A Methodology of Forecasting Demands of the Communication Traffic

The Traffic demand of the communication has strong relations to the gross domestic product (GDP). Some statistical models are well known for the demand forecast. As such models, there are the Linear regression Model (LM) and the Auto Regression model (AR). However the LM cannot apply analyzing a traffic demand, because its relations between a GDP and a traffic demand have the non linear shape. Also the AR has problems ,which cannot reflect the impact of social and economical events ,and have big forecasting errors, because a traffic demand has a trend component. Therefore this study considers new a methodology of forecasting demands of the communication traffic, which has high quality by resolving the above problems, by modeling and evaluating social and economical events.