LE ANH KHOA BABAIU14124

QUANTITATIVE METHODS FOR BUSINESS

ASSIGNMENT: FORECASTING

515: Data collected on the yearly demand for 50pound bags of fertilizer at Wallace Garden
Supply are shown in the following table. Develop a 3year moving average to forecast sales.
Then estimate demand again with a weighted moving average in which sales in the most recent
year are given a weight of 2 and sales in the other 2 years are each given a weight of 1. Which
method do you think is best?
DEMAND FOR FERTILIZER
YEAR
(1,000S OF BAGS)
1 4
2 6
3 4
4 5
5 10
6 8
7 7
8 9
9 12
10 14
11 15

Solution:
*We use the 3year moving average to forecast sales. We have the table as following:
FERTILIZER DEMAND
YEAR 3YEAR MOVING AVERAGE ERROR
(1,000S OF BAGS)
1 4
2 6
3 4
4 5 (4 + 6 + 4)/3 = 4.67 0.33
5 10 (6 + 4 + 5)/3 = 5.00 5.00
6 8 (4 + 5 + 10)/3 = 6.33 1.67
7 7 (5 + 10 + 8)/3 = 7.67 0.67
8 9 (10 + 8 + 7)/3 = 8.33 0.67
9 12 (8 + 7 + 9)/3 = 8.00 4.00
10 14 (7 + 9 + 12)/3 = 9.33 4.67
11 15 (12 + 14 + 15)/3 = 13.67 3.33
MAD
= 2.5425

*We use the weighted moving average to forecast sales. We have the table as following:
DEMAND FOR
YEA 3YEAR WEIGHTED MOVING
FERTILIZER ERROR
R AVERAGE
(1,000S OF BAGS)
1 4
2 6
LE ANH KHOA BABAIU14124

3 4
[(2 X 4) + (1 X 6) + (1 X 4)]/4
4 5 0.50
= 4.50
[(2 X 5) + (1 X 4) + (1 X 6)]/4
5 10 5.00
= 5.00
[(2 X 10) + (1 X 5) + (1 X 4)]/4
6 8 0.75
= 7.25
[(2 X 8) + (1 X 10) + (1 X 5)]/4
7 7 0.75
= 7.75
[(2 X 7) + (1 X 8) + (1 X 10)]/4
8 9 1.00
= 8.00
[(2 X 9) + (1 X 7) + (1 X 8)]/4
9 12 3.75
= 8.25
[(2 X 12) + (1 X 9) + (1 X 7)]/4
10 14 4.00
= 10.00
[(2 X 14) + (1 X 12) + (1 X 9)]/4
11 15 2.75
= 12.25
MAD
=
2.3125

Thus, since the MAD of the weighted moving average is smaller than of the 3year moving
average, the better method to forecast the sales will be the weighted moving average.
LE ANH KHOA BABAIU14124

516: Develop a trend line for the demand for fertilizer in Problem 515, using any computer
software.

Solution:
Let X be the time period (1, 2, ) and Y be the demand of fertilizer. Using the function Data
Analysis Regression of Excel, we have the following tables:
Regression Statistics
Multiple R 0.906914487
R Square 0.822493888
Adjusted R Square 0.802770986
Standard Error 1.712697677
Observations 11

ANOVA
df SS MS F Significance F
Regression 1 122.3272727 122.3272727 41.70247934 0.000117077
Residual 9 26.4 2.933333333
Total 10 148.7272727

Coefficients Standard Error tStat Pvalue

Intercept 2.218181818 1.107549848 2.002782829 0.076210093
YEAR (X) 1.054545455 0.163299316 6.457745686 0.000117077

So, based on the tables, we can see that the trend line for the demand for fertilizer will be:
= b0 + b1X = 2.22 + 1.05X
where X = time period (1, 2, ) and = predicted demand
LE ANH KHOA BABAIU14124

517: In Problems 515 and 516, three different forecasts were developed for the demand for
fertilizer. These three forecasts are a 3year moving average, a weighted moving average, and a
trend line. Which one would you use? Explain your answer.

Solution:
Based on the trend line ( = 2.22 + 1.05X), we have the following table:
YEAR PREDICTED
ACTUAL DEMAND ERROR
(X) DEMAND (Y)
1 4 3.27 0.73
2 6 4.32 1.68
3 4 5.38 1.38
4 5 6.44 1.44
5 10 7.49 2.51
6 8 8.55 0.55
7 7 9.60 2.60
8 9 10.65 1.65
9 12 11.71 0.29
10 14 12.76 1.24
11 15 13.82 1.18
MAD
= 1.39

Thus, since the MAD of the trend line is lowest, the trend line is the best method in order to
forecast the demand of fertilizer.
LE ANH KHOA BABAIU14124

518: Use exponential smoothing with a smoothing constant of 0.3 to forecast the demand for
fertilizer given in Problem 515. Assume that last periods forecast for year 1 is 5,000 bags to
begin the procedure. Would you prefer to use the exponential smoothing model or the weighted
average model developed in Problem 515? Explain your answer.

