UNIVERSITY OF TECHNOLOGY, JAMAICA

FACULTY OF BUSINESS AND MANAGEMENT

SCHOOL OF BUSINESS ADMINISTRATION
DECISION SCIENCE
TUTORIAL WORKSHEET – FORECASTING
1. Pineapple sales at Whit’s Pineapple Farm in St. Mary are shown in the left column in the table
below:

MOVING AVERAGE CALCULATIONS -  

PINEAPPLE SALES BY  
MONTH

MONTH ACTUAL FORECAST USING 3 - Absolute Error FOR 3 - 4 - WEEK MOVING

PINEAPPLE WEEK MOVING WEEK MOVING AVERAGE
SALES (1000’s) AVERAGE AVERAGE FORECAST FORECAST

January 10 -    
February 12 -    
March 13 -    
April 16 (13+12+10)/3 = (16 – 11.67) =  
11.67 4.33
May 19 (16+13+12)/3=13.67 (19 – 13.67) =  
5.33
June 23 (19+16+13)/3=16.00 (23 – 16) = 7  
July 26 (23+19+16)/3 = (26 – 19.33) =  
19.33 6.67
August 30 (26+23+19)/3=22.67 (30 – 22.67) =  
7.33
September 28 (30+26+23)/3 = (28 – 26.33) =  
26.33 1.67
October 18 (28+30+26)/3= (18 – 28) = 10  
28.00
November 16 (18+28+30)/3 = (16 – 25.33) =  
25.33 9.33
December 14 (16+18+28)/3 = (14 – 20.67) =  
20.67 6.67
SUM OF
ABSOLUTE
ERROR = 58.33.
Therefore MAD is
58.33/9 = 6.48
JANUARY -

a. Compute 4 – and 5 – month moving averages for the time series.

b. Compute the MAD for the 4 – and 5 – month moving averages forecasts.
c. Based on the MAD error analysis, which is the preferred forecasting
method? Remember that the MAD for the 3-month moving average is 6.48.
d. Use the preferred method to forecast sale of pineapples in January of the
following year.
e. Use a weight of 6 for the most recent observation, 3 for the second most
recent, and 2 for the third most recent to compute a 3-month weighted
moving average for the time series.
f. Compute the MAD for the weighted moving average in part (e). Do you
prefer this weighted moving average to the 3-month simple moving
average? Remember that the MAD for the 3-month simple moving average
is 6.48

g. Use Excel to do the calculations

1
 2. Port Authority of Jamaica has unloaded large quantities of rice from Guyana
from ships during the past eight quarters. The Port’s Operations Manager wants to
test the use of exponential smoothing to see how well the technique works in
predicting tonnage of rice unloaded. He assumes that the forecast of grain unloaded
in the first quarter was 175 tons. Two values of alpha are examined: α = 0.10 and α =
0.50. Using the information below on actual tonnage unloaded:

Quarter Actual Forecast MAD for α Forecast MAD for α

tonnage using α = = 0.10 using α = = 0.50
unloaded 0.10 0.50
1 180 175
2 168
3 159
4 175
5 190
6 205
7 180
8 182
9 -

a. Show the exponential smoothing forecasts using α = 0.1 and α = 0.5 .

b. (i). Evaluate the accuracy of each smoothing constant using the Mean
Absolute Deviation (MAD) criterion.
(ii). Which smoothing constant would you prefer? (Remember to include
why you prefer that smoothing constant).
c. Use the preferred method to forecast pineapple sale for quarter 9.
d. Use Excel to do the calculations.

3. KSAC building contracts for a 12-month period (in $millions) are 240,
350, 230, 260, 280, 320, 220, 310, 240, 310, 240, and 230.

a. Compare a 4-month moving average forecast with an exponential smoothing

forecast using ą = 0.3. Which provides the better forecasts?
b. What is the forecast for the next month?

FOR QUESTIONS 5 AN6 STUDENTS ARE EXPECTED TO USE BOTH

MANUAL AND EXCEL TO CONSTRUCT THEIR ANSWERS.

4. The Port Authority of Jamaica recently asked Correctional Services to

relocate its women prisoners from Fort August as it needs the space to
facilitate expansion of its container terminal. The port has experienced the
following container throughput during the past 12 years, expressed as TEUs
(i.e., 20-foot equivalent unit, a standard unit of measure for containers):

YEAR 1 2 3 4 5 6 7 8 9 10 11 12
TEUs 526.1 549.4 606.0 627.0 695.7 734.9 761.1 845.4 1021.1 1,137.6 1,173.6 1,233.4
(1,000s)

a) Develop the model (equation) for the linear trend line for these time series
using the formulae.

