Forecasting with Adaptive Extended Exponential Smoothing

Abstract

Much of product level forecasting is based upon time series techniques. However, traditional time series fore casting techniques have offered either smoothing constant adaptability or consideration of various time series compo nents, but not both. The purpose of this paper is to present a time series technique newly developed by the author that combines both the inclusion of level, trend, and seasonality and smoothing constant adaptability. Testing of this tech nique, managerial and research implications, and guide lines for use are also presented.