基于SSA和LWLR组合的网络延迟预测  

Internet Network Delay Prediction Using the Combination of Singular Spectrum Analysis and Locally Weighted Linear Regression

作　　者：唐政治 曾学文[1,2] 郭志川 TANG Zhengzhi;ZENG Xuewen;GUO Zhichuan(National Network New Media Engineering Research Center,Institute of Acoustics,Chinese Academy of Sciences,Beijing,100190,Chnia;University of Chinese Academy of Sciences,Beijing,100049,China)

机构地区：[1]中国科学院声学研究所,国家网络新媒体工程技术研究中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《网络新媒体技术》2021年第3期45-50,70,共7页Network New Media Technology

基　　金：国家科技重大专项“新一代宽带无线移动通信网”子课题“5G与信息中心网(ICN)融合技术研发”(编号:2017ZX03001019)。

摘　　要：针对互联网网络端到端延迟因其非线性和非平稳性(时变)特性而导致其准确性预测的难题,提出了一种基于滑动窗口的奇异谱分析和局部加权线性回归的组合预测算法。并采用互联网所采集的RTT数据来验证提出的预测算法,实验结果表明,预测算法具有快速、精确的特点,是一种很适合实际工程应用的预测算法。Aiming at the hard problem that the Internet end-to-end delay is difficult to predict due to its nonlinear and non-stationary(time-varying)characteristics,in this paper,we propose a novel prediction algorithm using the combination of the singular spectrum analysis and locally weighted linear regression based on a sliding window.Finally,the RTT data collected by the Internet is used to verify the proposed prediction algorithm.The experimental results show that the proposed prediction algorithm is fast and accurate,and it is a prediction algorithm suitable for practical engineering applications.

关 键 词：奇异谱分析 局部加权线性回归 端到端延迟 预测算法 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术]