What is Exponential Smoothing?
Exponential smoothing is a forecasting method in which current and past values of a time series are used to predict future values of the series. The forecast is based on a weighted average of current and past values, with the recent values carrying more weight. Example: Using historical values for gross fixed assets from 2002 – 2004 and a smoothing constant (alpha) of .5, what is the predicted value for gross fixed assets in 2005? 12/15/2004 12/15/2003 12/15/2002 Gross Fixed Assets $5,800 $4,500 $6,000 Forecasting with exponential smoothing (alpha = .5): Prediction for 12/2003 made in 12/2002 = $6000 Prediction for 12/2004 made in 12/2003 = .5 * $6000 + .5 * $4500 = $5250 Prediction for 12/2005 made in 12/2004 = .5 * $5250 + .5 * $5800 = $5525 [Top] What is adjusted Holt-Winters Exponential Smoothing? The adjusted Holt-Winters Exponential Smoothing algorithm is a forecasting method that extends simple exponential smoothing to allow for trends. [Top] How did we decide which prediction a
