A 3 Forecasting Final
A 3 Forecasting Final
A 3 Forecasting Final
(a) Three months weighted moving average method to forecast the sales for the months April
December:
Given weights are W1= 3/6, W2= 2/6 and W3= 1/6. We are giving more weight to more recent data.
Formula Used:
Months
Actual Sales, At
(Million)
January
February
March
20
24
27
Forecast
Sales, Ft
(Million)
-
April
31
May
Error (et)
At - Ft
Absolute Value of
Error | et |
{Error
(et)}2
24.83
6.17
6.17
38.07
37
28.50
8.50
8.50
72.25
June
47
33.33
13.67
13.67
186.87
July
53
41
12
12
144
August
52
48.33
3.67
3.67
13.47
September
54
51.50
2.5
2.5
6.25
October
36
53.17
-17.17
17.17
294.81
November
32
44.67
-12.67
12.67
160.53
December
29
37
-8
64
1. (b) Simple Exponential Smoothing ( = 0.6) Method to forecast the sales for the months April
December:
Given that, = 0.6 and the initial forecast for January was $22 million.
Formula used:
= $ 25.2880 Million
*** The formula has been used in the same way using the relevant data for forecasting sales for other
months.
Months
Actual Sales,
At (Million)
Forecast
Sales, Ft
(Million)
January
20
22.00
-2
February
24
20.80
3.2
3.2
10.24
March
27
22.72
4.28
4.28
18.32
April
31
25.29
5.71
5.71
32.6
May
37
28.72
8.28
8.28
68.56
June
47
33.69
13.31
13.31
177.16
July
53
41.67
11.33
11.33
128.37
August
52
48.47
3.53
3.53
12.46
September
54
50.59
3.41
3.41
11.63
October
36
52.64
-16.64
16.64
276.89
November
32
42.65
-10.65
10.65
113.42
December
29
36.27
-7.27
7.27
52.85
Error (et)
At - Ft
Absolute Value of
Error | et |
{Error
(et)}2
4
Actual
Months
Sales,
At (Million)
Forecast
Error (et)
Absolute Value
(Million)
At - Ft
of Error | et |
Sales,
Ft
April
31
25.29
5.71
5.71
May
37
28.72
8.28
8.28
June
47
33.69
13.31
13.31
July
53
41.67
11.33
11.33
August
52
48.47
3.53
3.53
September
54
50.59
3.41
3.41
October
36
52.64
-16.64
16.64
November
32
42.65
-10.65
10.65
December
29
36.27
-7.27
7.27
et = 11.01
| et | = 80.13
Months
Actual Sales, At
(Million)
Forecast
Sales, Ft
(Million)
Error (et)
At - Ft
Absolute Value of
Error | et |
April
May
June
July
August
September
October
November
December
31
37
47
53
52
54
36
32
29
24.83
28.50
33.33
41
48.33
51.50
53.17
44.67
37
6.17
8.50
13.67
12
3.67
2.5
-17.17
-12.67
-8
6.17
8.50
13.67
12
3.67
2.5
17.17
12.67
8
| et | = 84.35
Months
Actual Sales, At
(Million)
Forecast
Sales, Ft
(Million)
April
31
25.29
5.71
5.71
May
37
28.72
8.28
8.28
June
47
33.69
13.31
13.31
July
53
41.67
11.33
11.33
August
52
48.47
3.53
3.53
September
54
50.59
3.41
3.41
October
36
52.64
-16.64
16.64
November
32
42.65
-10.65
10.65
December
29
36.27
-7.27
7.27
Error (et)
At - Ft
Absolute Value of
Error | et |
| et | = 80.13
Recommendation:
Mean Absolute Deviation (MAD) for Simple Exponential Smoothing (=0.6) 8.90 is less than
Mean Absolute Deviation (MAD) for Three Months Weighted Moving Average 9.372. As Simple
Exponential Smoothing forecasts the sales for Dalworth Company with less error, we recommend
Simple Exponential Smoothing (=0.6) over the Three Months Weighted Moving Average
Method.
