文档介绍：2010届毕业生
毕业论文
题目: C++基于Garch方法的网络流量预测

院系名称: 计算机学院
专业班级: 计算机应用技术06级03班
学生姓名: 学号:
指导教师: 教师职称:

2010 年 6 月 2 日
C++基于garch方法的网络流量预测
摘要
随着网络技术发展、网络带宽的不断增加、用户数量的急速膨胀以及业务类型的多样化,人们经常会遇到网络拥塞和服务质量低等一系列问题,加强网络管理和改善网络的运行已成为当务之急。掌握网络行为的基本特征,发现网络行为变化的基本规律,构造出反映网络行为的数学模型是流量预测的内在需求。
本系统是基于广义自回归条件异方差garch(Generalized Autoregressive Conditional Heteroscedasty model)模型对流量进行预测。系统利用网络侦听原理通过套接字来实现对局域网流量的监控及截获并统计,再通过garch方法对网络实时流量计算,得出均值和条件方差来实现对网络流量的预测。
系统经过测试,能够较好的预测出流量区间。最后,本系统还构建了以UDP洪泛攻击的模拟测试实验,测试的结果系统正常显示由于攻击导致的流量异常,由此发出预警信息,告知用户可能受到网络攻击,从而做到自我保护,提高网络质量。
关键字:预测,garch,异常,网络流量,套接字
Network Traffic Prediction Based On Garch Model
Abstract
With the development work technology,work bandwidth,the rapid expansion of the number of users and diversification of business types,people often encounter the problem work congestion and low quality of service, strengthening work management and improveing the operation of work has e a top the basic characteristics work behavior,finding the basic rule of changes work behavior, creating a reflection of the mathematical model work behavior is the inherent demand work traffic prediction.
The system is based on the Generalized Autoregressive Conditional Heteroscedasticity garch (Generalized Autoregressive Conditional Heteroscedasty model) model to predict system uses the principle of work through the listening socket to achieve the monitoring of LAN’s traffic, interception work traffic and Statistics,puting real-time traffic through garch model to obtain the average and conditional variance to achieve the prediction of work traffic.
The system tested, that can better predict work traffic’s last,we set up a experiment based on UDP attacking to test the results show that the outlier traffic which was caused by attack, informing the user may be subject work attacks, in order to protective measures by yourself,work quality of service.
Key words: prediction,garch,work traffic,socket
目