2
 b) Comment on what is happening to container throughput at the Port
Authority.

c) Calculate the value of the Product Moment Co-efficient of

Correlation. Comment on the strength of the relationship between
the independent and dependent variables. (Remember -1 ≤ r ≤
+1)

d) What is the percentage of the variation in TEUs that is a result of

the behavior of time i.e., what is the value of the Co-efficient of
Determination? Is the model a good estimator? Explain.
(Remember 0 ≤ r² ≤ 1)

e) Use Excel add-in (Regression) to verify the model you developed.

5. The management of a chain of fast food restaurants wants to investigate the

relationship between the daily sales volume (in dollars) of a company restaurant and
the number of competitor restaurants within a 1 – kilometre radius. The following
data have been collected.

NUMBER OF
COMPETITORS SALES
WITHIN 1 KM
1 3600
1 3300
2 3100
3 2900
3 2700
4 2500
5 2300
5 2000

a. Develop the least squares estimated regression equation (using both

excel and manual computation) that relates daily sales volume to the
number of competitor restaurants within a 1-kilometre radius.

b. What is the value of the Product Moment Co-efficient of Correlation?

What is the relationship between the variables.

c. What is the value of the Co-efficient of Determination? Is the model a

good estimator? Explain.
d. Use the estimated regression equation developed in part (a) to forecast
the daily sales volume for a particular company restaurant that has seven
(7) competitors within a 1-kilometre radius.

3
6. Administrators at State University believes that decreases in the number of first year applications
that they have experienced are directly related to tuition increases. They have collected the
following enrollment and tuition data for the past decade

APPLICATION TUITION
6,050 $3,600
4,060 3,600
5,200 4,000
4,410 4,400
4,380 4,500
4,160 5,700
3,560 6,000
2,970 6,000
3,280 7,500
3,430 8,000

a) Develop a linear regression model for these data and forecast the number of applications for
State University if tuition increases to $9,000 per year and if tuition is lowered to $7,000 per
year.
b) Determine the strength and direction of the linear relationship between first year applications
and tuition.
c) Is this model a good estimator of the number of first year students who will apply to State
University?

7. Infoworks is a large computer discount store that sells computers and ancillary equipment and
software in the town where State University is located. Infoworks has collected historical data on
computer sales and printers as follows:

Personal computers sold Printers sold

1.045 326
1,610 510
860 296
1,211 478
975 305
1,117 506
1,066 612
1,310 560
1,517 590
1,246 676

e. Develop the least squares estimated regression equation (using both excel and
manual computation) that relates daily sales volume to the number of competitor
restaurants within a 1-kilometre radius.

f. What is the value of the Product Moment Co-efficient of Correlation? What is the
relationship between the variables.

g. What is the value of the Co-efficient of Determination? Is the model a good

estimator? Explain.

h. If 1,300 computers are sold what would be the expected demand for printers?

4
8. A prediction equation for starting salaries (in $1,000’s) and SAT scores was performed using

simple linear regression. In the regression printout shown below:

SUMMARY OUTPUT

Regression Statistics
Multiple R 0.9350181
23
R Square 0.8742588
9
Adjusted R 0.8602876
Square 55
Standard Error 3.3072959
49
Observations 11

ANOVA
  df SS MS F Significanc
eF
Regression 1 684.4652324 684.46 62.575 2.42144E-
52 64 05
Residual 9 98.44385847 10.938
21
Total 10 782.9090909      

  Coefficient Standard t Stat P-value Lower 95%

s Error
Intercept - 8.660080997 - 0.0083 -
29.140604 3.3649 24 48.7310838
68 3 7
SAT 0.0654438 0.008273059 7.9104 2.42E- 0.04672886
41 76 05 6

a. Develop the least squares estimated regression equation based on the

Summary output above, correct to 2 decimal places (3 marks)

b. Is SAT is a good predictor for starting salary? Why?

c. Use the equation developed in (a) to forecast starting salary when SAT is 700.
===============================

9. A prediction equation for starting salaries (in $1,000’s) and SAT scores was performed using

simple linear regression. In the regression printout shown below:

SUMMARY OUTPUT

Regression Statistics
Multiple R 0.9350181
23
R Square 0.8742588
9
Adjusted R 0.8602876
Square 55
Standard Error 3.3072959
49
Observations 11

5
ANOVA
  df SS MS F
Significance
F
Regression 1 684.4652324 684.465 62.5756 2.42144E-
2 4 05
Residual 9 98.44385847 10.9382
1
Total 10 782.9090909      

  Coefficient Standard t Stat P-value Lower 95%

s Error
Intercept - 8.660080997 - 0.00832 -
29.140604 3.36493 4 48.7310838
68 7
SAT 0.0654438 0.008273059 7.91047 2.42E- 0.04672886
41 6 05 6

a. Develop the least squares estimated regression equation based on the Summary output above, correct to 2
decimal places (3 marks)

b. Is SAT is a good predictor for starting salary? Why?

b. Use the equation developed in (a) to forecast starting salary when SAT is 700.
===============================