改进时间卷积网络下局域网异常状态预测方法  

LAN Anomaly Prediction Method Based on Improved Time Convolution Network

作　　者：葛昕[1] 岳敏楠 GE Xin;YUE Min-nan(Shanghai for Science and Technology,Shanghai 200093,China;University of Shanghai for Science and Technology,School of Energy and Power Engineering,Shanghai 200093,China)

机构地区：[1]上海理工大学,上海200093 [2]上海理工大学能源与动力工程学院,上海200093

出　　处：《计算机仿真》2024年第1期438-442,共5页Computer Simulation

摘　　要：局域网异常会阻碍网络运行速度,严重时会导致网络瘫痪。为了精准预测局域网是否存在异常,提出一种基于改进时间卷积网络的局域网异常预测方法。组建变分模态分解(Variational Mode Decomposition,VMD)高频噪声分量判定标准,剔除高频分量,将剩余VMD分量叠加重构,去除局域网数据中的噪声。建立局域网异常预测模型,将去噪后的局域网数据特征数值规约到和灰度图像像素值对应的范围内,形成局域网灰度图,并将其输入到改进时间卷积网络结构中训练和模型调优,完成局域网异常预测。经实验测试证明,所提方法可以获取高精度和高效率的局域网异常预测结果,在局域网异常预测领域具有广阔的发展前景。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 corresponding 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.Experimental 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.

关 键 词：改进时间卷积网络 局域网 改进灰狼优化算法 异常预测 变分模态分解 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]