法,提出一种高维模糊分类系统的设计方法.首先定义系统的精确性指标,给出解释性的必要条件,利用聚类算法辨识初始模型.相互协作的3类种群分别代表系统的特征变量、规则前件和模型隶属函数的参数,适应度函数采用3类种群合作计算的策略,在算法运行中利用基于相似性的模型简化技术约简模糊系统,最后利用该方法对W ine问题进行研究.仿真结果表明该方法能够对高维分类问题的特征变量进行选择,同时利用较少规则和模糊集合数达到较高的识别率.

An approach to construct high-dimensional fuzzy classification systems is proposed based on multi- objective cooperative coevolutionary algorithm. The precision index is defined, and the necessary conditions of interpretability are analyzed. The initial fuzzy system is identified using fuzzy clustering algorithm. Three cooperative species represent the relevant features, the structure and parameters of the fuzzy systems respectively. The fitness function is calculated on the cooperation of individuals from the three species. The similarity-driven rule based simplification method is applied to reduce the fuzzy model. By computational simulation on the wine classification problem, the proposed approach is able to generate compact fuzzy rule bases with high classification ability.