摘要： 针对粗集神经网络构建过程中的论域空间划分问题，提出一种基于模糊聚类的论域划分方法。将带交叉变异算子的粒子群优化算法(PSO)与模糊C-均值聚类算法(FCM)相结合，给出一种新的模糊聚类算法CMPSO-FCM，该算法具有良好的搜索能力和聚类效果。提出一种基于信息熵的模糊粗糙集决策规则获取方法，并用获取的规则指导粗集神经网络的构建。实验结果表明，该方法构造的神经网络具有更精简的结构、较好的分类精度和泛化能力。

关键词: 粗集神经网络, 模糊聚类, PSO算法, FCM算法, 信息熵, 属性约简

Abstract: Aiming at the problem of the universal space partition in the process of constructing rough set neural network, this paper proposes an universe of discourse method based on fuzzy clustering. A modified PSO algorithm with crossover and mutation operators is combined with FCM algorithm. And a new fuzzy clustering algorithm(CMPSO-FCM) is proposed. The searching capability and clustering effectiveness are improved by the new algorithm. A set of fuzzy rough decision rules are acquired by entropy method, and a rough set neural network is designed under these decision rules. Experimental results show that this method has superiorities at the aspect of structure, classification precision and generalization.

Key words: rough set neural network, fuzzy clustering, PSO algorithm, FCM algorithm, entropy, attribute reduction

中图分类号: 

• TP391