Original Article

Computer Communication Network Fault Detection Based on Improved Neural Network Algorithm


School of Computer Science and Technology, Henan Polytechnic University, Xinxiang


GNA University, Village Hargobindgarh, Phagwara, Punjab, India


Electronics & Communication Engineering Department, National Institute of Technology, Hamirpur, India


Department of Computing, Mathematics and Physics Høgskulen på Vestlandet Bergen Norway

ELECTRICA 2022; 22: 351-357
DOI: 10.54614/electrica.2022.21168
Read: 404 Downloads: 147 Published: 04 July 2022

To detect computer communication network failures, a computer communication network fault detection based on an improved neural network algorithm is proposed. A network fault diagnosis example is used to verify the effectiveness of the method. There are many network failure phenomena. Here, the author selected 13 network fault information parameters for a comprehensive diagnosis of network failures. The author designed a three-layer backpropagation (BP) neural network. There are 13 nodes in the input layer, corresponding to the above 13 network fault parameter information. The output layer has three nodes, which output the fault code sequence. The training of the network uses the trainlm() function. The performance function uses the mean square error performance function mse() and set e = 0.001; the network learning rate is set to a = 0.05. The author selects 100 failure data as the training set for network training and selects 10 sets of samples as the test set. The experimental data shows that after the network has been trained 25 times, the output error reaches the set precision e. After training the BP network using this algorithm 140 times, the output error reaches the set precision e. This method effectively improves the effectiveness and accuracy of S network fault diagnosis.

Cite this article as: D. Sun, P. Chopra, J. Bhola and R. Neware, "Computer communication network fault detection based on improved neural network algorithm," Electrica., 22(3), 351-357, 2022.

EISSN 2619-9831