Business Forecasting with Exponential Smoothing: Computation of Prediction Intervals

The problem considered in this paper is how to find reliable prediction intervals with simple exponential smoothing and trend corrected exponential smoothing. Methods for constructing prediction intervals based on linear approximation and bootstrapping are proposed. A Monte Carlo simulation study, in which the proposed methods are compared, indicates that the most reliable intervals can be obtained with a parametric form of the bootstrap method. An application of the method to predicting Malaysian GNP per capita is considered