Prediction for Network traffic Based on Modified Elman Neural Network

A modified Elman neural network model for the network system is proposed which is nonlinear, multivariable and time-varying. The learning method based on seasonal periodicity is introduced into the model training. And the output traffic of the backbone network at a certain university is selected for the test. Experimental results show this model has better accuracy of prediction. Compared with traditional linear model, BP neural network model and normal Elman neural network model, it has higher accuracy and better adaptability. Finally, abnormal behavior of network traffic can be found on time through the test of adaptive boundary value. And it shows that the model is feasible and effective.