LAN Anomaly Prediction Method Based on Improved Time Convolution Network
Currently,LAN anomalies can hinder network performance and even cause network paralysis.In order to accurately predict whether there are LAN anomalies,this paper proposed a method of predicting LAN anomaly based on improved time convolutional network.Firstly,we established a high-frequency noise component judgment standard based on Variational Mode Decomposition(VMD)to eliminate high-frequency components.Meanwhile,we superimposed and reconstructed the remaining VMD components,thus removing noise from LAN data.Secondly,we built a LAN anomaly prediction model,reducing the numerical values of denoised LAN data features to the range cor-responding to the grayscale image pixel,thus forming a LAN grayscale map.Then,we input it into the improved time convolutional network structure for training and model tuning,and finally completed the LAN anomaly prediction.Ex-perimental test results show that the proposed method can obtain high-precision and high-efficiency LAN anomaly prediction results,and it has broad development prospects in the field of LAN anomaly prediction.
Improved time convolution networkLANImproved grey wolf optimization algorithmAbnormal pre-dictionVMD
NETL
NSTL
万方数据
维普
