Research On Prediction Of Network Traffic Flow Based On Chaos And SVM Techniques

The birth of the Internet has brought great convenience to the daily lives of people especially on learning and work. However, with the rapid development of Internet, it also brings a series of problems:due to Internet business approximately to exponential explosion, and because the randomness and spontaneity, this to the network itself that limited load constitutes a serious challenge. As the network has the dynamic and uncertainty so sudden, etc, to the control condition, if it sets too loose, it will too hard to catch the ideal control effect; If set too tight, it could lead to cannot make full use of the network resources. So what time should be controlled in becomes an important problem in flow control related research. Researchers have put forward various related technology, one considered effective is chaos prediction technology. Chaos theory is better able to describe network flow characteristics, has the good prediction accuracy. However, whether and which algorithm with chaos theory can adopted to achieve better forecasting effect?This paper firstly introduces the present situation of network traffic prediction, and discusses the relevant prediction technology, due to the obvious sudden and spontaneous network traffic; this paper further proposed the network traffic chaos characteristics. Based on the simulation experiments, the simulation results predicted the network traffic well. Considering the chaos theory and network flow related properties, with support vector machine, this paper further proposed a regression theory based on chaos-support vector machine theory network traffic prediction. Simulation results show that this based on chaos-support vector machine theory network traffic prediction relative to traditional network traffic prediction and with much more precise forecasting ability. So it maybe has a better support to the network flow control of relevant research.