基于递阶遗传算法模糊加权神经网络的模糊规则自动获取

针对模糊规则的自动获取一直是模糊系统的一个瓶颈问题，提出一种基于递阶结构的混合编码遗传算法与进化规划相结合的模糊加权神经网络学习新算法，利用该算法同时优化模糊加权神经网络的结构和参数，最后说明了从网络中提取模糊规则的方法，从而自动获得最优的模糊规则。分析和实验结果表明，本文方法在规则提取和分类准确性等方面比其他方法更好。

Automatic fuzzy rule extraction based on hierarchical genetic algorithm weighted fuzzy neural networks

A Weighted fuzzy neural network learning algorithm which combines a genetic algorithm based on hierarchical structure and evolutionary programming was presented. The shooting algorithm is used to optimize the Weighted fuzzy neural network structure and train the connection weights, and presented fuzzy rule extraction method from weighted fuzzy neural networks. So fuzzy rules from the trained Weighted fuzzy neural network can be automatically extracted. The analysis and experiment result show that the method in this paper performs better than other method on the rule extraction and the best classifying veracity.