基于四阶累积量自适应特征提取网络流量预测

Network traffic prediction based on four order adaptive feature extraction

网络流量时间序列受到复杂背景信息干扰时,预测精度不高的问题,提出一种基于四阶累积量自适应特征提取的网络流量预测算法.构建网络流量数据传输结构模型,采用四阶累积量自适应特征提取方法,实现对流量准确预测估计.仿真结果表明,采用该算法进行流量预测,流量预测输出波束的指向性较好,对旁瓣干扰抑制效果较好,说明流量预测的抗干扰能力较强,预测精度高于传统方法.

The network time series is interfered by complex background information,and thus the forecasting is not precise, a network traffic prediction algorithm is thus proposed based on adaptive feature fourth-order cumulant, to make the model of network traffic data transmission structure, to use method of fourth-order cumulant adaptive feature extraction, and to realize ac- curate prediction and consistency estimates for flow . The simulation results show that traffic prediction, traffic prediction output beam directivity is better; sidelobe interference inhibition effect is desirable traffic prediction has strong anti-jamming capability; prediction precision is higher than that of traditional methods.