Prediction of internet traffic based on Elman neural network

Abstract

Predicting Internet traffic is needed for effective dynamic bandwidth allocation and for quality-of-service (QoS) control strategies implemented at the network edges. In this paper, a method is presented to model and predict the Internet traffic based on Elman neural network (Elman-NN). The traffic is viewed as a time series, which is nonlinear and variant functions. An Elman neural network is employed to model the relationship with a satisfactory accuracy, and the Elman NN-based traffic model is used to conduct prediction for the future traffic. The simulation results show that this method is feasible and efficient to model and predict the traffic.