论文部分内容阅读
为了解决自相似模型难以进行自相似网络流量趋势预测的问题,提出时间序列分析中短时相关模型(自适应自回归模型)的方法用于流量数据的估计;同时为了提高预测精度,提出改进的最小平方格型(modified least squarelattice,MLSL)算法,使模型参数不断递推修正,收敛到最佳值。仿真试验结果验证了短时相关模型在网络流量预测应用中的可行性,实现了自相似网络流量的短期预测,该算法比最小平方(least square,LS)算法均方误差减少20%,具有收敛快、预测精度高的优点,而该算法的计算量减少一半。
In order to solve the problem that the self-similar model is difficult to predict the trend of self-similar network traffic, a short-time correlation model (adaptive autoregressive model) in time series analysis is proposed for the estimation of traffic data. In order to improve the prediction accuracy, an improved Modified least square lattice (MLSL) algorithm, the model parameters continue recursive correction, converged to the best value. Simulation results validate the feasibility of short-term correlation model in network traffic prediction and achieve short-term prediction of self-similar network traffic. Compared with the least squares (LS) algorithm, the mean square error of this algorithm is reduced by 20% Fast, high prediction accuracy, and the algorithm reduces the amount of computation in half.