Week-4 Exponential Smoothing Methods – Short Version
Some of the slides are adapted from the lecture notes provided by Prof. Antoine Saure and Prof. Rob Hyndman
Business Forecasting Analytics
ADM 4307 – Fall 2021
Exponential Smoothing Methods
Ahmet Kandakoglu, PhD
04 October, 2021
Exponential Smoothing
• Exponential smoothing methods are weighted averages of past observations, with weights
decaying exponentially as the observations get older
• The most recent observations usually provide the best guide as to the future
Fall 2021 ADM 4307 Business Forecasting Analytics 2
Example: Australian Holiday Tourism
aus_holidays <- tourism %>% filter(Purpose == “Holiday”) %>% summarise(Trips = sum(Trips)/1000)
fit <- aus_holidays %>% %>% model(ETS(Trips))
Fall 2021 ADM 4307 Business Forecasting Analytics 3
> report(fit)
Series: Trips
Model: ETS(M,N,A)
Smoothing parameters:
alpha = 0.3484054
gamma = 0.0001000018
Initial states:
l[0] s[0] s[-1] s[-2] s[-3]
9.727072 -0.5376106 -0.6884343 -0.2933663 1.519411
sigma^2: 0.0022
AIC AICc BIC
226.2289 227.7845 242.9031
> fit %>% forecast(h = 8) %>% autoplot(aus_holidays)
Example: Australian Holiday Tourism
gg_tsresiduals(fit)
accuracy(fit)
Fall 2021 ADM 4307 Business Forecasting Analytics 4
# A tibble: 1 x 10
.model .type ME RMSE MAE MPE MAPE MASE RMSSE ACF1
1 ETS(Trips) Training 0.0520 0.428 0.331 0.334 3.45 0.798 0.793 -0.0642
Business Forecasting Analytics
ADM 4307 – Fall 2021
Exponential Smoothing Methods
Fall 2021 ADM 4307 Business Forecasting Analytics 5