based on a rolling avg. of past values for a series
easy to use
based on avg. of past values, so it lags a trend (if one exists) and tends to dampen/ smooth turning points, where seasonality or other irregular events
larger qty of average points = greater chance a random event is eliminated
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q
forecasting:
exponential smoothing method
A
fcst created using values from most recent period (extrapolation?)
fcst lags a trend in the series (if there is one)– like simple moving avg. method
mild trend = minimal lag effect, due to using most recent period values– unlike moving avg. method