In: Statistics and Probability
Consider the following gasoline sales time series. If needed, round your answers to two-decimal digits.
|
a)
α= | 0.1 | Exponential smoothing | ||||
period | demand | forecast | forecast error=demand value-forecast value | absolute forecast error | squared forcast error | Abs %error |
t | Dt | Ft | et=Dt-Ft | | et | | (et)² | | et/Dt | |
1 | 17 | |||||
2 | 21 | 17 | 4.00 | 4.00 | 16.0 | 19.05% |
3 | 16 | 17.4 | -1.40 | 1.40 | 2.0 | 8.75% |
4 | 24 | 17.26 | 6.74 | 6.74 | 45.4 | 28.08% |
5 | 17 | 17.934 | -0.93 | 0.93 | 0.9 | 5.49% |
6 | 18 | 17.8406 | 0.16 | 0.16 | 0.0 | 0.89% |
7 | 22 | 17.85654 | 4.14 | 4.14 | 17.2 | 18.83% |
8 | 20 | 18.270886 | 1.73 | 1.73 | 3.0 | 8.65% |
9 | 21 | 18.4437974 | 2.56 | 2.56 | 6.5 | 12.17% |
10 | 19 | 18.6994177 | 0.30 | 0.30 | 0.1 | 1.58% |
11 | 16 | 18.7294759 | -2.73 | 2.73 | 7.5 | 17.06% |
12 | 25 | 18.4565283 | 6.54 | 6.54 | 42.8 | 26.17% |
13 | 19.1108755 |
α= | 0.2 | Exponential smoothing | ||||
period | demand | forecast | forecast error=demand value-forecast value | absolute forecast error | squared forcast error | Abs %error |
t | Dt | Ft | et=Dt-Ft | | et | | (et)² | | et/Dt | |
1 | 17 | |||||
2 | 21 | 17 | 4.00 | 4.00 | 16.0 | 19.05% |
3 | 16 | 17.8 | -1.80 | 1.80 | 3.2 | 11.25% |
4 | 24 | 17.44 | 6.56 | 6.56 | 43.0 | 27.33% |
5 | 17 | 18.752 | -1.75 | 1.75 | 3.1 | 10.31% |
6 | 18 | 18.4016 | -0.40 | 0.40 | 0.2 | 2.23% |
7 | 22 | 18.32128 | 3.68 | 3.68 | 13.5 | 16.72% |
8 | 20 | 19.057024 | 0.94 | 0.94 | 0.9 | 4.71% |
9 | 21 | 19.2456192 | 1.75 | 1.75 | 3.1 | 8.35% |
10 | 19 | 19.5964954 | -0.60 | 0.60 | 0.4 | 3.14% |
11 | 16 | 19.4771963 | -3.48 | 3.48 | 12.1 | 21.73% |
12 | 25 | 18.781757 | 6.22 | 6.22 | 38.7 | 24.87% |
13 | 20.0254056 |
forecast for week 13, when α=0.1 = 19.1108755
forecast for week 13, when α=0.2 = 20.0254056
--------------------
b)
An α = 0.2 smoothing constant provides the more accurate forecast, with an overall MSE of 12.19
c)
An α = 0.2 smoothing constant provides the more accurate forecast, with an overall MAE of 2.83 |
d)
An α = 0.1 smoothing constant provides the more accurate forecast, with an overall MAPE of 13.34%