基于神经网络的网络故障诊断

Network Fault Diagnosis Based on Neural Network

针对传统故障知识库冗余度高和稳定性难以两全的缺陷，提出了RSFBP算法，综 合运用神经网络方法和粗糙集理论，实现对学习样本净化处理。实验表明，利用该方法实现 的系统在进行网络故障诊断时可以取得较好的效果。

This paper puts forward a design method for network fault diagnosis systems by proposing RSFBP algorithm, which tightly combines neural network and rough sets. It's capable of overcoming several shortcomings in existing diagnosi s systems, such as a dilemma between stability and redundancy. The experiment sy stem implemented by this method shows a good diagnostic ability.