|
Accurate prediction of the future network traffic plays an important role in various network problems (e.g. traffic engineering, quality of service provisioning, etc.). However, modern network communication is extremely complicated, which makes the tasks of modeling and predicting network behavior very difficult. To this end, besides the traditional approaches (e.g., ARIMA), there are some studies exploiting Deep Learning techniques such as Recurrent Neural Network to estimate the traffic volume. In this paper, we propose a highly accurate traffic prediction algorithm by leveraging the Diffusion Convolutional RNN, for spatial-temporal modeling and estimating the future traffic volume of the network's links. We have conducted experiments using the Abilene dataset and the results show that our proposed approach increases 15% in the prediction accuracy.
|
