网络时延的混沌特性分析及预测

针对互联网的传输时延处于不确定状态的问题，从网络时延序列表现出的性质出发，将混沌理论引入对网络时延的特性分析。通过对实际网络时延数据的处理，计算验证了网络时延具有混沌特性。并提出了采用混沌一支持向量机回归模型对互联网络时延进行预测的方法，通过实例分析对该方法进行了验证。分析结果证明了混沌—支持向量机模型可以对网络时延的变化趋势进行较为准确的预测，和BP神经网络回归预测相比，混沌—支持向量机对时延的短期预测具有更高精度。

Chaotic Analysis and Prediction of Internet Time-Delay

Aiming at the problem of indeterminacy of Internet time-delay,based on its character,the chaos theory was implemented.By analyzing the actual timedelay data,the validity of its chaotic character was demonstrated.Then a method of time-delay prediction using the chaos-support vector machine(SVM) model was put forward,and the feasibility of the method was validated by practical application.Results of the analysis indicated that the chaos-SVM model could predict the changing trend of network time-delay accurate...