Due to the complicated and ever-changing characteristics of Internet, 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), which updates the model parameters and improves the adaptive ability of the model. Generating Self-similar series to simulate the network traffic and do some test, the experimental results of simulating and actual network traffic data have proved that this algorithm has the advantage of high accuracy and fast convergence.
展开▼