1. (d) Comparison of the performances of Three Months Weighted Moving Average and Simple
Exponential Smoothing Method (=0.6) using Root Mean Squared Error (RMSE):
Root Mean Squared Error (RMSE) of Three Months Weighted Moving Average:
Months
Actual Sales, At
(Million)
Forecast
Sales, Ft
(Million)
Error (et)
At - Ft
Absolute
Value of
Error | et |
{Error (et)}2
April
31
24.83
6.17
6.17
38.07
May
37
28.50
8.50
8.50
72.25
June
47
33.33
13.67
13.67
186.87
July
53
41
12
12
144
August
52
48.33
3.67
3.67
13.47
September
54
51.50
2.5
2.5
6.25
October
36
53.17
-17.17
17.17
294.81
November
32
44.67
-12.67
12.67
160.53
December
29
37
-8
64
(et)2= 980.25
2
Actual Sales,
At (Million)
Forecast
Sales, Ft
Error (et)
At - Ft
Absolute Value
of Error | et |
{Error (et)}2
(Million)
April
31
25.29
5.71
5.71
32.6
May
37
28.72
8.28
8.28
68.56
June
47
33.69
13.31
13.31
177.16
July
53
41.67
11.33
11.33
128.37
August
52
48.47
3.53
3.53
12.46
September
54
50.59
3.41
3.41
11.63
October
36
52.64
-16.64
16.64
276.89
November
32
42.65
-10.65
10.65
113.42
December
29
36.27
-7.27
7.27
52.85
(et)2= 873.94
July 17
July 24
July 31
August 7
31
24
23
27
28
27
26
F July 10
F July 17
F July 24
F July 31
F August 7
Using Simple Exponential Smoothing Method with = 0.25, the forecasted number of calls for the
week of August 7th is 26 calls.
= 43.4 Surgeries
*** The formula has been used in the same way using the relevant data for forecasting surgeries for
other years.
Year
1993
1994
1995
1996
1997
1998
Actual Demand, At
45
50
52
56
58
Forecast Demand, Ft
41.0000
43.4000
47.3600
50.1440
53.6576
56.2630
Actual Demand, At
45
50
52
56
58
Forecast Demand, Ft
41.0000
44.6000
49.4600
51.7460
55.5746
57.7575
=(
= ( 52 + 50 + 45 ) / 3
= 49 Surgeries
*** The formula has been used in the same way using the relevant data for forecasting surgeries
for other years.
Year
Actual Demand, At
1993
1994
1995
1996
1997
1998
45
50
52
56
58
Forecast Demand,
Ft
49.0000
52.6667
55.3333
*** The formula has been used in the same way using the relevant data for forecasting surgeries for
other years.
Year
Actual Demand, At
1993
1994
1995
1996
1997
1998
45
50
52
56
58
Forecast Demand,
Ft
50.17
53.67
56.33
*** The formula has been used in the same way using the relevant data for forecasting surgeries for
other years.
Year
Actual Demand, At
1993
1994
1995
1996
1997
45
50
52
56
58
Forecast Demand,
Ft
45.8
49
52.2
55.4
58.6
1998
61.8
(f) Comparison of the performances of Simple Exponential Smoothing with = 0.6, Simple
Exponential Smoothing with = 0.9, Three Year Moving Average (MA), Three Year Weighted
Moving Average (WMA) and Regression Model (Y=42.6+3.2X) by using Mean Absolute Deviation
(MAD).
