Forecasting
Assignment 1
PROBLEM 1: MA & EXPONENTIAL SMOOTHING
For a product the demand data has been shown in the table below for the year.
Compare the forecasts using a MA forecasting method with a period of 5 months
and an Exponential smoothing Method with an of 1/3. For Exponential
Smoothing use the midpoint of first 5 month range of the average as the initial
forecast. (Hint: the Exponential Smoothing Forecast therefore in March 2013 for
April 2013 will be 4951)
Month, t
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Demand,
x(t)
4576
5568
3240
5978
5395
4644
5880
6096
5967
5828
5808
6076
A. From June to December, which method of forecasting, outperforms the
other?
a. Exponential
b. Moving Average
c. None
d. Not enough information
B. From June to December, for how many months does one method of
forecasting, outperform the other?
ANS: ____________________
PROBLEM 2: SUGAR BON-BON CEREAL FORECASTS
You are working for a retailer that sells a lot of breakfast cereals. One of the
cereals that you sell is the popular Sugar Bon-Bon brand that features twice the
sugar and even more caffeine. The monthly demand for the previous 24 months
is shown in the following data file: SugarBonBon_Data.xlsx.
A. You decide to initialize the model by taking the average of the demand of
the first 4 time periods.
What is the average demand for periods 1 through 4? Keep all estimates
to one decimal place for this problem.
ANS: ____________________
B. Using a simple (single) exponential smoothing model determine the
forecast for consumption in month 25. Assume that the level smoothing
parameter, Alpha, is 0.15. Also, use your answer to the first part (A.) as
the value for x^(4,5).
ANS: ________________________
C. Using the same model, what is the estimate for month 30?
ANS: _____________________
D. What is the MAPE for this forecast through period 24?
ANS: ____________________
E. Take the data and chart it so that you can see demand on the y axis and
each month on the x axis.
What patterns appear to be present in the demand for Sugar Bon-Bons?
a. Only level, it appears to be stationary demand
b. A positive trend
c. A negative trend
d. Seasonality across an entire year with some high months and some low
months
e. There is not enough data to even make a guess
F. We now want to apply Holt's model (level and trend) to the data where
alpha = 0.20 and Beta = 0.05. You have initialized over the first four
periods and have found that a^(4)=157.5 and b^(4)= 19.1
Now apply Holt's model (level and trend) to the data where alpha = 0.20
and Beta = 0.05. What is the forecast for period 5 using these initialized
values?
ANS: ______________________
G. What is the forecast for period 6 from period 5, that is, x^(5,6)?
ANS: ______________________
H. What is the forecast for period 25 from period 24, that is, x^(24,25)?
ANS: ______________________
I.
What is the forecast for period 30 from period 24, that is, x^(24,30)?
ANS: ______________________
J.
What is the MAPE for this forecast through period 24?
ANS: ______________________
K. Which model do you prefer?
a. Simple Exponential Model (B)
b. Holt Model with Level and Trend (F)
c. They are equal in performance
d. There is not enough data to decide
L. Using the model developed in part 1.3, you notice that using an Alpha
value of 0.99 while keeping Beta at 0.05 gives you a MAPE of 11.3%. What
do you recommend?
a. Stay with Alpha = 0.20
b. Go with Alpha = 0.99
c. Select an Alpha in the middle (~0.66)
d. There is not enough data to decide
PRACTICE PROBLEM 3. TRAINMAX
TrainMax is a manufacturer of high-end specialty engine equipment for high
speed trains. They produce parts that are sent to the original equipment
manufacturers (OEMs) for manufacturing new engines. They face a continuing
challenge of trying to forecast demand for their products. The demand for one
part in particular, XC-288, was highlighted as needing to be examined.
You can download the spreadsheet with the recent demand data for XC-288 from
here: TrainMax_Data.xlsx.
A. You want to develop an exponential smoothing model to forecast demand.
Start with a simple exponential smoothing model with just level. Assume
that the forecast for period 1 is the actual demand for period 1, that is,
x^(0,1) = x(1) = 1027. Set the Alpha to 0.12
What is the forecast for period 15 from period 14, that is, x^(14,15)?
ANS: _____________________
B. What is the MAPE for these forecasts?
ANS: _____________________
C. Looking at the data and the forecast, can you recommend a different,
better value for Alpha?
ANS: ____________________
