Answer: Demand Pattern A shows a strong linear trend from 103 up to 207 but Demand Pattern B
fluctuates around 169 and (with a standard deviation of 4) is purely random. Adjusted exponential
smoothing or linear regression are the only suitable methods for Demand Pattern A. A linear equation of
the form Y = 95.98 + 5.82X yields an R squared of 0.99. The randomness in Demand Pattern B is pure noise
and therefore not predictable. Moving averages with a large n or exponential smoothing with a low alpha
will smooth this randomness but don’t truly add much in the way of predictive ability.
Diff: 3
Reference: 9.5 Time Series Forecasting Models
Keywords: time series, trend, adjusted exponential smoothing, regression, moving average, randomness, exponential
smoothing
AACSB: Application of Knowledge
LO: 9.2: Apply a variety of time series forecasting models, including moving average, exponential smoothing, and
linear regression models.
30) Develop forecasts for periods 7 through 10 for the demand data in the table using a three period
moving average, a weighted moving average using 0.6, 0.3, and 0.1, and exponential smoothing with
alpha = 0.7. Use a 6th period forecast of 135 as the starting point for the exponential smoothing technique.
1
2
3
4
5
115
6
135
7
137
8
144
9
153