Abstract：Concerning the nonlinear, multivariable and time-varying qualities of neural network, a modified Elman neural network model was proposed. The learning method based on seasonal periodicity was introduced into the model training. And the output traffic of the backbone network of a certain university was given. The experimental results show that this model has better predication effect. Compared with the traditional linear model, the BP neural network model and the normal Elman neural network model, it has higher precision and better adaptability. Finally, abnormal behaviors of network traffic can be found on time through test of adaptive boundary value method, which proves that the model is feasible and effective.
2010, Vol. 30 
