【摘要】 针对网络流量高度自相关、随机性和非线性等时间序列特征,采用支持向量机(SVM)模型进行预测。针对SVM模型中参数难以确定的问题,采用粒子群(PSO)算法进行参数寻优,保证预测的精确度。将PSO-SVM模型预测结果与ARIMA自回归移动平均模型、BP神经网络模型预测结果进行比对,PSO-SVM模型具有更高的预测精度,能够更好地反映网络流量的变化规律。更多还原

【Abstract】 Network traffic has the time series characteristic of highly autocorrelation,randomness and nonlinear,so a support vector machine(SVM) model is proposed for network traffic forecasting.But parameters of SVM model are very difficult to determine,therefore particle swarm optimization(PSO) algorithm is used to search these parameters and make sure the accuracy of SVM model.Compared with autoregressive integrated moving average(ARIMA)model and BP Neural Networks,the forecast results show that the PSO-SVM model has higher prediction precision.It can reflect the variety regulation of the network traffic.更多还原