Lesson 7: Measuring Forecast Error

February 10, 2016

Previous | Next

Wed, Feb 10

Review:

• Time Series Forecasting

Presentation:

• Forecast Error
  • MAD – mean absolute deviation
    • MAD = ∑|Actual – Forecast|/(Number of Forecasts)
  • MSE – mean square error
    • MSE = ∑(Actual – Forecast)^2/(Number of Forecasts)
  • MAD easier to interpret than MSE
  • Example Prob. 15.1 and 15.3 on p. 607

Activity:

Time Period	Acres Harvested
1	140.0
2	141.7
3	134.6
4	131.7
5	131.9
6	134.3
7	135.2
8	131.0

• Use the data above to complete the following
  • Create a linear trend model
  • Generate forecasts for Time Periods 9, 10 and 11 (3 separate forecasts)
  • Calculate forecast error using MAD and MSE assuming the following “actual” amounts
    • Time Period 9: 120.6
    • Time Period 10: 115.2
    • Time Period 11: 114.5

Assignment:

• Repeat the activity above in Sheets or Excel.