The design of network fault diagnosis system based on PNN

In the process of computer network fault diagnosis, the fault´s characteristic data has a complicated nonlinear relationship with the faults´ reason. It´s usually adopting BP neural network to diagnose the network fault. This article firstly studies the PNN neural network, and then researches the characteristic parameters of the computer network. At last, the method of network fault diagnosis based on PNN is put forward. Experimental result indicates that PNN can well diagnose the faults of computer network, and has a better performance than many other neural networks.