基于粗糙集和BP神经网络算法的网络故障诊断模型研究

针对计算机网络故障诊断知识库冗余性高、神经网络与PCA、DS证据等理论相结合诊断精度不高的难题, 提出了一种新的基于粗糙集和BP神经网络的计算机网络故障诊断模型. 首先利用粗糙集算法对网络故障特征进行约简处理、提取最小诊断规则; 其次利用最小规则训练BP神经网络, 建立基于粗糙集和BP神经网络的计算机网络故障诊断模型; 最后将模型运用于真实网络故障数据诊断. 结果表明: 该模型具有学习效率高、诊断速度快、准确率高的特点, 能够快速诊断网络故障类型.

Network Fault Diagnosis Based on Rough Sets and BP Neural Network

To deal with the problems of redundancy of network fault diagnosis, the knowledge base and Low Accuracy of neural network model combined with PCAand DS evidence theory are presented in this paper. A new fault diagnosis model of computer network based on rough set and BP neural network is engineered, in which many fault features of computer network are retrieved. These features are then reduced to the minimum diagnosis rules using rough set. The minimum diagnosis rules are trained by BP neural network. The simulation results indicate that the new fault diagnosis model has higher learning efficiency, faster speed of diagnosis and higher detection accuracy.