摘要
通过将模糊集和粗集,神经网络结合,提出了一种基于模糊规则的新的粗模糊神经网络,它通过利用误差反向传播算法实时修正该新型网络中的权值参数,从而能被有效地应用于不确定系统的决策分类与模式识别问题.最后通过对一个不确定决策系统的模式识别的仿真结果表明该粗模糊神经网络能大大提高模式识别决策的准确率.
A new rough_fuzzy neural network based on extracting of fuzzy rules is proposed by integrating fuzzy set and rough set with neural network.It can be effectively used in decision controlling for any uncertain decision system by timely correcting the parameter value of network with error back propagation algorithm.The simulation results of a model recognition indicates that the construction approach of the rough_fuzzy neural network can improve model recognition probabilities.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2005年第2期330-334,共5页
Control Theory & Applications
基金
国家自然科学基金资助项目(10041005)
广东省自然科学基金资助项目(011221).
关键词
粗集
模糊神经网络
决策表
模式识别
rough set
rough_fuzzy neural network
decision table
pattern recognition
