Network Traffic Prediction Using a Nonlinear Combination Model

Abstract

The research proposed a nonlinear combination prediction model of network traffic flow in order to improve the prediction results. To get the final forecasting result, three prediction results were fed into a support vector machine for data fusion. Initially, the single models of ARIMA, ARMA, and GM (1, 1) were employed to predict the characteristics of network traffic. According to the results of the simulation, the nonlinear combination forecasting model is better at predicting network traffic than the other models because it can accurately capture the intricate variations in network traffic.