Improved MPEG video traffic model based on time series analysis

In this paper, a novel improved MPEG video traffic model is presented based on time series analysis, while elimination of trend component and MA process is used and the frame-size probability distribution of traffic predicted is transformed to that of real situation. Considered in prediction error, tail of probability distribution, autocorrelation and self-similarity, our model gives attention to both LRD and SRD, shares almost the same self-similarity with real traffic and can provide good result of prediction on both GOP and frame time scale according to a number of simulations.