Exponential Smoothing
Exponential Smoothing
Exponential Smoothing model is a “natural advancement” of moving average models. Simple exponential smoothing is the most basic smoothing model generating a stable forecast.
The smoothing co-efficient alpha (?) is the weight on the most recent observation and comparable to 1/M in moving average models. It can be any number between 0 to 1.
M = 4 ? weight is 25%
? = 25% ? recent observation is weighted 25% and the older observations are weighted 75%.
This results in a forecast that is a straight-line but adjusts every month with new data. It is appropriate when the underlying time series behaves like a constant + Noise; when the mean is moving very slowly over time.
Equation for Simple Exponential Smoothing