和声搜索算法优化支持向量机的网络流量预测  

Network Traffic Predicting Based on SVM Optimized by Harmony Search Algorithm

作　　者：丁春莉[1] 李林森[1] 

机构地区：[1]陕西交通职业技术学院,西安710021

出　　处：《微型电脑应用》2017年第1期67-70,共4页Microcomputer Applications

摘　　要：网络流量受到外界因素作用,具有复杂的变化规律,为了改善了网络流量的预测效果,设计了和声搜索算法优化支持向量机的网络流量预测模型(HS-SVM)。首先对当前网络流量预测研究现状进行深入分析,并指出了网络流量的混沌特性,然后采用混沌理论的相应方法确定网络流量的延迟时间和嵌入维数,并对原始网络流量数据进行重构,最后采用HS-SVM建立网络流量预测模型,并与当前其它网络流量预测模型进行了对照模拟测试。HS-SVM能够挖掘和分析网络流量的变化规律,网络流量预测结果要明显优于其它网络流量预测模型,测试结果验证了HS-SVM的可行性和优越性。Network flow is influenced by some external factors, and has complicated variation law. In order to improve prediction effect of network traffic, this paper puts forward a novel network traffic prediction model based on HS-SVM. First of all, current research status of network traffic prediction is analyzed deeply, and chaotic characteristics of network traffic are pointed out; Then, delay time and embedding dimension are determined by chaos theory to reconstruct original network traffic data; Finally, network traffic prediction model is established by HS-SVM and the simulation test is carried out compared with other network traffic prediction models. HS-SVM can mine and analyze the change law of network traffic, prediction results are better than that of other prediction models, and the test results verify feasibility and superiority of HS-SVM.

关 键 词：互联网 网络流量 混沌理论 和声搜索算法 参数选择 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]