基于粗糙集和信息熵的BP神经网络故障诊断  

Fault diagnosis of BP neural network based on rough set and information entropy

作　　者：王发兴[1] 沈永红[2] 

机构地区：[1]南京邮电大学通达学院,南京210046 [2]天水师范学院数学与统计学院,甘肃天水741000

出　　处：《计算机辅助工程》2009年第3期78-82,共5页Computer Aided Engineering

摘　　要：针对粗糙集只能处理量化数据,容错和推广能力较差的缺点以及BP神经网络的维数灾难问题,提出1种基于信息熵的粗糙集属性离散化方法.该方法利用粗糙集对属性进行约简,解决BP神经网络的维数灾难问题,并将BP神经网络用于模式分类补偿粗糙集属性约简用于模式分类时的不足.实例分析表明该方法具有较好的故障诊断效果.Rough set can only process quantization data, and the ability of fault-tolerant and generalization is weak, meanwhile, BP neural network has the dimension disaster problem. So a rough set attribute discretization method based on information entropy is proposed. The attribute is reduced to solve the dimension disaster problem of BP neural network. BP neural network is used to deal with the pattern classification to make up for the shortcoming brought by attribute reduction. The example result shows that the method has better performance for fault diagnosis.

关 键 词：粗糙集 信息熵 离散化 神经网络 故障诊断 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]