CAJ | 학술논문

网络流量是一种复杂、多变的非线性混沌系统,为了获得更加理想的网络流量预测结果,针对传统回声状态网络的不足,提出一种小波回声状态网络流量预测模型(WESN)。首先采用小波分解将网络流量数据分解成高频和低频,然后将高频和低频的数据同时输入回声状态网络进行学习,从而解决了 ESN 模型中存在的病态矩阵问题,最后对模型性能进行仿真测试。结果表明,WESN 可以很好地拟合网络流量变化,具有一定的实际应用价值。
망락류량시일충복잡、다변적비선성혼돈계통,위료획득경가이상적망락류량예측결과,침대전통회성상태망락적불족,제출일충소파회성상태망락류량예측모형(WESN)。수선채용소파분해장망락류량수거분해성고빈화저빈,연후장고빈화저빈적수거동시수입회성상태망락진행학습,종이해결료 ESN 모형중존재적병태구진문제,최후대모형성능진행방진측시。결과표명,WESN 가이흔호지의합망락류량변화,구유일정적실제응용개치。
Network traffic is a complex,non-linear chaotic system variable.In order to obtain a more rational network traffic prediction for the deficiencies in traditional echo networks,network traffic predic-tion model proposed wavelet echo state.Firstly,the wavelet decomposition to decompose the network traf-fic data onto high-frequency and low-frequency.Secondly high frequency and low frequency is inputted to the echo state network for learning.Finally the simulation experiments are carried out to test for the model performance.And this can solve the problem of morbid matrix models of ESN model.The results show that the proposed model can describe the change trend more accurately and improved the prediction accura-cy of network traffic.