A comprehensive review on hybrid network traffic prediction model

Network traffic is a typical nonlinear time series. As such, traditional linear and models are inadequate to describe the multi-scale characteristics of traffic, thus compromising prediction accuracy. Therefore, research date has tended focus on hybrid rather than non-linear ones. Generally, model adopts two or more methods as combined modelling analyze then predict network traffic. Against this backdrop, paper will review past conducted models. The concludes with summary strengths limitations existing which use optimization decomposition techniques, respectively. These techniques have been identified major contributing factors in constructing accurate fast response prediction.