基于神经网络模型的网络流量预测综述  

A Survey of Network Traffic Prediction Based on Neural Network Models

作　　者：杜爽 徐展琦[1] 马涛[1] 杨帆[1] DU Shuang;XU Zhanqi;MA Tao;YANG Fan(State Key Laboratory of Integrated Service Networks,Xidian University,Xi’an 710071,China)

机构地区：[1]西安电子科技大学综合业务网理论及关键技术国家重点实验室

出　　处：《无线电通信技术》2020年第2期216-222,共7页Radio Communications Technology

基　　金：国家自然科学基金项目(61572391)~~

摘　　要：网络流量预测技术可以帮助运营商准确预估网络的使用情况,合理分配并高效利用网络资源,以满足日益增长且多样化的用户需求。大量研究表明,神经网络的预测性能高于其他经典预测方法,在网络流量预测中的应用潜力巨大。总结网络流量预测的一些关键因素和预测误差评判方法,重点介绍近年提出的可用于预测的神经网络模型原理、相关预测方法及其属性比较,旨在为包括流量在内的各种网络参数预测技术的学术研究与实际应用提供可借鉴的方法和手段。Network traffic prediction technology can help operators accurately predict the status of networks usage,and operators could therefore rationally allocate and efficiently utilize network resources to meet the growing and diverse user needs.Numerous studies have shown that the prediction performance using neural networks is higher compared to other classical prediction methods,and they have the great potential in network traffic prediction.This paper summarizes some key factors and prediction error evaluation methods for network traffic prediction.It focuses on introducing principles,relevant prediction methods and characteristic comparison of several neural network models proposed for prediction in recent years.The efforts made in this paper aim to provide some reference methods and means for academic research and practical application of prediction technologies for network parameters like network traffic.

关 键 词：网络流量 预测模型 神经网络 网络优化 

分 类 号：TN927[电子电信—通信与信息系统]