In moving averages, all observations are weighted equally, whereas in exponential smoothing, the weights are assigned in exponentially decreasing order as the observation gets older. This ensures that the recent or latest observations are given more weightage as compared to older observations and thus can forecast on the basis of recent observations.
You have completed the preceding recipes and the AirPassengers
dataset is available or loaded in R.
Perform the following steps with R:
> library(forecast) > t = ets(AirPassengers) > t Output: ETS(M,Ad,M) Call: ets(y = AirPassengers) Smoothing parameters: alpha = 0.7322 beta = 0.0188 gamma = 1e-04 phi = 0.98 Initial states: l = 120.9759 b = 1.8015 s=0.8929 0.7984 0.9211 1.0604 1.2228 1.2324 1.1107 0.9807 0.9807 1.0106 0.8843 0.9051 sigma: 0.0368 AIC AICc BIC 1395.092 1400.564 1448.548...