problem:
Scott Armstrong, the managing editor of Your Horoscope magazine, needs to develop a forecasting system for monthly newsstand sales in order to schedule press runs. Sales in thousands of copies for the first 7 months of publication were:
Year
|
Month
|
Sales
|
2013
|
August
|
50
|
|
September
|
55
|
|
October
|
65
|
|
November
|
74
|
|
December
|
80
|
2014
|
January
|
76
|
|
February
|
86
|
Scott does not believe there is a seasonal pattern. He is considering three different forecasting models: three-period moving average, simple exponential smoothing with a = 0.7 , and time series regression (i.e., trend projection).
Required:
Question 1) Determine the best forecasting model among the above three methods, and develop a forecast for March 2014.
Question 2) Use the first four months as the warm-up sample period and the remaining months as the forecasting sample period. The criteria of measuring forecast accuracy to be used are Bias, MAD, MSE, and MAPE.
Solve the given numerical problem and illustrate step by step calculation.