基于 LMD 改进 GPR 优化的网络流量预测.

It is easy for ant colony algorithm to fall into local optimum and poor accuracy in network traffic prediction. An improved ant colony optimization GPR prediction algorithm based on LMD is proposesd. Firstly,according to the complexity of network traffic,local mean decomposition(LMD) is used to decompose the network traffic into multiple related subsequences. Secondly,Gaussian process regression(GPR) is applied to model and analyze the network traffic subsequences. Ant colony algorithm is used to optimize the super parameters. Thirdly,by introducing the line of sight angle parameter to control the line of sight of ants in search,the local search ability of ants is improved. Fourthly,the step size of ant colony algorithm is updated by Levy flight to improve the global search. Experimental results show that the improved ant colony algorithm finds a better value. Compared with the original GPR algorithm,the improved ant colony optimization GPR algorithm after LMD decomposition can predict network traffic,better serve the trend of network traffic and improve the prediction effect,in a maintaining network security role. [ABSTRACT FROM AUTHOR]