利用IBBO优化LSSVM的网络流量预测模型  

Network Traffic Forecasting Model Based on IBBO-LSSVM

作　　者：嵇可可[1] 张光会[2] 

机构地区：[1]江苏食品药品职业技术学院信息工程系,江苏淮安223003 [2]湖南大学信息科学与工程学院,长沙410082

出　　处：《小型微型计算机系统》2014年第11期2454-2458,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60702065)资助

摘　　要：针对网络流量预测过程中的LSSVM参数优化问题,提出一种改进生物地理学(IBBO)算法优化最小二乘支持向量机(LSSVM)的网络流量预测模型(IBBO-LSSVM).该模型采用相空间重构网络流量学习样本,利用LSSVM对网络流量进行建模,并运用改进生物地理学算法优化模型参数,最后进行网络流量预测实例分析.结果表明,IBBO-LSSVM可以对复杂、多变的网络流量变化特点进行拟合,获得了较高的预测精度,为具有混沌性的网络流量提供了一种新的预测模型.Network traffic is nonlinear and random change rule. In order to obtain better forecasting results, a novel network traffic forecasting model based on improved biogeography-based optimization algorithm and least squares support vector machine （ IBBO- LSSVM ） was proposed. Firstly, the learning samples of network traffic were obtained by phase space reconstruction, and then the east squares support vector machine was used to establish the network traffic forecasting model which parameters were optimized by im- proved biogeography-based optimization algorithm. The simulation results show that IBBO-LSSVMN can improve the forecasting ac- curacy of the network traffic,and provide a new research approach for the network traffic forecasting.

关 键 词：网络流量 最小二乘支持向量机 生物地理学算法 预测精度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]