Application of an Improved BP Neural Network in Local Area Network Fault Diagnosis

The methods of existing network fault diagnosis have many deficiencies.In order to exclude the fault of network accurately and effectively,in this paper,artificial neural network method is applied in fault diagnosis on the LAN.An improved BP neural network is proposed for the purpose of overcoming the slow convergence and existence of local minimum in conventional BP neural network.Both the conventional and the improved BP neural network are applied in the LAN fault diagnosis respectively.The result of the simulation shows that the improved BP neural network is more fast and effective than the conventional BP neural network in the task of fault diagnosis,thus,this new method has practicability.