For Exercises 13.46 to 13.49, see pages 693694; for 13.50, see page 699; and for 13.51 and 13.52, see pages 705706.

Question 13.49

13.49 Comparing models for hires rate with MAPE.

MAD (Exercise 13.47) measures the average magnitude of the prediction errors, and MSE (Exercise 13.48) measures the average squared magnitude of the prediction errors. To put the prediction errors in perspective, it can be useful to measure the errors in terms of percents. One such measure is mean absolute percentage error (MAPE):

where is the residual for period , is the actual observation for period , and is the number of available residuals.

mean absolute percentage error (MAPE)

  1. Compute MAPE to the hundredth place for residuals from the 1-month, 2-month, 3-month, and 4-month moving-average models determined in Exercise 13.46, part (c).
  2. What moving-average span value corresponds with the smallest MAPE?

13.49

(a) The MAPEs are: 1-month: 22.0042, 2-month: 17.3309, 3-month: 16.5842, 4-month: 17.2786. (b) The 3-month MAPE is the smallest.

hires