Econometric Fits of a Linear-Integral-Equation VEC Model to Produce Short-Term Postal Forecasts
摘要
The chapter illustrates how linear integral equations may be specified and econometrically fit to time series to improve predictions of postal variables over a fixed forecasting horizon. The method is demonstrated by fitting a version of the Vector Error Correction model to produce continuous forecasts of US commercial mail volumes over a 1-year horizon.