In: Statistics and Probability
Question2:
A manager use his company profit time series for one goods to calculate forecast for upcoming years. He use
F1 = Moving Average Method for 3 years
F2 = Weight Moving Average for 3 years
F3 = Exponential Smoothing Method with α = 0.2
F4 = Exponential Smoothing Method with α = 0.7
His calculation represent in the following table. Gide the manager to chose the better method by calculate MAD and MES for each method
t |
y |
F1=MA(3) |
F2=WMA(3) |
F3=exp(0.2) |
F4=exp(0.7) |
1 |
58 |
- |
- |
- |
- |
2 |
46 |
- |
- |
58.0 |
58.0 |
3 |
55 |
- |
- |
55.6 |
49.6 |
4 |
39 |
53.0 |
53.5 |
55.5 |
53.4 |
5 |
42 |
46.7 |
42.9 |
52.2 |
43.3 |
6 |
63 |
45.3 |
42.7 |
50.1 |
42.4 |
7 |
54 |
48.0 |
56.4 |
52.7 |
56.8 |
8 |
55 |
53.0 |
54.6 |
53.0 |
54.8 |
9 |
61 |
57.3 |
55.6 |
53.4 |
55.0 |
10 |
52 |
56.7 |
59.1 |
54.9 |
59.2 |
11 |
- |
56.0 |
54.1 |
54.3 |
54.2 |
F1:
period | demand | forcast | forecast error=demand value-forecast value | absolute forecast error | squared forcast error |
t | Dt | Ft | et=Dt-Ft | | et | | (et)² |
1 | 58 | ||||
2 | 46 | ||||
3 | 55 | ||||
4 | 39 | 53.000 | -14.00 | 14.00 | 196.00 |
5 | 42 | 46.700 | -4.70 | 4.70 | 22.09 |
6 | 63 | 45.300 | 17.70 | 17.70 | 313.29 |
7 | 54 | 48.000 | 6.00 | 6.00 | 36.00 |
8 | 55 | 53.000 | 2.00 | 2.00 | 4.00 |
9 | 61 | 57.300 | 3.70 | 3.70 | 13.69 |
10 | 52 | 56.700 | -4.70 | 4.70 | 22.09 |
MAD/MAE= Σ |et|/n = 7.54
MSE= Σ(et)²/n = 86.74
---------------------------------------------------------------------------
F2:
period | demand | forcast | forecast error=demand value-forecast value | absolute forecast error | squared forcast error |
t | Dt | Ft | et=Dt-Ft | | et | | (et)² |
1 | 58 | ||||
2 | 46 | ||||
3 | 55 | ||||
4 | 39 | 53.500 | -14.50 | 14.50 | 210.25 |
5 | 42 | 42.900 | -0.90 | 0.90 | 0.81 |
6 | 63 | 42.700 | 20.30 | 20.30 | 412.09 |
7 | 54 | 56.400 | -2.40 | 2.40 | 5.76 |
8 | 55 | 54.600 | 0.40 | 0.40 | 0.16 |
9 | 61 | 55.600 | 5.40 | 5.40 | 29.16 |
10 | 52 | 59.100 | -7.10 | 7.10 | 50.41 |
MAD/MAE= Σ |et|/n = 7.29
MSE= Σ(et)²/n = 101.23
----------------------------------------------
F3
period | demand | forcast | forecast error=demand value-forecast value | absolute forecast error | squared forcast error |
t | Dt | Ft | et=Dt-Ft | | et | | (et)² |
1 | 58 | ||||
2 | 46 | 58.000 | -12.00 | 12.00 | 144.00 |
3 | 55 | 55.600 | -0.60 | 0.60 | 0.36 |
4 | 39 | 55.500 | -16.50 | 16.50 | 272.25 |
5 | 42 | 52.200 | -10.20 | 10.20 | 104.04 |
6 | 63 | 50.100 | 12.90 | 12.90 | 166.41 |
7 | 54 | 52.700 | 1.30 | 1.30 | 1.69 |
8 | 55 | 53.000 | 2.00 | 2.00 | 4.00 |
9 | 61 | 53.400 | 7.60 | 7.60 | 57.76 |
10 | 52 | 54.900 | -2.90 | 2.90 | 8.41 |
MAD/MAE= Σ |et|/n = 7.33
MSE= Σ(et)²/n = 84.32
-----------------------------
F4
period | demand | forcast | forecast error=demand value-forecast value | absolute forecast error | squared forcast error |
t | Dt | Ft | et=Dt-Ft | | et | | (et)² |
1 | 58 | ||||
2 | 46 | 58.000 | -12.00 | 12.00 | 144.00 |
3 | 55 | 49.600 | 5.40 | 5.40 | 29.16 |
4 | 39 | 53.400 | -14.40 | 14.40 | 207.36 |
5 | 42 | 43.300 | -1.30 | 1.30 | 1.69 |
6 | 63 | 42.400 | 20.60 | 20.60 | 424.36 |
7 | 54 | 56.800 | -2.80 | 2.80 | 7.84 |
8 | 55 | 54.800 | 0.20 | 0.20 | 0.04 |
9 | 61 | 55.000 | 6.00 | 6.00 | 36.00 |
10 | 52 | 59.200 | -7.20 | 7.20 | 51.84 |
MAD/MAE= Σ |et|/n = 7.77
MSE= Σ(et)²/n = 100.25
--------------------------------------------------------
Hence we can say that F3 is the best method
Please revert back in case of any doubt.
Please upvote. Thanks in advance.