Network traffic prediction and applications based on time series model

Network traffic prediction is a very complex and difficult issue in the network management and design. This paper shows a model with a new algorithm (MLSL), and the model parameters can be modified by the new algorithm, which improves the adaptive ability of the model and makes the model adaptive function. Simulation and actual network traffic data experiment has proved that this algorithm has the advantage of high prediction accuracy and fast convergence, and its computing complexity is lower than other related algorithms.