Solution:
We use the exponential smoothing with a smoothing constant of 0.3 to forecast the demand for
fertilizer. We have the table as following (with the new assumption):
FORECASTED
YEAR DEMAND ERROR
DEMAND
1 4,000 5,000 1,000
5,000 + (0.3)*(4,000 5,000)
2 6,000 1,300
= 4,700
4,700 + (0.3)*(6,000 4,700)
3 4,000 1,090
= 5,090
5,090 + (0.3)*(4,000 5,090)
4 5,000 237
= 4,763
4,763 + (0.3)*(5,000 4,763)
5 10,000 5,166
= 4,834
4,834 + (0.3)*(10,000 4,834)
6 8,000 1,616
= 6,384
6,384 + (0.3)*(8,000 6,384)
7 7,000 131
= 6,869
6,869 + (0.3)*(7,000 6,869)
8 9,000 2,092
= 6,908
6,908 + (0.3)*(9,000 6,908)
9 12,000 4,464
= 7,536
7,536 + (0.3)*(12,000 7,536)
10 14,000 5,125
= 8,857
8,875 + (0.3)*(14,000 8,875)
11 15,000 4,588
= 10,412
MAD (in thousands)
= 2,437

Thus, since the MAD of the weighted moving average is smaller than of the exponential
smoothing, the better method to forecast the sales will be the weighted moving average.
LE ANH KHOA BABAIU14124

525: Sales of industrial vacuum cleaners at R. Lowenthal Supply Co. over the past 13 months
are as follows:
SALES ($1,000s) MONTH SALES ($1,000s) MONTH
11 January 14 August
14 February 17 September
16 March 12 October
10 April 14 November
15 May 16 December
17 June 11 January
11 July

(a) Using a moving average with three periods, determine the demand for vacuum cleaners for
next February.
(b) Using a weighted moving average with three periods, determine the demand for vacuum
cleaners for February. Use 3, 2, and 1 for the weights of the most recent, second most recent, and
third most recent periods, respectively. For example, if you were forecasting the demand for
February, November would have a weight of 1, December would have a weight of 2, and January
would have a weight of 3.
(c) Evaluate the accuracy of each of these methods.
(d) What other factors might R. Lowenthal consider in forecasting sales?

Solution:
(a), (b)
Perio Mont Deman 3period 3period Weighted
Error Error
d h d Moving Average Moving Average
1 Jan. 11
2 Feb. 14
3 Mar. 16
(11 + 14 + 16)/3 [(3 X 16) + (2 X 14) + (1 X
4 Apr. 10 3.67 4.50
= 13.67 11)]/6 = 14.50
(14 + 16 + 10)/3 [(3 X 10) + (2 X 16) + (1 X
5 May 15 1.67 2.33
= 13.33 14)]/6 = 12.67
(16 + 10 + 15)/3 [(3 X 15) + (2 X 10) + (1 X
6 June 17 3.33 3.50
= 13.67 16)]/6 = 13.50
(10 + 15 + 17)/3 [(3 X 17) + (2 X 15) + (1 X
7 July 11 3.00 4.17
= 14.00 10)]/6 = 15.17
(15 + 17 + 11)/3 [(3 X 11) + (2 X 17) + (1 X
8 Aug. 14 0.33 0.33
= 14.33 15)]/6 = 13.67
(17 + 11 + 14)/3 [(3 X 14) + (2 X 11) + (1 X
9 Sept. 17 3.00 3.50
= 14.00 17)]/6 = 13.50
(11 + 14 + 17)/3 [(3 X 17) + (2 X 14) + (1 X
10 Oct. 12 2.00 3.00
= 14.00 11)]/6 = 15.00
(14 + 17 + 12)/3 [(3 X 12) + (2 X 17) + (1 X
11 Nov. 14 0.33 0
= 14.33 14)]/6 = 14.00
(17 + 12 + 14)/3 [(3 X 14) + (2 X 12) + (1 X
12 Dec. 16 1.67 2.17
= 14.33 17)]/6 = 13.83
LE ANH KHOA BABAIU14124

(12 + 14 + 16)/3 [(3 X 16) + (2 X 14) + (1 X

13 Jan. 11 3.00 3.67
= 14.00 12)]/6 = 14.67
(14 + 16 + 11)/3 [(3 X 11) + (2 X 16) + (1 X
14 Feb.
= 13.67 14)]/6 = 13.17
MAD
= MAD
= 2.72
2.20

(c) As can be seen from the table, the MAD for the moving average is 2.20, and the MAD for the
weighted moving average is 2.72. Thus, based on this analysis, the moving average appears to
be more accurate; and, the forecast for February is 13.67 (about 14).
(d) There are many other factors to consider, including seasonality and any underlying causal
variables such as advertising budget.
LE ANH KHOA BABAIU14124

527: Emergency calls to Winter Park, Floridas 911 system, for the past 24 weeks are as follows:
WEEK CALLS WEEK CALLS WEEK CALLS
1 50 9 35 17 55
2 35 10 20 18 40
3 25 11 15 19 35
4 40 12 40 20 60
5 45 13 55 21 75
6 35 14 35 22 50
7 20 15 25 23 40
8 30 16 55 24 65

(a) Compute the exponentially smoothed forecast of calls for each week. Assume an initial
forecast of 50 calls in the first week and use =0.1 . What is the forecast for the 25th week?
(b) Reforecast each period using =0.6
(c) Actual calls during the 25th week were 85. Which smoothing constant provides a superior
forecast?

