In: Statistics and Probability
Month | ||||||
1 | 37 | |||||
2 | 44 | |||||
3 | 35 | |||||
4 | 50 | |||||
5 | 34 | |||||
6 | 30 | |||||
7 | 50 | |||||
8 | 29 | |||||
9 | 36 | |||||
10 | 35 | |||||
11 | 41 | |||||
12 | 45 |
a. Show the naive forecast, an exponential smoothing forecasts using α = 0.2, and a 3-month moving average forecast.
b. Compare the MFE, MSE, and MAPE on the models
c. Make a conclusion on which model to use.
d. Find the alpha (smoothing constant) that minimizes the MSE.
a)
exponential smoothing α=0.2
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 | 37 | |||||
2 | 44 | 37 | 7.00 | 7.00 | 49.00 | 15.91% |
3 | 35 | 38.4 | -3.40 | 3.40 | 11.56 | 9.71% |
4 | 50 | 37.72 | 12.28 | 12.28 | 150.80 | 24.56% |
5 | 34 | 40.176 | -6.18 | 6.18 | 38.14 | 18.16% |
6 | 30 | 38.9408 | -8.94 | 8.94 | 79.94 | 29.80% |
7 | 50 | 37.15264 | 12.85 | 12.85 | 165.05 | 25.69% |
8 | 29 | 39.722112 | -10.72 | 10.72 | 114.96 | 36.97% |
9 | 36 | 37.5776896 | -1.58 | 1.58 | 2.49 | 4.38% |
10 | 35 | 37.2621517 | -2.26 | 2.26 | 5.12 | 6.46% |
11 | 41 | 36.8097213 | 4.19 | 4.19 | 17.56 | 10.22% |
12 | 45 | 37.6477771 | 7.35 | 7.35 | 54.06 | 16.34% |
MFE= Σet/n = 0.96
MSE= Σ(et)²/n = 62.61
MAPE= Σ | et/Dt |/n = 18.02%
==============================
naive forecast
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 | 37 | |||||
2 | 44 | 37 | 7.00 | 7.00 | 49.00 | 15.91% |
3 | 35 | 44 | -9.00 | 9.00 | 81.00 | 25.71% |
4 | 50 | 35 | 15.00 | 15.00 | 225.00 | 30.00% |
5 | 34 | 50 | -16.00 | 16.00 | 256.00 | 47.06% |
6 | 30 | 34 | -4.00 | 4.00 | 16.00 | 13.33% |
7 | 50 | 30 | 20.00 | 20.00 | 400.00 | 40.00% |
8 | 29 | 50 | -21.00 | 21.00 | 441.00 | 72.41% |
9 | 36 | 29 | 7.00 | 7.00 | 49.00 | 19.44% |
10 | 35 | 36 | -1.00 | 1.00 | 1.00 | 2.86% |
11 | 41 | 35 | 6.00 | 6.00 | 36.00 | 14.63% |
12 | 45 | 41 | 4.00 | 4.00 | 16.00 | 8.89% |
MFE= Σet/n = 0.73
MSE= Σ(et)²/n = 142.73
MAPE= Σ | et/Dt |/n = 26.39%
=====================
3 month moving average-
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 | 37 | |||||
2 | 44 | |||||
3 | 35 | |||||
4 | 50 | 38.6666667 | 11.33 | 11.33 | 128.44 | 22.67% |
5 | 34 | 43 | -9.00 | 9.00 | 81.00 | 26.47% |
6 | 30 | 39.6666667 | -9.67 | 9.67 | 93.44 | 32.22% |
7 | 50 | 38 | 12.00 | 12.00 | 144.00 | 24.00% |
8 | 29 | 38 | -9.00 | 9.00 | 81.00 | 31.03% |
9 | 36 | 36.3333333 | -0.33 | 0.33 | 0.11 | 0.93% |
10 | 35 | 38.3333333 | -3.33 | 3.33 | 11.11 | 9.52% |
11 | 41 | 33.3333333 | 7.67 | 7.67 | 58.78 | 18.70% |
12 | 45 | 37.3333333 | 7.67 | 7.67 | 58.78 | 17.04% |
MFE= Σet/n = 0.81
MSE= Σ(et)²/n = 72.96
MAPE= Σ | et/Dt |/n = 20.29%
b)
MSE is smallest for using exponential smooting
MFE is smallest using naive method
MAPE is smallest for exponential method
c) exponential smooting method should be use
d)
alpha (smoothing constant) = 0.10 minimizes the MSE.