Network Traffic Prediction Based on Machine Learning Model

Abstract

Network traffic engineering is of great significance to large-scale network planning and design, network resource management and network intrusion detection. Traffic modeling and prediction is an important part of network traffic engineering. The traditional traffic time series model is only suitable for analyzing stationary and special non-stationary processes, and it is difficult to describe the complex traffic behavior of large-scale networks. The idea of using support vector machine to estimate function is to select a non-linear map** to map the sample vector from the original space to the high-dimensional feature space, and construct the optimal decision function in the high-dimensional feature space. The principle of structural minimization is used, the loss function is introduced, and the kernel function of the original space is used to replace the point product operation in the high-dimensional feature space for avoiding the complexity. In this paper, support vector machine (SVM) is used to predict network traffic. The network traffic is preprocessed, the parameters are dynamically adjusted, and then the network traffic is predicted. From the actual prediction results, this method has a good prediction effect.