data(airpass)
summary(airpass)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 104.0 180.0 265.5 280.3 360.5 622.0
data(airpass)
plot(airpass)
As you can see by the plot this series shows a definite positive trend between 1960 and 1971 and also some seasonality. Based on the above chart, it seems the count of international airline passengers spikes during the summertime through the holiday season, and then dips back down near the end of each year.
fsmoothing=ses(airpass, h=1)
plot(fsmoothing)
summary(fsmoothing)
##
## Forecast method: Simple exponential smoothing
##
## Model Information:
## Simple exponential smoothing
##
## Call:
## ses(y = airpass, h = 1)
##
## Smoothing parameters:
## alpha = 0.9999
##
## Initial states:
## l = 111.9892
##
## sigma: 33.8299
##
## AIC AICc BIC
## 1733.787 1733.958 1742.696
##
## Error measures:
## ME RMSE MAE MPE MAPE MASE ACF1
## Training set 2.22249 33.59418 25.68134 0.3758224 8.957112 0.8017825 0.3028609
##
## Forecasts:
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jan 1972 431.9958 388.641 475.3506 365.6904 498.3012
round(accuracy(fsmoothing))
## ME RMSE MAE MPE MAPE MASE ACF1
## Training set 2 34 26 0 9 1 0
fc1=holt(airpass, h=1)
plot(fc1)
summary(fc1)
##
## Forecast method: Holt's method
##
## Model Information:
## Holt's method
##
## Call:
## holt(y = airpass, h = 1)
##
## Smoothing parameters:
## alpha = 0.9999
## beta = 1e-04
##
## Initial states:
## l = 119.7517
## b = 1.5963
##
## sigma: 34.0125
##
## AIC AICc BIC
## 1737.295 1737.729 1752.144
##
## Error measures:
## ME RMSE MAE MPE MAPE MASE
## Training set 0.5677345 33.53682 25.61104 -0.3624936 8.996892 0.7995878
## ACF1
## Training set 0.3024925
##
## Forecasts:
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jan 1972 433.6004 390.0116 477.1892 366.937 500.2638
round(accuracy(fc1))
## ME RMSE MAE MPE MAPE MASE ACF1
## Training set 1 34 26 0 9 1 0
fc2=hw(airpass, seasonal="additive", h=1)
fc3=hw(airpass, seasonal="multiplicative", h=1)
plot(fc2)
plot(fc3)
summary(fc2)
##
## Forecast method: Holt-Winters' additive method
##
## Model Information:
## Holt-Winters' additive method
##
## Call:
## hw(y = airpass, h = 1, seasonal = "additive")
##
## Smoothing parameters:
## alpha = 0.9935
## beta = 2e-04
## gamma = 6e-04
##
## Initial states:
## l = 120.9608
## b = 1.3934
## s = -29.1816 -54.3842 -20.7169 15.0727 65.1554 66.1846
## 33.5822 -4.232 -8.0946 -3.8205 -34.3364 -25.2288
##
## sigma: 18.0471
##
## AIC AICc BIC
## 1565.872 1570.729 1616.359
##
## Error measures:
## ME RMSE MAE MPE MAPE MASE ACF1
## Training set 0.9638247 17.01495 12.81203 0.3539818 5.224715 0.3999971 0.1875311
##
## Forecasts:
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jan 1972 437.2863 414.158 460.4145 401.9146 472.6579
summary(fc3)
##
## Forecast method: Holt-Winters' multiplicative method
##
## Model Information:
## Holt-Winters' multiplicative method
##
## Call:
## hw(y = airpass, h = 1, seasonal = "multiplicative")
##
## Smoothing parameters:
## alpha = 0.3146
## beta = 0.0071
## gamma = 0.5977
##
## Initial states:
## l = 120.3796
## b = 1.7757
## s = 0.9298 0.7946 0.9024 1.0451 1.1338 1.1388
## 1.0529 0.9638 1.0349 1.0807 0.9854 0.9378
##
## sigma: 0.0407
##
## AIC AICc BIC
## 1405.654 1410.511 1456.141
##
## Error measures:
## ME RMSE MAE MPE MAPE MASE ACF1
## Training set 1.256973 10.63256 7.790649 0.2182707 2.914411 0.2432275 0.2135914
##
## Forecasts:
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jan 1972 445.8901 422.6577 469.1225 410.3592 481.421
Using simple exponential smoothing method the alpha R chose was a value near 1 so it is implying more recent history is being used. the RMSE is 33.59418
Using Hotls method the alpha R chose was a value near 1 so it is implying more recent history is being used. The beta is close to zero indicating the trend is weak. The RMSE is 33.53682
Using Holt-Winter additive method the alpha R chose was a value near 1 so it is implying more recent history is being used. The beta is close to zero indicating the trend is weak. The gamma captures the strength of the seasonailty so with this method there is no seasonality. The RMSE is 17.01495
Using Holt-Winter multiplicative method the alpha R chose is closer to zero implying R went further back into the history. The beta is close to zero indicating the trend is weak. The gamma captures the strength of the seasonailty so with this method there is some seasonality. The RMSE is 10.63256
In conclusion based on the RMSE we will go with the Holt-Winter multiplicative method as is has the lowest RMSE.