(i) Mean Absolute Deviation of Simple Exponential Smoothing with = 0.6
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand, Ft
41.0000
43.4000
47.3600
50.1440
53.6576
56.2630
Error (et),
At - Ft
4.0000
6.6000
4.6400
5.8560
4.3424
Absolute Value of
Error, | et |
4.0000
6.6000
4.6400
5.8560
4.3424
| et | = 25.4384
Actual
Demand, At
45
50
52
56
58
Forecast
Demand, Ft
41.0000
44.6000
49.4600
51.7460
55.5746
57.7575
Error (et),
At - Ft
4.0000
5.4000
2.5400
4.2540
2.4254
Absolute Value of
Error, | et |
4.0000
5.4000
2.5400
4.2540
2.4254
| et | = 18.6194
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
49.0000
52.6667
55.3333
Error
(et),
At - Ft
7.0000
5.3333
Absolute Value of
Error,
| et |
7.0000
5.3333
| et | = 12.3333
Error
(et),
At - Ft
5.8335
4.3334
Absolute Value of
Error,
| et |
5.8335
4.3334
| et | = 10.1669
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
45.8
49
52.2
55.4
58.6
61.8
Error
(et),
At - Ft
-0.8
1
-0.2
0.6
-0.6
Absolute Value of
Error,
| et |
0.8
1
0.2
0.6
0.6
| et | = 3.2
Recommendation:
From our calculation, we have found that,
MAD 3 year MA (6.1667) > MAD SES (=0.6) (5.0877) > MAD 3 year WMA (5.0835) > MAD SES (=0.9) (3.7239) >
MAD RM(Y=42.6+3.2X) (0.64)
Therefore, we recommend the Regression Model (Y=42.6+3.2X) Method to be used to forecast the
surgeries for the year 1998 as it has the least error.
(g) Comparison of the performances of Simple Exponential Smoothing with = 0.6, Simple
Exponential Smoothing with = 0.9, Three Year Moving Average (MA), Three Year Weighted
Moving Average (WMA) and Regression Model (Y=42.6+3.2X) by using Root Mean Squared Error
(RMSE).
(i) Root Mean Squared Error (RMSE) of Simple Exponential Smoothing (=0.6):
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
41.0000
43.4000
47.3600
50.1440
53.6576
56.2630
Error (et),
At - Ft
4.0000
6.6000
4.6400
5.8560
4.3424
Absolute Value
of Error,
| et |
4.0000
6.6000
4.6400
5.8560
4.3424
{Error (et)}2
16.0000
43.5600
21.5296
34.2927
18.8564
(et)2 = 134.2388
(ii) Root Mean Squared Error (RMSE) of Simple Exponential Smoothing (=0.9):
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
41.0000
44.6000
49.4600
51.7460
55.5746
57.7575
Error (et),
At - Ft
4.0000
5.4000
2.5400
4.2540
2.4254
Absolute
Value of
Error,
| et |
4.0000
5.4000
2.5400
4.2540
2.4254
{Error (et)}2
16.0000
29.1600
6.4516
18.0965
5.8826
(et)2 = 75.5907
(iii) Root Mean Squared Error (RMSE) of Three Year Moving Average (MA):
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
49.0000
52.6667
55.3333
Error (et),
At - Ft
7.0000
5.3333
Absolute Value
of Error,
| et |
7.0000
5.3333
{Error (et)}2
49.0000
28.4444
(et)2 = 77.4444
(iv) Root Mean Squared Error (RMSE) of Three Year Weighted Moving Average (WMA):
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
50.1665
53.6666
56.3332
Error (et),
At - Ft
5.8335
4.3334
Absolute Value
of Error,
| et |
5.8335
4.3334
{Error (et)}2
34.0297
18.7784
(et)2 = 52.8081
Year
1993
1994
1995
1996
1997
1998
Actual
Demand, At
45
50
52
56
58
Forecast
Demand,
Ft
45.8
49
52.2
55.4
58.6
61.8
Error
(et),
At - Ft
-0.8
1
-0.2
0.6
-0.6
Absolute Value
of Error,
| et |
0.8
1
0.2
0.6
0.6
{Error (et)}2
0.64
1
0.04
0.36
0.36
(et)2 = 2.4
Recommendation:
From our calculation, we have found that,
RMSE 3 year MA (6.2227) > RMSE SES (=0.6) (5.1815) > RMSE 3 year WMA (5.1385) > RMSE SES (=0.9) (3.8882)
> RMSE RM(Y=42.6+3.2X) (0.6928)
Therefore, we recommend the Regression Model (Y=42.6+3.2X) Method to be used to forecast the
surgeries for the year 1998 as it has the least error.