Solution:
(a), (b), (c)
We have the following table:
Actua Forecas Forecas
Smoothed Value Smoothed Value
Week l t t
( =0.1 ) ( =0.6 )
Value Error Error
1 50 50.00 0 50.00 0
50.00 + (0.1)*(50 50.00) 50.00 + (0.6)*(50 50.00)
2 35 15.00 15.00
= 50.00 = 50
50.00 + (0.1)*(35 50.00) 50.00 + (0.6)*(35 50.00)
3 25 23.50 16.00
= 48.50 = 41.00
48.50 + (0.1)*(25 48.50) 41.00 + (0.6)*(25 41.00)
4 40 6.15 8.60
= 46.15 = 31.40
46.16 + (0.1)*(40 46.15) 31.40 + (0.6)*(40 31.40)
5 45 0.53 8.44
= 45.53 = 36.56
45.53 + (0.1)*(45 45.53) 36.56 + (0.6)*(45 36.56)
6 35 10.47 6.62
= 45.47 = 41.62
45.47 + (0.1)*(35 45.47) 41.62 + (0.6)*(35 41.62)
7 20 24.42 17.65
= 44.42 = 37.65
44.42 + (0.1)*(20 44.42) 37.65 + (0.6)*(20 37.65)
8 30 11.98 2.94
= 41.98 = 27.06
41.98 + (0.1)*(30 41.98) 27.06 + (0.6)*(30 27.06)
9 35 5.78 6.18
= 40.78 = 28.82
40.78 + (0.1)*(35 40.78) 28.82 + (0.6)*(35 28.82)
10 20 20.20 12.53
= 40.20 = 32.53
40.20 + (0.1)*(20 40.20) 32.53 + (0.6)*(20 32.53)
11 15 20.18 10.01
= 38.18 = 25.01
38.18 + (0.1)*(15 38.18) 25.01 + (0.6)*(15 25.01)
12 40 4.14 21.00
= 35.86 = 19.00
35.86 + (0.1)*(40 35.86) 19.00 + (0.6)*(40 19.00)
13 55 18.73 23.40
= 36.27 = 31.60
14 35 36.27 + (0.1)*(55 36.27) 3.14 31.60 + (0.6)*(55 31.60) 10.64
LE ANH KHOA BABAIU14124

= 38.14 = 45.64
38.14 + (0.1)*(35 38.14) 45.64 + (0.6)*(35 45.64)
15 25 12.83 14.26
= 37.83 = 39.26
37.83 + (0.1)*(25 37.83) 39.26 + (0.6)*(25 39.26)
16 55 18.45 24.30
= 36.55 = 30.70
36.55 + (0.1)*(55 36.55) 30.70 + (0.6)*(55 30.70)
17 55 16.60 9.72
= 38.40 = 45.28
38.40 + (0.1)*(55 38.40) 45.28 + (0.6)*(55 45.28)
18 40 0.06 11.11
= 40.06 = 51.11
40.06 + (0.1)*(40 40.06) 51.11 + (0.6)*(40 51.11)
19 35 5.50 9.44
= 40.05 = 44.44
40.05 + (0.1)*(35 40.05) 44.44 + (0.6)*(35 44.44)
20 60 20.45 21.22
= 39.55 = 38.78
39.55 + (0.1)*(60 39.55) 38.78 + (0.6)*(60 38.78)
21 75 33.40 23.49
= 41.60 = 51.51
41.60 + (0.1)*(75 41.60) 51.51 + (0.6)*(75 51.51)
22 50 5.06 15.60
= 44.94 = 65.60
44.94 + (0.1)*(50 44.94) 65.60 + (0.6)*(50 65.60)
23 40 5.45 16.24
= 45.45 = 56.24
45.45 + (0.1)*(40 45.45) 56.24 + (0.6)*(40 56.24)
24 65 20.09 18.50
= 44.91 = 46.50
44.91 + (0.1)*(65 44.91) 46.50 + (0.6)*(65 46.50)
25 85 38.08 27.40
= 46.92 = 57.60
MAD MAD
= 13.61 = 14.01

Thus, since the MAD of =0.1 is smaller than of =0.6 , the better smoothing constant
should be use to forecast is =0.1 .