Journal of Forecasting (1995),14, 431-441.
Rob J Hyndman
Abstract: Many modern time series methods, such as those involving non-linear models or non-normal data, frequently lead to forecast densities which are asymmetric or multi-modal. The problem of obtaining forecast regions in such cases is discussed and it is proposed that highest density forecast regions be used. A graphical method for presenting the results is discussed.
Keywords: non-linear time series, non-normal time series, highest density regions, forecast intervals, threshold models.