Business Statistics: For Contemporary Decision Making, 7th Edition

SUPPLEMENT 3

ADVANCED EXPONENTIAL SMOOTHING

The exponential smoothing technique presented in the previous section is a mechanism for combining information from present actual data and forecasts of present data (which incorporate information from previous data) to forecast data for ensuing time periods. Weights are assigned by the forecaster to the importance of the current actual data versus the forecast of the current data in making the forecast. There are other factors that can be taken into consideration in exponential smoothing. In particular, the time-series components, trend, and seasonal effects can be incorporated into the forecast when they exist. First, we will consider exponential smoothing which includes trend effects.

EXPONENTIAL SMOOTHING WITH TREND EFFECTS:: HOLT's METHOD

One exponential smoothing technique which includes trend effects is Holt's two-parameter method. Holt's technique uses weights to smooth the trend in a manner similar to that which we used in exponential smoothing in the previous section. Holt's two-parameter method uses the following three equations to accomplish this:

In simple exponential smoothing, the smoothed values were the forecasts. The next smoothed value was determined by weighing the actual value by α and the previous forecast (the last smoothed value) by 1 – α. Now we have introduced trend into consideration. Notice from the last two ...