Which of the 3 smoothing models should be used and how does one know which to use other than the fact that the most accurate model must be selected?There is no cook-book formula to make that decision.Remember that software and technology is used to make these determinations. So, the task is not so demanding and time-intensive.In general, the simple moving avg is used when there is true random behavior in the data and no pattern that might be important over time.If there is a relationship that might be expected with passage of time, the weighted moving average may be more applicable so that you can weigh the more recent data differently than the older data.Single exponential smoothing is just a form of weighted moving average in which alpha (the smoothing coefficient) is your weight. The most recent data get a weight of alpha. The prior data values are also weighted but their importance in the analysis drops at an exponetial rate. That is because they go into calculation of a forecast value. But then that forecast value goes into
This is the end of the preview.
access the rest of the document.
Holt, Time series analysis, smoothing models, exp smoothing model