Exponential Smoothing

Exponential smoothing is a method applied to time series data. Exponential smoothing can be defines as the one that is either used to create smoothed data for presentation, or to make forecasts. The time series data is a series of observations. They may be in random order or in noisy process. The single moving average weighs the past observation equally whereas exponential smoothing allocates exponential decreasing weights over time. To further explain, current observations are given comparatively more weight in forecasting than the older observations.

• Exponential smoothing is generally practical to financial market and economic data and it can also be used with separate set of repetitive measurements.
• Exponential smoothing is of two trends such as, double exponential smoothing and triple exponential smoothing. Double exponential smoothing is enhanced to handle inclination and developments whereas triple exponential smoothing is improved to handle parabola trends.
• an exponential weighted moving average with a smoothing constant a represents roughly to simple moving average of length n, where n and a are related by: α = 2/ (n+1) or
  n= (2- α) / α.
• In exponential smoothing older is given relative weight while newer data is specified more weight.
• It is also called averaging it is engaged in making short term forecasts.
• Exponential smoothing has a wait and see approach to change which makes it as an impulsive way and makes people employ exponential smoothing in their daily living.

Procedure:

New forecast = old forecast + a(actual forecast). it gives heaviest weight to more recent information and smaller weight to observations in the more distant past. The motive behind this is that the future may be more reliant upon the current past than on the distant past.

• This method is effective when there is indiscriminate demand and no recurrent variation in the data.
• Exponential smoothing is well known technique for short run forecasting by business interpreters.
• Every fresh forecast is established on the prior forecast in addition to a percentage of the variation between that forecast and the authentic assessment of the time series at that point.

Questions:

• What is exponential smoothing?
• Explain the uses of exponential smoothing?